vim/YouCompleteMe
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
1,025 Commits 1 Branch 0 Tags 38 MiB
7f5dd0fcf0
Commit Graph

417 Commits

Author SHA1 Message Date
Strahinja Val Markovic
a5fb6b7509 Removing unsupported signal() calls on Windows 
SIGQUIT and SIGHUP are not supported.

For details, see Python docs: http://docs.python.org/2/library/signal.html#signal.signal
 2013-10-17 20:16:44 -07:00
Strahinja Val Markovic
6e9a16e90e Signal handler must take 2 params 
Otherwise we get a TypeError
 2013-10-17 20:14:56 -07:00
Strahinja Val Markovic
acae8e4e9d Tweaked the request retries logic. 
We now send more retry requests that are less spaced apart.
 2013-10-17 14:45:53 -07:00
Strahinja Val Markovic
40464f6e0d Resolving issues with event requests timing out 
It appears that the issue comes from sending a None timeout to Requests. It
seems it's a bug in Requests/urllib3. So we just pick an arbitrary long timeout
of 30s as the default.
 2013-10-17 14:21:37 -07:00
Strahinja Val Markovic
5070835d01 Ensuring server cleanup on OS signals 
atexit won't run registered functions for SIGTERM which we send to the server.
This prevents clean shutdown.

Also making sure that the server logfiles are deleted as well.
 2013-10-17 12:21:59 -07:00
Strahinja Val Markovic
b4d5b4ffbb More logging for extra conf store 2013-10-17 12:21:33 -07:00
Strahinja Val Markovic
5b37a2e36d Don't print traceback when no detailed diagnostic 
We should only print the message.
 2013-10-16 14:37:42 -07:00
Strahinja Val Markovic
18002e17df Better handling of crashed ycmd; restart command 
Previously the YCM Vim client would go bonkers when ycmd crashed. Now the user
can continue using Vim just without YCM functionality.

Also added a :YcmRestartServer command to let the user restart ycmd if it
crashed. With a little luck, this will be rarely necessary.
 2013-10-15 15:27:54 -07:00
Strahinja Val Markovic
436017bd4d Now using new ycm_client_support shared lib 
This means we can now load just ycm_client_support (which is a much smaller
library) into Vim and ycm_core into ycmd. Since ycm_client_support never depends
on libclang.so, we never have to load that into Vim which makes things much,
much easier.
 2013-10-15 14:15:04 -07:00
Strahinja Val Markovic
016434e846 Deleting some dead code 2013-10-14 20:46:42 -07:00
Strahinja Val Markovic
98f549aeae More robust way of picking an unused local port 
Fixes #584.
 2013-10-14 20:38:45 -07:00
Strahinja Val Markovic
b903867cdd ycm_core now imported after extra conf preload 2013-10-14 15:29:00 -07:00
Strahinja Val Markovic
bc607724f0 Ensuring ident completion works always 
A bug turned it off when omni completion was available.

