See the test for usage example.
wraphybridvalue() takes a key/value pair because a hybrid dict passes a key
to its makemap() function. Since makemap() of showmanifest() doesn't need
a key, it's set to None.
This is minimal and non-controversial implementation of extdata() template
function. Originally extdata sources were exposed to the keyword namespace,
but I've changed it to a plain function for simplicity.
The goal is to allow us to easily access to nested data. The dot operator
will be introduced later so we can write '{p1.files}' instead of
'{revset("p1()") % "{files}"}' for example.
In the example above, 'p1' needs to carry a mapping dict along with its
string representation. If it were a list or a dict, it could be wrapped
semi-transparently with the _hybrid class, but for non-list/dict types,
it would be difficult to proxy all necessary functions to underlying value
type because several core operations may conflict with the ones of the
underlying value:
- hash(value) should be different from hash(wrapped(value)), which means
dict[wrapped(value)] would be invalid
- 'value == wrapped(value)' would be false, breaks 'ifcontains'
- len(wrapped(value)) may be either len(value) or len(iter(wrapped(value)))
So the wrapper has no proxy functions and its scope designed to be minimal.
It's unwrapped at eval*() functions so we don't have to care for a wrapped
object unless it's really needed:
# most template functions just call evalfuncarg()
unwrapped_value = evalfuncarg(context, mapping, args[n])
# if wrapped value is needed, use evalrawexp()
maybe_wrapped_value = evalrawexp(context, mapping, args[n])
Another idea was to wrap every template variable with a tagging class, but
which seemed uneasy without a static type checker.
This patch updates {manifest} to a mappable as an example.
This makes 'foo|bar%baz' parsed as '(foo|bar)%baz', not 'foo|(bar%baz)'.
Perhaps it was a mistake that '%' preceded '|'. Both '|' and '%' can be
considered a kind of function application, and '|' is more like a '.' operator
seen in OO languages. So IMHO '|' should have the same (or higher) binding as
'%'.
The BC breakage should be minimal since both '|' and '%' operators have
strict requirements for their operands and 'foo|bar%baz' was invalid:
- right-hand side of '|' must be a symbol
- left-hand side of '%' must be a dict or list
- right-hand side of '%' must be a string or symbol
Before, iter(hybrid) was proxied to hybrid.gen, which generated formatted
strings. That's why we had to apply joinfmt() to the dicts generated by
hybrid.itermaps(). Since this weird API was fixed at 1906bcc0f923, we can
get rid of the makemap() calls from join().
It can be useful for extensions to be able to produce the shortest
unambiguous hash (including the in-tree "show" extension). That logic
is currently inside the shortest() template function. Let's move it
out of the templater. I've put it on revlog since it's closely related
to revlog._partialmatch. We may also want a convenience method on
context, but I'll leave that for a later patch.
Differential Revision: https://phab.mercurial-scm.org/D724
Extract, deduplicate users informations from obs markers in order to display
them.
Print all users for the moment, we might want to display users only in verbose
mode later.
Add a template function obsfateverb which use the markers information to
compute a better obsfate verb.
The current logic behind the obsfate verb is simple for the moment:
- If the successorsets is empty, the changeset has been pruned, for example:
Obsfate: pruned
- If the successorsets length is 1, the changeset has been rewritten without
divergence, for example:
Obsfate: rewritten as 2:337fec4d2edc, 3:f257fde29c7a
- If the successorsets length is more than 1, the changeset has diverged, for
example:
Obsfate: split as 2:337fec4d2edc, 3:f257fde29c7a
As the divergence might occurs on a subset of successors, we might see some
successors twice:
Obsfate: split as 9:0b997eb7ceee, 5:dd800401bd8c, 10:eceed8f98ffc; split
as 8:b18bc8331526, 5:dd800401bd8c, 10:eceed8f98ffc
The repo instance is currently only used to provide a changeset
lookup function as part of parsing revsets. I /think/ this allows
node fragments to resolve. I'm not sure why we wouldn't want this
to always "just work" if parsing a revset string.
Plus, an upcoming commit will introduce a new consumer that needs a
handle on the repo. So passing it more often will make that code
work more.
Passing a repo instance in all callers of revset.match* results in
a bunch of test changes. Notably, branch and tags caches get
populated as part of evaluating revsets. I'm not sure if this is
desirable. So this patch takes the conservative approach and only
passes the repo if we're passing a ui instance.
Differential Revision: https://phab.mercurial-scm.org/D97
We were using str because on Python 2, str were bytes but now we have to use
bytes. Otherwise the if conditions fails and we have weird results from commands
on Python 3.
This provides a simpler API for callers which don't need full templating
stack. Instead of storing the given template as the name specified by topic,
use '' as the default template to be rendered.
The idea is simple. If the given node id prefix is 'ff...f', add +1 to the
number of matches (e.g. ambiguous if partial + maybewdir > 1).
This patch also fixes id() revset and shortest() template since _partialmatch()
can raise WdirUnsupported exception.
See the previous commit for why.
splitlines() does not pass a mapping dict, which would probably mean the
legacy template didn't work from the beginning.
Like field init shorthand of Rust. This is convenient for building a JSON
object from selected keywords.
This means dict() won't support Python-like dict(iterable) syntax because
it's ambiguous. Perhaps it could be implemented as 'mapdict(xs % (k, v))'.
Before, it could spill an internal representation of compiled template such
as [(<function runsymbol at 0x....>, 'extras'), ...]. Show less cryptic
message if no symbol found.
New findsymbolicname() function will be also used by dict() constructor.
Unlike revset, function arguments are pre-processed in templater. That's why
we need to define argspec per function. An argspec field looks somewhat
redundant in @templatefunc definition as a name field contains human-readable
list of arguments. I'll make function doc be built from argspec later.
Ported separate() function as an example.
The goal is to fix "{hybrid_obj|json}" output.
A _hybrid object must act as a list or a dict as well as a generator of
legacy template strings. Before, _hybrid.__iter__() was assigned for legacy
template, which conflicted with list.__iter__() API.
This patch drops _hybrid.__iter__() and makes stringify/flatten functions
unwrap a generator instead.
My initial expectation was that the list would be empty, and therefore
detectable with {if()}. The map for {latesttag()} is populated with real values
in this case (except {tag}), so it probably doesn't make any sense to change
this to an empty list.
This was originally written for JSON templating where we would have to be
careful to not add extra comma, but seems generally useful.
Inner loop started by % operator has its own counter.
This slightly reduces the difference between changeset_templater and formatter,
and helps extending formatter to support changeset templating.
New formatnode() is not a template filter, but a function since a filter
cannot access to ui. And it's marked as DEPRECATED since I think it exists
only for compatibility reasons.
New revsetlang module hosts parser, tokenizer, and miscellaneous functions
working on parsed tree. It does not include functions for evaluation such as
getset() and match().
2288 mercurial/revset.py
684 mercurial/revsetlang.py
2972 total
get*() functions are aliased since they are common in revset.py.