urbit/shrub
1
1
Fork 0
mirror of https://github.com/urbit/shrub.git synced 2024-12-19 08:32:39 +03:00
Code Issues Packages Projects Releases Wiki Activity
3f63aa00e6
shrub/pkg/arvo/lib/language-server/complete.hoon

387 lines
9.0 KiB
Plaintext
Raw Normal View History

language-server: multiple files and ford Handle multiple files by keeping a map of text buffers. Also use the Ford parser so we can parse ford runes. At some point we should load in libraries when that happens so we have the appropriate types. This corresponds to hoon-language-server 0.1.1
2019-11-05 05:12:23 +03:00
/+ language-server-parser
auto: fix some crashes on strange wet gates
2019-11-01 07:37:24 +03:00
:: Autocomplete for hoon.
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
::
language-server: cleanup and incremental text sync
2019-11-04 08:32:42 +03:00
=/ debug |
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
|%
language-server: genericise searching gates genericise searching gates, as we are not always searching against a [term type]. e.g. chat-cli autocomplete
2019-11-27 05:49:31 +03:00
+* option [item]
language-server: revive rune/symbol completion
2020-01-28 05:24:39 +03:00
[term=cord detail=item]
dojo, auto: move insert-magic logic to lib/auto This changes the entry-points in lib/auto so that clients never have to handle magic-spoon. You can specify either a tape of code with a position index or a preparsed hoon (presumably you ran +insert-magic before parsing).
2019-11-01 00:56:54 +03:00
::
language-server: initial commit A simple language server engine, for use with hoonls.py, which presents the RPC interface expected by editors. Features: - Syntax error detection - Rune snippets - Autocomplete
2019-11-04 06:18:40 +03:00
:: Like +rose except also produces line number
::
++ lily
various: move away from {type} syntax In favor of [type] syntax. Turns a bunch of ++ into +$ along the way.
2020-11-26 17:09:31 +03:00
|* [los=tape sab=rule]
language-server: initial commit A simple language server engine, for use with hoonls.py, which presents the RPC interface expected by editors. Features: - Syntax error detection - Rune snippets - Autocomplete
2019-11-04 06:18:40 +03:00
=+ vex=(sab [[1 1] los])
?~ q.vex
[%| p=p.vex(q (dec q.p.vex))]
?. =(~ q.q.u.q.vex)
[%| p=p.vex(q (dec q.p.vex))]
[%& p=p.u.q.vex]
::
dojo, auto: move insert-magic logic to lib/auto This changes the entry-points in lib/auto so that clients never have to handle magic-spoon. You can specify either a tape of code with a position index or a preparsed hoon (presumably you ran +insert-magic before parsing).
2019-11-01 00:56:54 +03:00
:: Get all the identifiers accessible if this type is your subject.
::
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
++ get-identifiers
|= ty=type
%- flop
language-server: genericise searching gates genericise searching gates, as we are not always searching against a [term type]. e.g. chat-cli autocomplete
2019-11-27 05:49:31 +03:00
|- ^- (list (option type))
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
?- ty
%noun ~
%void ~
[%atom *] ~
[%cell *]
%+ weld
$(ty p.ty)
$(ty q.ty)
::
[%core *]
%- weld
auto: don't look in context of non-gold cores
2019-11-01 02:38:47 +03:00
:_ ?. ?=(%gold r.p.q.ty)
~
$(ty p.ty)
language-server: genericise searching gates genericise searching gates, as we are not always searching against a [term type]. e.g. chat-cli autocomplete
2019-11-27 05:49:31 +03:00
^- (list (option type))
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
%- zing
%+ turn ~(tap by q.r.q.ty)
|= [term =tome]
%+ turn
~(tap by q.tome)
|= [name=term =hoon]
^- (pair term type)
~| term=term
[name ~(play ~(et ut ty) ~[name] ~)]
::
[%face *]
?^ p.ty
~
[p.ty q.ty]~
::
[%fork *]
auto: support forks
2019-11-01 09:21:10 +03:00
%= $
ty
=/ tines ~(tap in p.ty)
?~ tines
%void
|- ^- type
?~ t.tines
i.tines
(~(fuse ut $(tines t.tines)) i.tines)
==
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
::
[%hint *] $(ty q.ty)
[%hold *] $(ty ~(repo ut ty))
==
::
language-server: add function signature on hover Add function signature on hover. Fixes bug around autocomplete not working properly when the cursor was mid-word.