Fixes #583.
 2013-10-14 13:29:28 -07:00
Strahinja Val Markovic
a534a58477 Checking if ultisnips data present before using it 
Mentioned in issue #583, but it's not the root cause.
 2013-10-14 12:32:18 -07:00
Strahinja Val Markovic
c3fcaf2b29 Strict 0.5s time budget for completion requests 
Vim still loves to block the main GUI thread on occasion when asking for
completions... to counteract this stupidity, we enforce a hard budget of 0.5s
for all completion requests. If the server doesn't respond by then (it should,
unless something really bad happened), we give up.
 2013-10-14 11:08:15 -07:00
Strahinja Val Markovic
a04ae37ead Client & server threads increased from 4 to 10 2013-10-11 20:12:04 -07:00
Strahinja Val Markovic
f6432e1498 Releasing Python's GIL in C++ code where possible 
Without this, all requests to the server become effectively serialized.
 2013-10-11 19:27:04 -07:00
Strahinja Val Markovic
3ae10395ea Preventing Vim thread block on file save 
Syntastic would run SyntasticCheck on file save, which would unconditionally
call _latest_file_parse_request.Response() and thus block until the request
returned from the server. We don't want that, so we throw in an explicit check
for the request being ready.
 2013-10-11 19:09:21 -07:00
Strahinja Val Markovic
db45e243dd Getting path to python exe on Windows correctly 
Fixes #581.
 2013-10-11 11:11:02 -07:00
Strahinja Val Markovic
78e3607b00 We now only run extra conf preload for global file 
This changes functionality, but since this is an undocumented, non-public API,
it's fine. The reason this is required is because of issue #579; if we try to
run extra conf preload on non-global extra conf, we might not have the
permission to load it. The global extra conf is something the user explicitly
has to set so it's always fine to load that.
 2013-10-10 12:55:49 -07:00
Strahinja Val Markovic
7a73eb14d8 Fix problems with unknown extra conf at ycmd start 
Fixes #579.
 2013-10-10 11:32:20 -07:00
Strahinja Val Markovic
8ce07f508c <c-space> again filters the semantic completions 
Fixes #576.
 2013-10-09 20:20:34 -07:00
Strahinja Val Markovic
98ef568359 Refactored the popen call for ycmd 
Also removed shell = true in the other branch that calls popen
 2013-10-09 19:28:27 -07:00
Strahinja Val Markovic
9482ad189e Using full path to Python in popen call 
This is the root of the problem in issue #577.

Fixes #577.
 2013-10-09 19:18:36 -07:00
Strahinja Val Markovic
a25ed01a7c Starting ycmd without shell = True 
This should make it easier to shut down the server on some machines.

Fixes #577.
 2013-10-09 17:00:28 -07:00
Strahinja Val Markovic
70a51be209 Making the cs_completer test less flaky 2013-10-09 13:17:53 -07:00
Strahinja Val Markovic
3b057cc667 Basic integration test for the cs_completer 2013-10-08 21:30:53 -07:00
Strahinja Val Markovic
d84f2b0e8e cs_completer works again 2013-10-08 20:52:04 -07:00
Strahinja Val Markovic
7afd76b4d4 Better url joining in base_request 2013-10-08 20:49:00 -07:00
Strahinja Val Markovic
daa0d50655 Removed the logging code from Clang completer 
The exceptions are already logged by Bottle so what's the point.
Also removed an outdated TODO.
 2013-10-08 20:20:15 -07:00
Strahinja Val Markovic
3dbd407f7a Adding forgotten test file 2013-10-08 19:02:20 -07:00
Strahinja Val Markovic
3d55748400 Correctly handling ycm_extra_conf files now 
The user is asked about loading unknown extra conf files, as they were before.
 2013-10-08 16:21:43 -07:00
Strahinja Val Markovic
f0650ddc7f Minor style fixes 2013-10-07 16:53:41 -07:00
Strahinja Val Markovic
ebb1627f2e Resolving a minor TODO 
/filetype_completion_available handler renamed to /semantic_completion_available
 2013-10-07 16:28:40 -07:00
Strahinja Val Markovic
5000d2e4ae NativeFiletypeCompletionAvailable now a local call 
It used to block on the server to get the data. Now it doesn't anymore. This
speeds up Vim startup.
 2013-10-07 16:10:48 -07:00
Strahinja Val Markovic
c7be1f1b47 Omni completion works again 2013-10-07 15:47:48 -07:00
Strahinja Val Markovic
4b73739b09 Removing some obsolete todos 2013-10-07 14:21:46 -07:00
Strahinja Val Markovic
1db0e720bc Refactored the server tests for simplicity 2013-10-07 13:59:01 -07:00
Strahinja Val Markovic
ff7fa74fc9 <c-space> works again (forces semantic completion) 2013-10-07 13:09:34 -07:00
Strahinja Val Markovic
99b0f018b3 Adding a minor TODO 2013-10-07 11:03:30 -07:00
Strahinja Val Markovic
a836f68142 Fileype completer calls InCFamilyFile correctly 2013-10-07 11:03:30 -07:00
Strahinja Val Markovic
a9d7105e1b YCM now working on new buffers with ft set 
We used to demand a name be set for the buffer.

