[PATCH] D49175: [clangd] Ignore sema code complete callback with recovery context.

[Eric Liu via cfe-commits]

Sounds good. Thanks for the suggestion!

On Wed, Jul 11, 2018 at 3:45 PM Sam McCall <sammccall at google.com> wrote:

>
>
> On Wed, Jul 11, 2018, 15:18 Eric Liu via Phabricator <
> reviews at reviews.llvm.org> wrote:
>
>> ioeric added a comment.
>>
>> In https://reviews.llvm.org/D49175#1158562, @sammccall wrote:
>>
>> > I like this idea, of course hard to know how it will affect all
>> practical cases.
>> >
>> > Is it easy to get have internal stats on how many wins/losses this has?
>>
>>
>> This would fix bad completion results when sema could recover to the
>> normal mode (e.g. completion in macros like EXPECT_THAT, expressions in if
>> statements). From the code completion metrics I collected from existing
>> code, the number of completions where no result was found dropped by 30%.
>> The loss seems to be we will get no result in recovery mode rather than bad
>> results (e.g. in excluded preprocessor branches).
>>
> Thanks! I have no doubt this is positive overall, so just thinking out
> loud here.
>
> For changes like this, we can show that quantitavely by running
> before/after completions at the same sites, and reporting e.g.
> Percent of all sites that are better vs worse vs neutral (in terms of rank
> of desired result)
> Change in MRR: mean/std deviation
>
> For our internal quality evaluation tools, we could think about how to
> make it easy to get these numbers for a candidate change.
> Having the raw results as before/after pairs is great for investigating
> wins/losses too.
>
>
>
>>
>> Repository:
>>   rCTE Clang Tools Extra
>>
>> https://reviews.llvm.org/D49175
>>
>>
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20180711/095b3241/attachment.html>