0fea20c821
Prepare for v1.41.2 and indicate vm changes for stack correction.
2026-02-18 08:27:10 -06:00
91cc499e77
1.42.2 patch.
2026-02-18 08:19:55 -06:00
68850a0a05
Update for 1.41.2 patch.
2026-02-18 08:19:13 -06:00
d3d7c675a8
Update CHANGELOG.md
2026-02-17 18:56:21 -06:00
b2c9fc123c
Generate JOP_PUT_INDEX in the compiler when possible.
2026-02-17 09:07:01 -06:00
fa0c039cd3
Add regression test for issue #1714
2026-02-17 07:59:08 -06:00
78ef9d1733
Initialize memory allocated by put ( #1715 )
2026-02-17 07:50:50 -06:00
b6676f350c
Use snprintf instead of sprintf - sequel ( #1713 )
...
Co-authored-by: sogaiu <983021772@users.noreply.github.com >
2026-02-16 09:12:02 -06:00
0299620a2d
Code defensively with regard to stack resizes from custom get and put
...
for abstract types.
Abstract types whose get/put/length/etc. implementations allocated fiber
slots could break the VM by invalidating the stack pointer in the
interpreter. This is admittedly a bit unusual but is something most
users would probably expect to work. Debugging this would be a real pain.
2026-02-16 09:02:07 -06:00
739d9d9fe3
Expose module/add-syspath and update CHANGELOG.md
2026-02-15 22:20:54 -06:00
1557f9da78
Don't reference argv after fiber may have been resized.
2026-02-15 21:54:42 -06:00
529d8c9e4a
Improve ability to load modules by full path.
...
Be explicit when we are including this functionality. Add a function
module/add-file-extension that can do this programatically.
2026-02-15 21:10:26 -06:00
2df16e5a48
Allow garbage collection to be called inside the module entry.
...
This usually shouldn't be needed, but in the case that it is, or if
garbage collection is triggered manually, we can prevent use-after-free.
2026-02-15 18:46:21 -06:00
b0db2b22d6
Remove macos-13
2026-02-15 11:18:40 -06:00
8b6d56edae
Patch release to 1.41.1
2026-02-15 10:36:46 -06:00
a2a7e9f01e
Add explicit include of inttypes.h
2026-02-15 10:24:56 -06:00
4b078e7a45
Use correct format specifier on windows if missing message.
2026-02-15 10:17:16 -06:00
5c0bb4b385
Cosmo libc builds were not working.
2026-02-15 09:57:10 -06:00
2aaa7dfa10
Sort keys when compiling struct and table literals.
...
Order of evaluation becomes more clear in some cases.
2026-02-15 09:42:55 -06:00
10bb17199c
Prepare for 1.41.0 release
2026-02-15 08:45:56 -06:00
0aa7dfeb9a
Work on windows for WSAConnect not working.
...
For remote connections, if you did not manually wait for the connection
to settle, the programmer would see unspecified network errors. This is
is because the underlying TCP Connection had not been established yet.
The correct way to deal with this is to use ConnectEx if available
instead of WSAConnect.
2026-02-14 21:11:28 -06:00
8f7c32e5cb
Update for msvc build.
2026-02-14 18:31:57 -06:00
abd7bb1110
Use snprintf instead of sprintf ( #1711 )
...
Co-authored-by: sogaiu <983021772@users.noreply.github.com >
2026-02-14 09:04:33 -06:00
d81512723b
When pretty printing, don't sort keys for huge tables.
...
This was exposed when printing `(invert (range 200000))`, which
isn't so large that we shouldn't be able to sort it, but was taking
far too long to compute.
2026-02-14 08:57:27 -06:00
2a54154976
Don't use preload on absolute paths.
...
When importing full paths, the old preload code was preventing
(import <fullpath> :fresh true) from working as expected.
2026-02-13 19:36:09 -06:00
306ce892ea
Merge branch 'make-modules-easier'
2026-02-06 00:31:08 -06:00
c7c3821aa6
Remove extra output from peg test.
2026-02-06 00:23:45 -06:00
d2685594f9
VERBOSE=1 caused tests to fail as we did more asserts inside the assert.
