These are chat archives for coala/coala-bears

16th
Apr 2018
Dopamine Gamer
@DopamineGamer_twitter
Apr 16 2018 01:45
@jayvdb Would you know how the confidence parameter affects the results/reports? I couldn’t find in documentation.
Boxuan Li
@li-boxuan
Apr 16 2018 01:48
confidence is just a number between 0 and 100 describing the likelihood of this result being a real issue. What do you mean by how it affects the results?
Dopamine Gamer
@DopamineGamer_twitter
Apr 16 2018 02:19
@li-boxuan How is it used by the system? Does it report this to the user?
Boxuan Li
@li-boxuan
Apr 16 2018 02:20
If you use --json flag, then yes
Dopamine Gamer
@DopamineGamer_twitter
Apr 16 2018 02:21
Ah, okay. Thank you.
Boxuan Li
@li-boxuan
Apr 16 2018 02:21
I'm not sure how it is used by system. Maybe you can look into code :wink:
Dopamine Gamer
@DopamineGamer_twitter
Apr 16 2018 02:21
Of course. Thanks.
@li-boxuan Just looked. Seems like it’s only for reporting.
Dopamine Gamer
@DopamineGamer_twitter
Apr 16 2018 02:29
@Makman2 So, coming back to the patch conflict and multiprocessing issue, why not apply the concurrency across source files instead of bears? Can coala be configured to do that now?
Dopamine Gamer
@DopamineGamer_twitter
Apr 16 2018 03:56
@Makman2 pip show coala is raising an exception because of your name :smile:
Sangam Kumar
@sangamcse
Apr 16 2018 07:34
@newbazz, @bekicot, have a look at coala/coala-bears#2235.
Lasse Schuirmann
@sils
Apr 16 2018 09:10
hey @DopamineGamer_twitter what question do you have about libclang?
John Vandenberg
@jayvdb
Apr 16 2018 09:30
clang ast being shared by multiple bears
not derived stats, but the ast data structures itself
Mischa Krüger
@Makman2
Apr 16 2018 17:18
@DopamineGamer_twitter I'm not sure if that makes sense, and from implementation perspective this would highly complicate things (job queues for each file? hmmm...). And no coala can't do it, and the NextGen-Core can't do that either. You could work around that by using locks I guess, but please don't do that, that is just not clean, breaks quickly, doesn't match our concept. The question would be also then: What bear gets priority working on a file when multiple bears want to work on it? And what if you run two bears, and both emit patches, and you don't accept the first one, what happens with the second patch? Because it builds on the first^^
In ideal cases we need to write bears where analysis doesn't conflict. If there are bears that handle this by being applied sequentially, run coala with a section order or even run coala multiple times.

pip show coala is raising an exception because of your name :smile:

Then you should use pip3 ;)

Rohan
@rohanrk
Apr 16 2018 18:51
May I be assigned to coala/coala#5383?
Dopamine Gamer
@DopamineGamer_twitter
Apr 16 2018 22:09

@Makman2 Implementation difficulties aside, why do you think it doesn’t make sense, and doesn’t align with coala’s concepts? I’m not pushing for a new radical feature here. Just trying to understand why things are the way they are. It’ll help me set things up.

What bear gets priority working on a file when multiple bears want to work on it?

Is there a concept of priority now? The bears could run in any order, considering the dependencies.

And what if you run two bears, and both emit patches, and you don't accept the first one, what happens with the second patch? Because it builds on the first

I don’t see a problem. Why wouldn’t it work the same way as now as if --jobsis set to 1?

I have a small bunch of bears which perform simple analysis. I have a large set of files waiting for analysis. Instead of splitting these simple tasks into multiple bears to achieve concurrency (with the associated problem of patch conflicts), I was just wondering why not assign each file’s processing to a process in a pool, wherein the bears will be run sequentially on the file. There wouldn’t be any patch conflict this way. Usually, the set of bears is limited, but the number of source files is more.

@sils Hi Lasse. Thanks for getting to me. I figured a few things in the last few days. I’ll get back to you later.

@jayvdb

clang ast being shared by multiple bears

I know this can’t happen with the process pool.

Then you should use pip3 :wink:

Same issue with pip3 as well. This is trivial, but does it work for you, @Makman2?