Improve struct and pointer autocompletion in C #4231
Conversation
|
I tested this with the sample code in the issue and auto-completion works fine with the chages to fix "1.". The changes done for "2." break auto-completion and from my testing are not required. Note that I use asynccomplete for auto-completion instead of ALE's built in auto-completion. |
|
@hsanson unfortunate to hear that patch 2 breaks autocompletion with asynccomplete. I do not know enough about how it interacts with ALE. Can you try with ale's built-in autocompletion function? One more question, I tried to look into the appveyor results, but for the life of me I couldn't find which tests broke. What should I be looking for? |
|
@s-marios seems that the new patterns need to be also added to this test: |
|
Ah, sorry about that. Amended the test, run locally all tests passed. History rewritten, rebased, it should be ready to go! |
* Add explicit trigger characters for C (dense-analysis#4226) * Stop completion before issuing subsequent requests (dense-analysis#4226) Co-authored-by: Marios Sioutis <[email protected]>
* Add explicit trigger characters for C (dense-analysis#4226) * Stop completion before issuing subsequent requests (dense-analysis#4226) Co-authored-by: Marios Sioutis <[email protected]>
In dense-analysis#4231 some code was added to stop the completion menu if any when opening a new one. This resulted in an issue in Vim that fills the buffer with Ctrl-Z characters when deleting to the end of a line in a position that triggers auto-completion. Since auto-completion seems to work fine on all my tests I am reverting this specific change.
In #4231 some code was added to stop the completion menu if any when opening a new one. This resulted in an issue in Vim that fills the buffer with Ctrl-Z characters when deleting to the end of a line in a position that triggers auto-completion. Since auto-completion seems to work fine on all my tests I am reverting this specific change.
In dense-analysis#4231 some code was added to stop the completion menu if any when opening a new one. This resulted in an issue in Vim that fills the buffer with Ctrl-Z characters when deleting to the end of a line in a position that triggers auto-completion. Since auto-completion seems to work fine on all my tests I am reverting this specific change.
This pull request addresses #4226 and comes in two patches:
For the first patch, ALE currently does not use the trigger characters advertised by the LSP. Thus it is necessary to hard-code trigger characters. This patch introduces
->in order to make autocompletion with pointers possible.The second patch is a proof-of-concept on how ALE should behave when a newer autocompletion request has been sent to the server. Currently, if there was a previous autocomplete request, the autocomplete operation will not stop if the user enters the full name of a variable and then a trigger character, which results in no completion candidates even though ALE did talk to the LSP and retrieved completion candidates successfully (see #4226 for this description). This patch stops the previous completion operation, and ALE is able to show a new autocompletion menu with the updated, more recent completion candidates.
Review and further comments on getting this merged are greatly appreciated!