2019-11-12 18:57:01 +03:00
++ search-exact
language-server: genericise searching gates genericise searching gates, as we are not always searching against a [term type]. e.g. chat-cli autocomplete
2019-11-27 05:49:31 +03:00
|* [sid=term options=(list (option))]
=/ match
%+ skim options
language-server: revive rune/symbol completion
2020-01-28 05:24:39 +03:00
|= [id=cord *]
language-server: add function signature on hover Add function signature on hover. Fixes bug around autocomplete not working properly when the cursor was mid-word.
2019-11-12 18:57:01 +03:00
=(sid id)
?~ match
~
[~ i.match]
::
dojo, auto: move insert-magic logic to lib/auto This changes the entry-points in lib/auto so that clients never have to handle magic-spoon. You can specify either a tape of code with a position index or a preparsed hoon (presumably you ran +insert-magic before parsing).
2019-11-01 00:56:54 +03:00
:: Get all the identifiers that start with sid.
::
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
++ search-prefix
language-server: revive rune/symbol completion
2020-01-28 05:24:39 +03:00
|* [sid=cord ids=(list (option))]
language-server: genericise searching gates genericise searching gates, as we are not always searching against a [term type]. e.g. chat-cli autocomplete
2019-11-27 05:49:31 +03:00
^+ ids
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
%+ skim ids
language-server: revive rune/symbol completion
2020-01-28 05:24:39 +03:00
|= [id=cord *]
^- ?(%.y %.n)
hoon: renames +new-end, +new-lsh, +new-rsh
2020-12-02 11:47:29 +03:00
=(sid (end [3 (met 3 sid)] id))
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
::
dojo, auto: move insert-magic logic to lib/auto This changes the entry-points in lib/auto so that clients never have to handle magic-spoon. You can specify either a tape of code with a position index or a preparsed hoon (presumably you ran +insert-magic before parsing).
2019-11-01 00:56:54 +03:00
:: Get the longest prefix of a list of identifiers.
::
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
++ longest-match
language-server: genericise searching gates genericise searching gates, as we are not always searching against a [term type]. e.g. chat-cli autocomplete
2019-11-27 05:49:31 +03:00
|= matches=(list (option))
language-server: revive rune/symbol completion
2020-01-28 05:24:39 +03:00
^- cord
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
?~ matches
''
=/ n 1
=/ last (met 3 term.i.matches)
|- ^- term
?: (gth n last)
term.i.matches
hoon: renames +new-end, +new-lsh, +new-rsh
2020-12-02 11:47:29 +03:00
=/ prefix (end [3 n] term.i.matches)
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
?: |- ^- ?
?| ?=(~ t.matches)
hoon: renames +new-end, +new-lsh, +new-rsh
2020-12-02 11:47:29 +03:00
?& =(prefix (end [3 n] term.i.t.matches))
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
$(t.matches t.t.matches)
== ==
$(n +(n))
hoon: renames +new-end, +new-lsh, +new-rsh
2020-12-02 11:47:29 +03:00
(end [3 (dec n)] term.i.matches)
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
::
dojo, auto: move insert-magic logic to lib/auto This changes the entry-points in lib/auto so that clients never have to handle magic-spoon. You can specify either a tape of code with a position index or a preparsed hoon (presumably you ran +insert-magic before parsing).
2019-11-01 00:56:54 +03:00
:: Run +find-type safely, printing the first line of the stack trace on
:: error.
::
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
++ find-type-mule
|= [sut=type gen=hoon]
^- (unit [term type])
=/ res (mule |.((find-type sut gen)))
?- -.res
%& p.res
language-server: initial commit A simple language server engine, for use with hoonls.py, which presents the RPC interface expected by editors. Features: - Syntax error detection - Rune snippets - Autocomplete
2019-11-04 06:18:40 +03:00
%| ((slog (flop (scag 10 p.res))) ~)
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
==
::
:: Get the subject type of the wing where you've put the "magic-spoon".
