Java completer can include FixIts which are applied when a completion
entry is selected. We use the existing mechanism implemented for c-sharp
to perform these edits using the CompleteDone autocommand.
However, the existing mechanism relies on pattern matching the source to
work out which item was completed. Vim patch 8.0.1493 introduces support
for user_data on completion items, so when available we populate it with
the completion array index of the item and use that to get the exact
element that was selected. This is both a lot faster and a lot more
accirate.
Of course when applying these 'FixIts' we don't interrupt the user with
confirmation or the quickfix list as this would just be annoying. If the
server reports that an edit must be made, we just make the edit. This is
achieved by adding a silent flag to ReplaceChunks.
When no argument is given to the YcmToggleLogs command, instead of displaying
the list of available logfiles, prompt the user to open (or close if already
open) one of them.
Do not call user's completion function if the start column is after the current
column or if there are no candidates. This avoids keeping the user in
completion mode even if there is no completion menu.
Use the timers feature to display diagnostics asynchronously instead of waiting
for an autocommand to trigger.
Increase Vim version requirement to 7.4.1578.
Drop the CursorHold and CursorHoldI autocommands.
Parse buffer on the TextChanged autocommand instead of CursorMoved.
[READY] Add client logfile
We have a lot of issue reports with Python exceptions that interrupt user workflow or even worse make the editor unusable and force users to restart it (e.g. issue https://github.com/Valloric/YouCompleteMe/issues/2192). We really want to avoid that but at the same time we can't just silence these exceptions because they are useful to debug the issue. Logging them to Vim `:messages` is not practical because we can't write to it without displaying messages in the status line which, in addition to distract users, may lead to various issues like the infamous `Press ENTER or type command to continue` message. This is why a logfile is needed.
For now, only server crashes are logged but more logging will be added: connection issues with the server (`ConnectTimeout`, `ReadTimeout`, etc. exceptions), UltiSnips unavailability, requests sent to the server, etc.
The behavior of the `:YcmToggleLogs` command is changed to accept multiple arguments where each argument is a logfile name. Each of these files is opened in a separate window or closed if already open. When no argument is given, the list of available logfiles is displayed to the user. Example:
```
Available logfiles are:
ycm_pz83u7.log
ycmd_23830_stderr_gf6j3i.log
ycmd_23830_stdout_gmpa_k.log
```
With this change and PR #2342, we will add the completers logfiles to the list of files that can be opened with the `:YcmToggleLogs` command.
A bunch of tests are added that cover almost all changes introduced by this PR.
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/valloric/youcompleteme/2430)
<!-- Reviewable:end -->
[READY] Bumps required Vim to 7.4.143 and adopts TextChangedI.
Long and personal experience, when TextChangedI gets used, YCM seems
to perform better, diagnostics will trigger much less frequently at
inappropriate occasions, even less with whitespace agnostic triggers,
if I recall correctly...
There's a previous discussion at #1337. At the time this change was bundled in
a single pull request with other changes.
It has gone through the test of the time, sole issue I've stumbled upon was reported
by @puremourning at https://github.com/oblitum/YouCompleteMe/issues/10. This
is what me and other users of my fork use daily.
<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/valloric/youcompleteme/1901)
<!-- Reviewable:end -->