...
Thenc checked stderr for messages. Since the helper makes assert write
to stderr, this caused extra cruft in our test output.
2026-02-06 00:17:11 -06:00
ca5c617fba
More tweaks to peg suite for arm32 failures.
2026-02-06 00:04:18 -06:00
16b449a137
Print "actual" output on verbose mode for suite-peg
2026-02-05 23:52:28 -06:00
2e8dd90a51
Line endings in tests.
2026-02-05 23:38:55 -06:00
196f27af3d
Update CHANGELOG.md
2026-02-05 20:01:00 -06:00
42c0096ce7
Allow simpler loading of modules.
...
E.g.
(import project/file.janet :as file)
(import ./project/file.janet :as file)
(import /home/me/project/file.janet :as file)
Will now do what you expect. When loading with a file extension, this
will first check directly if the relative (or absolute) path exists.
Also expands `module/find` to be able to show all matches for a given
module name without needing to try and catch import errors. Lastly, one
can also override the module/find functionality by passing `:loader`
directly.
E.g.
(import ./macoslib.dylib :loader :native :as my-thing)
Which will allow easily bypassing all module/find logic;
"./macoslib.dylib" will be passed directly to the :native loader
`(get module/loaders :native)`.
The module system was previously left open to customization but with
defaults to encourage a more typical style. However, these qol
improvements are more than worth it and can in some cases let people
stop fighting the module system.
2026-02-05 19:45:30 -06:00
0194115412
Auto-format janet.
2026-02-05 00:33:52 -06:00
f33697d6a0
Show accumulation buffer and tagged capture in peg debug rule.
2026-02-04 21:17:08 -06:00
b2bf70eace
Move peg debug output to stderr in line with other internal debug tools.
...
Also allow disabling color in the debug output.
2026-02-04 18:52:20 -06:00
855d1f2940
Update changelog.
2026-02-04 18:46:09 -06:00
416bba9bd9
Add ??/debug PEG rule ( #1710 )
...
* Add debug (short-form ??) PEG rule
* Display capture stack to debug PEG rule
* Revise spec_debug to emit just the rule
* Update to proper indentation
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com >
* Add a marshal/unmarshal test for ??/debug
---------
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com >
2026-02-04 18:44:46 -06:00
517e40a17b
Speed up symbol resolution.
...
In my testing of `janet -l spork -e '(os/exit)'`, this speeds up
compilation by ~5%. Allocating and interning symbols from C is often
slow, so avoid it.
Add utility function for looking up keywords in tables more quickly.
More measurement on compiling spork for compilation speed, win
over original code closer to between 1-2% improvement
2026-02-02 21:19:48 -06:00
4f9a2af357
Add unmarshal sandbox.
2026-02-02 18:28:08 -06:00
a37752708e
Add :compile, :asm, and :threads to sandbox flags.
...
These can be used to limit programs in a very expansive way.
2026-02-02 18:13:46 -06:00
5042ad6d4b
More tweaks to changelog.
2026-02-01 14:46:20 -06:00
643c0b4976
Update CHANGELOG.md
2026-02-01 10:54:31 -06:00
ecb72c9c9a
Revert changes to peg grammar.
2026-02-01 08:08:50 -06:00
a95546ff16
Merge pull request #1703 from amano-kenji/peg
...
Make peg-grammar available in all threads for peg/compile
2026-02-01 07:42:28 -06:00
d47f82713b
More tweaks for named arity linting.
...
Handle more edge cases where &named is combined with &opt
2026-01-31 23:57:08 -06:00
497e363401
Make sure named argumnets are compiled before destructured params.
...
This handles an edge-case when combing destructured and named
parameters that produced spurious and confusing compiler lints.
2026-01-31 19:34:26 -06:00
8481da18d0
Add another bounds check for updating labels. Related to #1700
2026-01-31 13:20:45 -06:00
8f8382eead
Partially address issue #1654 - better linting for named arguments.
...
While there are still cases that we could improve here, this change will
help in most cases with minimal overhead.
2026-01-31 13:16:02 -06:00
8e2ec997f0
Update copyright.
2026-01-31 10:33:57 -06:00
Calvin Rose