The ultimate goal is to reduce the method calls necessary for `Vector.map`.
# Important Notes
- I managed to reduce the number of Java stack frames needed for each `Vector.map` call from **150** to **22** (See https://github.com/enso-org/enso/pull/11363#issuecomment-2432996902)
- Introduced `Stack_Size_Spec` regression test that will ensure that Java stack frames needed for `Vector.map` method call does not exceed **40**.
- Linting updates.
- Add an `Examples.welcome` and adjust the start up project to use it.
- Merge all of Cass's work into the source code.
- Make example render in mono space font.
Enables `engine.TruffleCompilation` in `std-benchmarks`, collects the logs and dumps compilation into to `System.err` when a benchmark is influenced by dynamic compilation.
Fix#10503 by creating a benchmark and then speeding it up by making sure usage of `InteropLibrary` reminds in partially evaluated code and isn't hidden behind `@TruffleBoundary`.
- Removes `First` and `Last` from the `Standard.Base` exports.
- Enable auto-scoping for all `Index_Sub_Range` and `Text_Sub_Range`.
- Update all use of those methods to use auto-scoping.
- Fix `Excel_Workbook.sheet` and add a test.
- Add icon for `Table.row_count` and `DB_Table.row_count`.
- Make `join_kind` widget `Display.Always`.
- Add expression as an option to `Aggregate_Column`.
- Add `Simple_Calculation.Copy` to create a copy.
- Add defaults to `Simple_Expression` so less errory.
- Set period to default to day for `date_diff` allowing use in expressions.
- Add `Text_Left`, `Text_Right`, `Text_Length` and `Format` to `Simple_Expression`.
Goal of this PR is to refactor the design of OrderMask and avoid copying arrays or lists wherever possible.
We have removed a few legacy functions which were not being used.
On a poor mans benchmark seems to be quicker (13s vs 16s) and memory usage should be lower.
Adds new benchmark joining a large table to a small table in preparation for a coming optimisation that will only index the smaller of the 2 tables in #8342
Implements `Warnings.get_all wrap_errors=True` which wraps warnings attached to values inside vectors with `Map_Error`, which includes the position of the value within the vector. See [the documentation](https://github.com/enso-org/enso/blob/develop/docs/semantics/wrapped-errors.md) for more details.
`get_all wrap_errors=True` does not change the warnings that are attached to values -- it wraps them before returning them to the caller, but does not change the original warnings attached to the values.
Wrapped warnings only appear attached to the vector itself. The values inside the vector do not have their warnings wrapped.
Warning propagation is not changed at all; `Warnings.get_all` (with default `wrap_errors=False`) behaves as before. `get_all wrap_errors=True` is meant to be used primarily by the IDE, although it can be used anywhere this wrapping is desired.
* tests
* wip
* wip
* additional warnings
* wip
* wip
* cleanup
* nested wrapping
* multiple nestings
* wraps_error uses looks_for, test for should_fail_with
* wip
* stack trace line fix
* use catch_primitive internally
* fix warning mapping, dtf spec
* just one wrapper checker, vector spec
* missing ctor, back to non-primitive catch
* back to c_p
* put old map back
* wip
* unnest tests
* Array.map on_problems
* wip
* Revert "wip"
This reverts commit c30d171457.
* better test names
* warning logging
* wip
* wip
* move logic into ALH
* doc
* constant
* My_Error.Error
* nested
* doc
* map_primtiive in warning mapper
* composition
* ref spec
* Remove warnings prior to matching on the value
If an expression has warnings and is matched we:
1) extract the warnings
2) execute the branch of a pattern that matches the value
3) attach extracted warnings to the result
This caused warnings to reappear when doing the custom warnings
manipulation.
This is also consistent with how `CaseNode`'s `doWarning` specialization
is defined.
* fix 1
* do not auto unwrap in test error checkers
* nested error matcher
* in problems too
* dtf
* v
* statistics
* wip
* Table_Spec, map_with_index_primitive
* Column_Operations_Spec
* disable warning wrapping and Report_Warning
* unimpl test
* Warnings_Spec
* DCS
* ACG JP
* zip_primitive
* join_helpers
* Lookup_Helpers
* Table
* Data_Formatter
* Value_Type_Helpers
* revert check types changes
* table_helpers
* table tests
* remove st
* do not remove warnings from value
* vec docs, tests for zip, mwi, flat_map
* docs, fixes
* remove nested_error_matcher
* cleanup
* benchmark
* one error
* alter
* add bench to main
* review
* review
* review
* tail call
* changelog
* tail call was not a tail call
* ws
* bad import
* Added missing import
* Update distribution/lib/Standard/Base/0.0.0-dev/src/Data/Array.enso
Co-authored-by: Radosław Waśko <radoslaw.wasko@enso.org>
* review, ref example
* lazy benchmark data
* extra paren
* check outside of catch
* review
* vector too
* actually lazy
* disambiguate Map_Error
* finish rename
* move to extensions
* combine Additional_Warnings error
* rename to map_no_wrap
* do not catch and rethrow
* review
* wip
* remove _primitives entirely
* remove unused should_fail_with function options
* remove expected_warning as function in Problems
---------
Co-authored-by: Hubert Plociniczak <hubert.plociniczak@gmail.com>
Co-authored-by: Radosław Waśko <radoslaw.wasko@enso.org>
With herein proposed change one can pass an optional filter to `enso --run test/Benchmarks` to execute only groups and specs that contain given string in its name.