Fixes #568.
 2013-10-07 11:03:30 -07:00
Strahinja Val Markovic
915d092364 Prefixing clang version with "Clang version" 
It's easier to know what that magic string is actually referring to.
 2013-10-07 11:03:30 -07:00
Strahinja Val Markovic
1bba4a38cc Ensuring the temp dir exists before returning it 2013-10-07 11:03:30 -07:00
Strahinja Val Markovic
f0c9878f87 Ensuring we are serializing real Python objects 
Trying to json.dump Vim dictionaries fails.
 2013-10-07 11:03:30 -07:00
Strahinja Val Markovic
cb98dc8537 Fixing python 2.6 compatibility 
string.format() requires the number inside '{}' for Python 2.6.
 2013-10-07 11:03:30 -07:00
Strahinja Val Markovic
070d39b2a9 Adding a run_tests script 
This is now also used by Travis CI.
 2013-10-07 11:03:30 -07:00
Strahinja Val Markovic
9747bbc26f Removing the sys.path changes from jedi_completer 
This not needed anymore, the correct path to jedi is added to sys.path in
ycmd.py
 2013-10-07 11:03:30 -07:00
Strahinja Val Markovic
da723b1425 Moving Jedi under third_party 2013-10-07 11:03:30 -07:00
Strahinja Val Markovic
5902a5e521 frozendict now as submodule in third_party 2013-10-07 11:03:30 -07:00
Strahinja Val Markovic
88a260d448 ycmd now uses Waitress instead of CherryPy 2013-10-07 11:03:29 -07:00
Strahinja Val Markovic
8bc888d711 Vim now loads most defaults from the json file 2013-10-07 11:03:29 -07:00
Strahinja Val Markovic
afa1afc49b Handling no diagnostic data from event response 2013-10-07 11:03:29 -07:00
Strahinja Val Markovic
c43327d176 YcmShowDetailedDiagnostic works again 2013-10-07 11:03:29 -07:00
Strahinja Val Markovic
159a8ecdfa YcmDiags and ForceCompilation work again 2013-10-07 11:03:29 -07:00
Strahinja Val Markovic
62f813367d settings.json has integers for 1/0, not strings 2013-10-07 11:03:29 -07:00
Strahinja Val Markovic
54255318f1 Changing line endings for settings.json to unix 2013-10-07 11:03:29 -07:00
Strahinja Val Markovic
c274b7d4f8 Fixing the diagnostic-related race conditions 
Now, every FileReadyToParse event returns diagnostics, if any. This is instead
of the previous system where the diagnostics were being fetched in a different
request (this caused race conditions).
 2013-10-07 11:03:29 -07:00
Strahinja Val Markovic
b9bb788a2a Diagnostics work again... somewhat. 
There appear to be timing issues for the diag requests. Somehow, we're sending
out-of-date diagnostics and then not updating the UI when things change.

That needs to be fixed.
 2013-10-07 11:03:29 -07:00
Strahinja Val Markovic
e44fd87582 Added 'gitcommit' to filetype-completion blacklist 2013-10-07 11:03:29 -07:00
Strahinja Val Markovic
5ef945fa0a Adding 'qf' (quickfix) filetype to ignore list 2013-10-07 11:03:29 -07:00
Strahinja Val Markovic
7248979bb4 We now handle the starting FileReadyToParse event 
The problem was that when you start vim like "vim foo.cc", the FileReadyToParse
event is sent to the server before it's actually started up. Basically, a race
condition.

We _really_ don't want to miss that event. For C++ files, it tells the server to
start compiling the file.

So now PostDataToHandlerAsync in BaseRequest will retry the request 3 times
(with exponential backoff) before failing, thus giving the server time to boot.
 2013-10-07 11:03:29 -07:00
Strahinja Val Markovic
9742302cbd 'get_completions' handler is now 'completions' 2013-10-07 11:03:28 -07:00
Strahinja Val Markovic
9d0a6c96d7 Event and completion request are now async 
This results in a much snappier Vim.
 2013-10-07 11:03:28 -07:00
Strahinja Val Markovic
fe94ed6b1c Removing an outdate TODO 2013-10-07 11:03:28 -07:00
Strahinja Val Markovic
ec920b2758 Adding WebTest to pip test dependencies 2013-10-07 11:03:28 -07:00
Strahinja Val Markovic
6c53bad58f No more threading in completers! 
The server is multi-threaded and will spawn a new thread for each new request.
Thus, the completers need not manage their own threads or even provide async
APIs; we _want_ them to block because now were implementing the request-response
networking API.