::
++ find-type
|= [sut=type gen=hoon]
=* loop $
|^
^- (unit [term type])
?- gen
[%cnts [%magic-spoon ~] *] `['' sut]
[%cnts [%magic-spoon @ ~] *] `[i.t.p.gen sut]
auto: support autocomplete inside wings
2019-11-01 08:43:50 +03:00
[%cnts [%magic-spoon @ *] *]
%= $
sut (~(play ut sut) wing+t.t.p.gen)
t.p.gen t.p.gen(t ~)
==
::
[%cnts [%magic-fork @ ~] *]
`['' (~(play ut sut) wing+t.p.gen)]
::
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
[^ *] (both p.gen q.gen)
[%brcn *] (grow q.gen)
hoon: bscl to bccl, etc
2020-11-17 10:05:17 +03:00
[%brpt *] (grow q.gen)
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
[%cnts *]
|- ^- (unit [term type])
=* inner-loop $
?~ q.gen
~
%+ replace
loop(gen q.i.q.gen)
|. inner-loop(q.gen t.q.gen)
::
[%dtkt *] (spec-and-hoon p.gen q.gen)
[%dtls *] loop(gen p.gen)
[%rock *] ~
[%sand *] ~
[%tune *] ~
[%dttr *] (both p.gen q.gen)
[%dtts *] (both p.gen q.gen)
[%dtwt *] loop(gen p.gen)
[%hand *] ~
[%ktbr *] loop(gen p.gen)
[%ktls *] (both p.gen q.gen)
hoon: bscl to bccl, etc
2020-11-17 10:05:17 +03:00
[%ktpm *] loop(gen p.gen)
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
[%ktsg *] loop(gen p.gen)
[%ktwt *] loop(gen p.gen)
[%note *] loop(gen q.gen)
[%sgzp *] (both p.gen q.gen)
hoon: bscl to bccl, etc
2020-11-17 10:05:17 +03:00
[%sggr *] loop(gen q.gen) :: should check for hoon in p.gen
[%tsgr *] (change p.gen q.gen)
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
[%tscm *]
%+ replace
loop(gen p.gen)
|.(loop(gen q.gen, sut (~(busk ut sut) p.gen)))
::
[%wtcl *] (bell p.gen q.gen r.gen)
[%fits *] (both p.gen wing+q.gen)
[%wthx *] loop(gen wing+q.gen)
[%dbug *] loop(gen q.gen)
[%zpcm *] (both p.gen q.gen)
[%lost *] loop(gen p.gen)
[%zpmc *] (both p.gen q.gen)
[%zpts *] loop(gen p.gen)
hoon: bscl to bccl, etc
2020-11-17 10:05:17 +03:00
[%zppt *] (both q.gen r.gen)
language-server: prevent %play-open on %zpgl
2021-01-29 04:44:37 +03:00
[%zpgl *] (spec-and-hoon p.gen q.gen)
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
[%zpzp *] ~
*
=+ doz=~(open ap gen)
?: =(doz gen)
~_ (show [%c 'hoon'] [%q gen])
~> %mean.'play-open'
!!
loop(gen doz)
==
::
++ replace
|= [a=(unit [term type]) b=(trap (unit [term type]))]
^- (unit [term type])
?~(a $:b a)
::
++ both
|= [a=hoon b=hoon]
(replace loop(gen a) |.(loop(gen b)))
::
++ bell
|= [a=hoon b=hoon c=hoon]
auto: gain and lose types on ?:
2019-11-02 04:02:20 +03:00
%+ replace loop(gen a)
|. %+ replace loop(gen b, sut (~(gain ut sut) a))
|. loop(gen c, sut (~(lose ut sut) a))
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
::
++ spec-and-hoon
|= [a=spec b=hoon]
(replace (find-type-in-spec sut a) |.(loop(gen b)))
::
++ change
|= [a=hoon b=hoon]
(replace loop(gen a) |.(loop(gen b, sut (~(play ut sut) a))))
::
++ grow
|= m=(map term tome)
=/ tomes ~(tap by m)
|- ^- (unit [term type])
=* outer-loop $
?~ tomes
~
=/ arms ~(tap by q.q.i.tomes)
|- ^- (unit [term type])
=* inner-loop $
?~ arms
outer-loop(tomes t.tomes)
%+ replace
loop(gen q.i.arms, sut (~(play ut sut) gen))
|. inner-loop(arms t.arms)
--
::
:: Not implemented yet. I wonder whether we should modify types found
:: in spec mode such that if it's a mold that produces a type, it
:: should just display the type and not that it's technically a
:: function.
