Commit graph

433 commits

Author SHA1 Message Date
0095511cc7
fix: &mut object references
it just works?? maybe it did all the time...
2024-09-24 03:35:42 +02:00
678f8bd9ea
chore: updated jni-toolbox usage 2024-09-23 18:12:13 +02:00
1d0d87d521
feat(java): converted remaining glue to toolbox 2024-09-23 00:26:11 +02:00
7056dc341c
feat: add jni-toolbox, convert client 2024-09-22 02:22:51 +02:00
5a8b4ae800
chore: bump versions, fix docs 2024-09-18 16:11:12 +02:00
7c92b94a07
fix(java): fixed broken config and textchange conversions 2024-09-18 01:54:20 +02:00
c5ab72280a
fix(lua): dont throw error, return a string
while less "correct", dealing with the error in lua is really awful.
also its too late to deal with that error anyway, it was in another
thread
2024-09-18 00:30:46 +02:00
d215b4ab0b
chore(lua): renamed some top level methods 2024-09-17 23:33:37 +02:00
3047d21870
feat(lua): also pass errors in callbacks 2024-09-17 23:27:27 +02:00
27b56cbd03
docs(java): written javadocs, removed unused class 2024-09-17 23:17:09 +02:00
54db029ecc
chore(lua): split down the monofile 2024-09-17 23:00:30 +02:00
864348bef4
fix: dont use color codes in logger 2024-09-17 22:17:47 +02:00
7d53f61f0a
feat(lua): pass logs via callback channel, fix hints 2024-09-17 18:53:52 +02:00
6e63468e48
feat(java): general cleanup and overhaul of glue code 2024-09-17 17:37:22 +02:00
42ae4247ce
feat(lua): reintroduce and_then safely 2024-09-17 14:34:25 +02:00
d87b0923d9
feat(java): poll, stop, activeBuffers, general fixes and improvements 2024-09-17 02:40:18 +02:00
13ef05f8a2
fix(lua): event type field 2024-09-17 02:02:38 +02:00
7df85fb13e
chore: rename Workspace::try_new -> connect
more appropriate: we pass a ton of parameters and a new connection is
established. try_new was too generic
2024-09-16 21:11:42 +02:00
7aff141e78
fix(lua): cursor fields must be tuples too 2024-09-16 19:58:52 +02:00
cd1fc21df7
feat(lua): alternative getters, annotations
also cleaned up features a bit and updated annotations
oh and run CI too once
2024-09-16 18:24:24 +02:00
5cf6c4d4eb
fix(java): errors in send buffer glue, safer error handling 2024-09-16 02:59:43 +02:00
795423de2a
feat(java): callback API
Co-authored-by: alemi <me@alemi.dev>
2024-09-16 00:20:03 +02:00
05fbbe6f9d
feat(java): get_user, active_workspaces 2024-09-15 02:00:04 +02:00
963f2b698c
feat(java): callback API, buffer send
Co-authored-by: alemi <me@alemi.dev>
2024-09-15 01:56:51 +02:00
a59d2c4648
fix: feature name, disable workflow 2024-09-14 18:54:32 +02:00
cschen
c1ed0d45f3 feat(python): add config support to python with a getter/setter approach 2024-09-14 00:17:46 +02:00
5701a0c49e
chore: merge branch 'config' 2024-09-13 22:57:53 +02:00
97c57a81d6
docs: moved around documentation
on building and using

Co-authored-by: zaaarf <me@zaaarf.foo>
2024-09-13 22:56:00 +02:00
ac94fb86fd
chore: rename feature to py for consistency 2024-09-13 22:21:53 +02:00
92d181246e
feat: strict filetree() 2024-09-13 20:02:42 +02:00
c0b8906043
chore: move ffi infos in root mod
because i actually dont want docs.rs to build with ffi flags anymore:
docs get "littered" with napi auto methods and pyo3 stuff, which is
annoying and needless noise when reading the doc. I would much rather
only have relevant stuff, to not make doc consumers hop around guessing
what is relevant and what not
2024-09-13 12:04:27 +02:00
2f68006d07
fix: glue for js and py 2024-09-11 18:07:55 +02:00
422566a145
fix: proto code was mixing different ideas 2024-09-11 17:50:40 +02:00
314b163b05
fix: more verbose error strings 2024-09-11 17:50:05 +02:00
981e7218b5
fix(lua): actually still need userdata
what's up with weird userdata NULL ?? lets actually get serialized but
return userdata so we get best of all worlds. in the future we could
make overloads with multiple args which bypass serialization
2024-09-11 17:49:09 +02:00
6559fcd5f3
feat(java): use config struct for connect 2024-09-11 15:51:07 +02:00
123f11854c
fix(lua): normalize workspace method names
no _buffer
2024-09-11 15:50:05 +02:00
5e95ddde2a
feat(lua): add config, serialize/deserialize args
before we held all structs on the rust side, which probably means faster
passing and smaller objects but also very inconvenient because must be
constructed on the rust side. all api structs are now serialized back
and forth, which should become fast enough if jitted, allows for simpler
and smaller code and all around probably is a negligible overhead since
we were creating lua function tables and functions every time anyway.
could actually be faster to just serialize simple primitives rather than
creating a function to get it? idk lua api changed too btw
2024-09-11 15:46:05 +02:00
e91a504588
feat: add serialize feature for api structs 2024-09-11 15:45:35 +02:00
d57fb2c4b6
feat: added codemp config, changed connect methods 2024-09-11 15:12:31 +02:00
ba566439d3
docs: add favicon 2024-09-11 15:04:34 +02:00
65721f3e6a
feat(lua): allow loading with many more names
codemp, libcodemp, codemp_native, codemp_lua all work now
2024-09-10 12:29:11 +02:00
e05e273ee2
fix(lua): use patched crate by us, add features
basically upstream lua pushed a change making `send` and `module`
features incompatible, which is not ok for us as we need the LuaFunction
to be sync to send it on the channel. Also added features for all lua
versions, it's a bit messy on the Cargo.toml but now we support really
all lua versions and can probably publish on luarocks
2024-09-10 12:12:59 +02:00
f6ee2f0806
fix(java): borked glue methods 2024-09-09 16:06:05 +02:00
cschen
f2a2892674 fix(python): fix macros also for client.rs 2024-09-08 11:25:33 +02:00
cschen
f116f27b39 fix(python): remove the macro_export and use pub(crate) instead so that it doesn't leak into the documentations. 2024-09-08 11:03:10 +02:00
b968945a37
chore: codemp.dev -> code.mp 2024-09-07 22:10:47 +02:00
cschen
c1c9902b45 feat(python): switched to abi3 mode to allow multiple python versions with single wheel.
abi3 does not support `PyFunction` so added workaround checking for `is_callable()`
2024-09-06 00:24:37 +02:00
fb5bb51e65
fix: uuid was copy, string is not 2024-09-06 00:12:31 +02:00
917a6b96c2
feat: automatically map cursor uuids to names
not a fan of passing an Arc<DashMap> down to the cursor worker, but it
needs to access a mapping managed by the workspace so not sure if it can
be better. Into and From protocol types and Cursor are gone: do things
manually (since user is now a different thing, it can't be auto). Also
api::Cursor got changed: user field is Option<String> now
2024-09-05 23:59:05 +02:00