The client gets the async API through the network (i.e., it can do something
else while the request is pending).
 2013-10-07 11:03:28 -07:00
Strahinja Val Markovic
e38d145a47 Added debug info about the server 
Location where running + logfiles location
 2013-10-07 11:03:28 -07:00
Strahinja Val Markovic
3ca758a581 Getting debug info works again 2013-10-07 11:03:28 -07:00
Strahinja Val Markovic
c9b4e9ed3a Minor code style changes 2013-10-07 11:03:28 -07:00
Strahinja Val Markovic
1027de73fb Using a constant for server error instead of 500 2013-10-07 11:03:28 -07:00
Strahinja Val Markovic
446d02f66e Subcommand name completion works again 2013-10-07 11:03:28 -07:00
Strahinja Val Markovic
718e9974b7 Resolving symlinks in GoTo target filepath 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
a6a364de41 Fix bug in cache clear clang completer subcommand 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
4a95c2fc7c GoTo commands for clang completer work again 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
fe0c0a1607 GoTo commands work again 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
13b86563f0 Minor code formatting changes 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
5b4b3ed281 LOGGER now set to None by default 
Init happens in test setup func or Main
 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
4c6a69b432 The UltiSnips completer works again 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
bf5708c422 NativeFiletypeCompletionAvailable works again 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
bdd769d043 default_settings.json now actually has defaults 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
4ac5466d80 Renaming server.py to ycmd.py 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
dd2445db06 Passing user options as file to server 
This is a much better idea than starting the server in a default state, and then
resetting the state with a POST request.
 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
10469d318d Passing user options to server now 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
387621d957 Moving SendCommandRequest into appropriate module 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
c01bc0481a Refactoring the server tests to use a helper func 
This makes the tests smaller, less repetitive and easier to maintain.
 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
3b9b9ed036 Ident completer loads syntax keywords again 2013-10-07 11:03:27 -07:00
Strahinja Val Markovic
c527cda436 Only calling Shutdown on valid Completers 2013-10-07 11:03:26 -07:00
Strahinja Val Markovic
c29dc44b38 Whitespace fix 2013-10-07 11:03:26 -07:00
Strahinja Val Markovic
9a4707f2c6 Ignoring failed event notifications 
These happen rarely and are not a big deal when they do. We still log them to
the Vim message area, but we don't annoy the user with the default, in-your-face
Python traceback.
 2013-10-07 11:03:26 -07:00
Strahinja Val Markovic
17716ff51f Whitespace fix 2013-10-07 11:03:26 -07:00
Strahinja Val Markovic
088eb4d0d2 Cleaner server shutdown 
Listening for VimLeave was sub-optimal. popen.terminate() is much cleaner.
 2013-10-07 11:03:26 -07:00
Strahinja Val Markovic
f51a687297 Server now shuts down cleanly on VimClose 2013-10-07 11:03:26 -07:00
Strahinja Val Markovic
1730660555 A (barely) working version of ycmd + client 
Still a lot of work to do.
 2013-10-07 11:03:26 -07:00
Strahinja Val Markovic
02b88dccf1 extra conf store now vim-free 2013-10-07 11:03:26 -07:00
Strahinja Val Markovic
fba4477ca3 Deleting a stray 'print' statement 
This was put in for debugging at some point.
 2013-10-07 11:03:26 -07:00
Strahinja Val Markovic
1d29a9a3bd Updating Jedi to v0.7.0 2013-10-07 11:03:26 -07:00
Strahinja Val Markovic
f5ad981f67 Converting Jedi unicode data to strings 2013-10-07 11:03:26 -07:00
Strahinja Val Markovic
9698bf8789 Bump ycm_core.so API version 2013-10-07 11:03:26 -07:00
Strahinja Val Markovic
08a9ff59b6 Using os.getcwd instead of getcwd from vimscript 2013-10-07 11:03:26 -07:00