::
++ find-type-in-spec
|= [sut=type pec=spec]
^- (unit [term type])
language-server: initial commit A simple language server engine, for use with hoonls.py, which presents the RPC interface expected by editors. Features: - Syntax error detection - Rune snippets - Autocomplete
2019-11-04 06:18:40 +03:00
~
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
::
language-server: break up parser from shoe
2021-01-29 05:01:33 +03:00
++ get-id-sym
|= [pos=@ud =tape]
%^ get-id pos tape
^- $-(nail (like (unit @t)))
;~(sfix (punt sym) (star ;~(pose prn (just `@`10))))
::
++ get-id-cord
|= [pos=@ud =tape]
%^ get-id pos tape
^- $-(nail (like (unit @t)))
;~(sfix (punt (cook crip (star prn))) (star ;~(pose prn (just `@`10))))
::
language-server: add function signature on hover Add function signature on hover. Fixes bug around autocomplete not working properly when the cursor was mid-word.
2019-11-12 18:57:01 +03:00
++ get-id
language-server: break up parser from shoe
2021-01-29 05:01:33 +03:00
|= [pos=@ud txt=tape seek=$-(nail (like (unit @t)))]
^- [forward=(unit @t) backward=(unit @t) id=(unit @t)]
=/ forward=(unit @t)
(scan (slag pos txt) seek)
=/ backward=(unit @t)
%- (lift |=(t=@t (swp 3 t)))
(scan (flop (scag pos txt)) seek)
=/ id=(unit @t)
auto: handle tab in middle of symbol
2019-11-01 10:18:34 +03:00
?~ forward
?~ backward
~
`u.backward
?~ backward
`u.forward
`(cat 3 u.backward u.forward)
language-server: add function signature on hover Add function signature on hover. Fixes bug around autocomplete not working properly when the cursor was mid-word.
2019-11-12 18:57:01 +03:00
[forward backward id]
::
:: Insert magic marker in hoon source at the given position.
::
++ insert-magic
|= [pos=@ud txt=tape]
^- [back-pos=@ud fore-pos=@ud txt=tape]
:: Find beg-pos by searching backward to where the current term
:: begins
language-server: break up parser from shoe
2021-01-29 05:01:33 +03:00
=+ (get-id-sym pos txt)
auto: handle tab in middle of symbol
2019-11-01 10:18:34 +03:00
=/ back-pos
?~ backward
pos
(sub pos (met 3 u.backward))
=/ fore-pos
?~ forward
dojo, auto: move insert-magic logic to lib/auto This changes the entry-points in lib/auto so that clients never have to handle magic-spoon. You can specify either a tape of code with a position index or a preparsed hoon (presumably you ran +insert-magic before parsing).
2019-11-01 00:56:54 +03:00
pos
auto: handle tab in middle of symbol
2019-11-01 10:18:34 +03:00
(add pos (met 3 u.forward))
:+ back-pos fore-pos
dojo, auto: move insert-magic logic to lib/auto This changes the entry-points in lib/auto so that clients never have to handle magic-spoon. You can specify either a tape of code with a position index or a preparsed hoon (presumably you ran +insert-magic before parsing).
2019-11-01 00:56:54 +03:00
:: Insert "magic-spoon" marker so +find-type can identify where to
:: stop.
::
;: weld
auto: handle tab in middle of symbol
2019-11-01 10:18:34 +03:00
(scag back-pos txt)
auto: support autocomplete inside wings
2019-11-01 08:43:50 +03:00
?: &(?=(~ id) ?=([%'.' *] (slag pos txt)))
"magic-fork"
dojo, auto: move insert-magic logic to lib/auto This changes the entry-points in lib/auto so that clients never have to handle magic-spoon. You can specify either a tape of code with a position index or a preparsed hoon (presumably you ran +insert-magic before parsing).
2019-11-01 00:56:54 +03:00
"magic-spoon"
?~ id
""
"."
auto: handle tab in middle of symbol
2019-11-01 10:18:34 +03:00
(slag back-pos txt)
dojo, auto: move insert-magic logic to lib/auto This changes the entry-points in lib/auto so that clients never have to handle magic-spoon. You can specify either a tape of code with a position index or a preparsed hoon (presumably you ran +insert-magic before parsing).
2019-11-01 00:56:54 +03:00
"\0a"
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
==
::
dojo, auto: move insert-magic logic to lib/auto This changes the entry-points in lib/auto so that clients never have to handle magic-spoon. You can specify either a tape of code with a position index or a preparsed hoon (presumably you ran +insert-magic before parsing).
2019-11-01 00:56:54 +03:00
:: Produce the longest possible advance without choosing between
:: matches.