While trying to speed `MetadataStorage` up - #8324 - I felt the need to have an independent (on my computer) measurement of startup time. Here is a benchmark that measures how long execution of two simple hello world programs take.
# Important Notes
There are two benchmarks:
- `empty_startup` measures the time needed to boot without using any `Standard` library - basically _an overhead of the JVM and engine_
- `hello_world_startup` measures the use of `IO.println` - it shall take longer than `empty_startup` and show the overhead we have while processing the standard library
Upgrade to GraalVM JDK 21.
```
> java -version
openjdk version "21" 2023-09-19
OpenJDK Runtime Environment GraalVM CE 21+35.1 (build 21+35-jvmci-23.1-b15)
OpenJDK 64-Bit Server VM GraalVM CE 21+35.1 (build 21+35-jvmci-23.1-b15, mixed mode, sharing)
```
With SDKMan, download with `sdk install java 21-graalce`.
# Important Notes
- After this PR, one can theoretically run enso with any JRE with version at least 21.
- Removed `sbt bootstrap` hack and all the other build time related hacks related to the handling of GraalVM distribution.
- `project-manager` remains backward compatible - it can open older engines with runtimes. New engines now do no longer require a separate runtime to be downloaded.
- sbt does not support compilation of `module-info.java` files in mixed projects - https://github.com/sbt/sbt/issues/3368
- Which means that we can have `module-info.java` files only for Java-only projects.
- Anyway, we need just a single `module-info.class` in the resulting `runtime.jar` fat jar.
- `runtime.jar` is assembled in `runtime-with-instruments` with a custom merge strategy (`sbt-assembly` plugin). Caching is disabled for custom merge strategies, which means that re-assembly of `runtime.jar` will be more frequent.
- Engine distribution contains multiple JAR archives (modules) in `component` directory, along with `runner/runner.jar` that is hidden inside a nested directory.
- The new entry point to the engine runner is [EngineRunnerBootLoader](https://github.com/enso-org/enso/pull/7991/files#diff-9ab172d0566c18456472aeb95c4345f47e2db3965e77e29c11694d3a9333a2aa) that contains a custom ClassLoader - to make sure that everything that does not have to be loaded from a module is loaded from `runner.jar`, which is not a module.
- The new command line for launching the engine runner is in [distribution/bin/enso](https://github.com/enso-org/enso/pull/7991/files#diff-0b66983403b2c329febc7381cd23d45871d4d555ce98dd040d4d1e879c8f3725)
- [Newest version of Frgaal](https://repo1.maven.org/maven2/org/frgaal/compiler/20.0.1/) (20.0.1) does not recognize `--source 21` option, only `--source 20`.
- Closes#5303
- Refactors `JoinStrategy` allowing us to 'stack' join strategies on top of each other (to some extent) - currently a `HashJoin` can be followed by another join strategy (currently `SortJoin`)
- Adds benchmarks for join
- Due to limitations of the sorting approach this will still not be as fast as possible for cases where there is more than 1 `Between` condition in a single query - trying to demonstrate that in benchmarks.
- We can replace sorting by d-dimensional [RangeTrees](https://en.wikipedia.org/wiki/Range_tree) to get `O((n + m) log^d n + k)` performance (where `n` and `m` are sizes of joined tables, `d` is the amount of `Between` conditions used in the query and `k` is the result set size).
- Follow up ticket for consideration later:
#8216
- Closes#8215
- After all, it turned out that `TreeSet` was problematic (because of not enough flexibility with duplicate key handling), so the simplest solution was to immediately implement this sub-task.
- Closes#8204
- Unrelated, but I ran into this here: adds type checks to other arguments of `set`.
- Before, putting in a Column as `new_name` (i.e. mistakenly messing up the order of arguments), lead to a hard to understand `Method `if_then_else` of type Column could not be found.`, instead now it would file with type error 'expected Text got Column`.
After a discussion, I was really curious that our panics are supposed to be almost free - and while trusting that statement, it was really hard to believe - so I wanted to see for myself - knowing that an experiment is the most robust source of this kind of information - testing that in practice.
So I wrote a benchmark comparing various ways of reporting errors, also testing them both at 'shallow' and 'deep' stack traces (adding 200 additional frames) - to see how stack depth affects them, if at all.
The panics are indeed blazing fast! Kudos to the engine team. However, it seems that our dataflow errors are relatively slow (and we tend to use them _more_ than panics and want to be using them more and more). This uncovers a possible optimization opportunity. Can we make them as fast as panics??
Analysis of the benchmark results in comment below.
- Follow-up of #8055
- Adds a benchmark comparing performance of Enso Map and Java HashMap in two scenarios - _only incremental_ updates (like `Vector.distinct`) and _replacing_ updates (like keeping a counter for each key). These benchmarks can be used as a metric for #8090
- Fixes#7352 by remembering original value types in type inference mode to be able to reconstruct them for Mixed.
- Added more benchmarks for comparing performance of constructing columns.
- Fixes missing implementations that caused `Table.union` crashing on some type pairs.
- Ensures that `Loss_Of_Integer_Precision` warning is not swallowed when numeric columns are unioned to create a `Float` column.
- Adds test for all of the above cases.
- Allow to output benchmark results to a CSV by setting an environment variable - useful for quickly comparing benchmarks, e.g. in Enso.
* Transform Range.iterate test to a benchmark
a) it slows down regular unit testing
b) it is not a unit test
c) it behaves like a benchmark
So it should be a benchmark.
* missing Range import