::
:: Takes a +hoon which has already has a magic-spoon marker. Useful if
:: you want to handle your own parsing.
::
++ advance-hoon
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
|= [sut=type gen=hoon]
%+ bind (find-type-mule sut gen)
|= [id=term typ=type]
language-server: genericise searching gates genericise searching gates, as we are not always searching against a [term type]. e.g. chat-cli autocomplete
2019-11-27 05:49:31 +03:00
=/ matches=(list (option type))
(search-prefix id (get-identifiers typ))
(longest-match matches)
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
::
dojo, auto: move insert-magic logic to lib/auto This changes the entry-points in lib/auto so that clients never have to handle magic-spoon. You can specify either a tape of code with a position index or a preparsed hoon (presumably you ran +insert-magic before parsing).
2019-11-01 00:56:54 +03:00
:: Same as +advance-hoon, but takes a position and text directly.
::
++ advance-tape
|= [sut=type pos=@ud code=tape]
(advance-hoon sut (scan txt:(insert-magic pos code) vest))
::
:: Produce a list of matches.
::
:: Takes a +hoon which has already has a magic-spoon marker. Useful if
:: you want to handle your own parsing.
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
::
dojo, auto: move insert-magic logic to lib/auto This changes the entry-points in lib/auto so that clients never have to handle magic-spoon. You can specify either a tape of code with a position index or a preparsed hoon (presumably you ran +insert-magic before parsing).
2019-11-01 00:56:54 +03:00
++ tab-list-hoon
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
|= [sut=type gen=hoon]
language-server: genericise searching gates genericise searching gates, as we are not always searching against a [term type]. e.g. chat-cli autocomplete
2019-11-27 05:49:31 +03:00
^- (unit (list (option type)))
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
%+ bind (find-type-mule sut gen)
|= [id=term typ=type]
(search-prefix id (get-identifiers typ))
::
dojo, auto: move insert-magic logic to lib/auto This changes the entry-points in lib/auto so that clients never have to handle magic-spoon. You can specify either a tape of code with a position index or a preparsed hoon (presumably you ran +insert-magic before parsing).
2019-11-01 00:56:54 +03:00
:: Same as +advance-hoon, but takes a position and text directly.
::
++ tab-list-tape
language-server: cleanup and incremental text sync
2019-11-04 08:32:42 +03:00
|= [sut=type pos=@ud code=tape]
language-server: genericise searching gates genericise searching gates, as we are not always searching against a [term type]. e.g. chat-cli autocomplete
2019-11-27 05:49:31 +03:00
^- (each (unit (list (option type))) [row=@ col=@])
language-server: cleanup and incremental text sync
2019-11-04 08:32:42 +03:00
~? > debug %start-magick
language-server: initial commit A simple language server engine, for use with hoonls.py, which presents the RPC interface expected by editors. Features: - Syntax error detection - Rune snippets - Autocomplete
2019-11-04 06:18:40 +03:00
=/ magicked txt:(insert-magic pos code)
language-server: cleanup and incremental text sync
2019-11-04 08:32:42 +03:00
~? > debug %start-parsing
language-server: convert to fusion
2020-05-14 09:26:52 +03:00
=/ res (lily magicked (language-server-parser *path))
language-server: initial commit A simple language server engine, for use with hoonls.py, which presents the RPC interface expected by editors. Features: - Syntax error detection - Rune snippets - Autocomplete
2019-11-04 06:18:40 +03:00
?: ?=(%| -.res)
language-server: cleanup and incremental text sync
2019-11-04 08:32:42 +03:00
~? > debug [%parsing-error p.res]
language-server: initial commit A simple language server engine, for use with hoonls.py, which presents the RPC interface expected by editors. Features: - Syntax error detection - Rune snippets - Autocomplete
2019-11-04 06:18:40 +03:00
[%| p.res]
:- %&
language-server: cleanup and incremental text sync
2019-11-04 08:32:42 +03:00
~? > debug %parsed-good
clay: added mark runes; can boot off solid pill
2021-02-17 22:31:20 +03:00
((cury tab-list-hoon sut) hoon:`pile:clay`p.res)
dojo: add autocomplete for various dojo syntaxes Adds autocomplete for - :app poke syntax - :app|gen poke generator syntax - +gen naked generator syntax - =var variable syntax
2020-04-01 04:22:21 +03:00
::
:: Generators
++ tab-generators
|= [pfix=path app=(unit term) gens=(list term)]
^- (list (option tank))
%+ turn gens
|= gen=term
^- (option tank)
=/ pax=path
(weld pfix ~[gen %hoon])
=/ file
.^(@t %cx pax)
:_ (render-help file)
?~ app
(cat 3 '+' gen)
?: =(%hood u.app)
(cat 3 '|' gen)
:((cury cat 3) ':' u.app '|' gen)
:: Stolen from +help
++ render-help
|= a=@t
^- tank
:- %leaf
=/ c (to-wain:format a)
?~ c "~"
hoon: renames +new-end, +new-lsh, +new-rsh
2020-12-02 11:47:29 +03:00
?. =(':: ' (end [3 4] i.c))
dojo: add autocomplete for various dojo syntaxes Adds autocomplete for - :app poke syntax - :app|gen poke generator syntax - +gen naked generator syntax - =var variable syntax
2020-04-01 04:22:21 +03:00
"<undocumented>"
(trip i.c)
dojo: add tab completion This is initial support for type-aware tab completion. When you hit tab, it tries to complete the word you're in the middle of using a face or arm in the subject at that point in the code. It also shows all possible matches and their associated types. It's nearly instantaneous. Notes: - It advances to the longest common prefix, so if you hit tab on `ab` and the only possible results are `abcde` and `abcdz`, then it'll write `abcd` and print both out (with their types). - If there are fewer than ten matches, it prints the type along with the face. Printing types is too slow to use all the time, but with 10 it's essentially instantaneous. - The match closest in the subject to you (i.e. smallest axis number) is displayed lowest (closest to your focus). Examples below, where `<TAB>` represents me hitting tab while my cursor is at that position (the line with the `<TAB>` is not preserved in the actual output). ``` ~zod:dojo> eth<TAB> ----- ethereum #t/<11.qcl {<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> <21.yeb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}> ethereum-types #t/<3.ltb 27.ipf 7.ecf 36.uek 92.bjk 247.ows 51.mvt 126.xjf 41.mac 1.ane $141> ~zod:dojo> ethereum ~zod:dojo> |= zong=@ud z<TAB> ----- zing #t/<1.dqs {* <126.xjf 41.mac 1.ane $141>}> zap #t/<1.iot {tub/{p/{p/@ud q/@ud} q/""} <1.rff {daf/@t <247.ows 51.mvt 126.xjf 41.mac 1.ane $141>}>}> zuse #t/$309 zong #t/@ud ~zod:dojo> |= zong=@ud zo<TAB> ----- zong #t/@ud ~zod:dojo> |= zong=@ud zong ~zod:dojo> <TAB> hoon-version trel quip pole unit qual lone ... about 600 more lines ... unity html zuse eny now our ~zod:dojo> ``` Functionally, this is in a state where I'd be comfortable shipping it. It doesn't interfere with anything if you don't press tab, and it's perfectly OTA-able. I do think its output is a little verbose, but that can be tuned over time as people try it and determine what feels good in practice. Additional notes: - There are plenty of similar systems for other languages, but my most direct inspiration is Idris's editor tools. This is implemented for the dojo, but I actually want it in my editor, which is why the meat is all defind in a library. I've only tested on dojo one-liners, so I don't know the performance on large blocks of code. - The default type printer isn't great for this use case. In particular, - Cores should not print anything about their context - The `#t/` should go away - If it looks like a gate, we should print its return value - Maybe special handling for molds, but if the above is done, then for example `bone` is `* -> @ud`. - The worst part about our wing ordering is that it really screws up tab completion. You want to do `point.owner-address` instead of `owner-address.point` because that lets you type `point.ow<TAB>`. I weakly prefer reading it how we do it now, but it's really not great. You could do an (dojo-specific?) alternate syntax of `point;owner-address`; this is a simple transformation. - Regardless of the above, this should handle the case where we're in the middle of defining a wing; it doesn't right now. - When a variable is shadowed, we show both of them. We should probably show the shadowed one with a `^`. - We probably shouldn't print out hundreds of results. Maybe just the closest 50 with ellipses. - This gets you any face in your subject, regardless of whether its type is reasonable. We could limit that some by copying the `gol` logic in mint, so that if the pseudo-backward-inference engine happens to know what type it should be, you can filter the tab results according to if they nest in that type. This would be "strongly type-aware".
2019-10-31 06:39:02 +03:00
--
Reference in New Issue Copy Permalink