Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Testmatrix: Add a build where C extensions are compiled without optimizations #8994

Open
MSeifert04 opened this issue Jul 14, 2019 · 6 comments

Comments

@MSeifert04
Copy link
Contributor

The issue in #8986 is probably something that we should've catched with out CI. The problem there was that the implementation assumed that the function is always going to be inlined, however when there is a debug build or when the compiler decides not to inline it then the code is going to fail.

I assumed that we used debug builds for the C coverage report but given that these didn't fail we probably have at least some optimizations "on" there.

@mhvk
Copy link
Contributor

mhvk commented Jul 14, 2019

Good idea to have a debug build - in numpy those often found bugs. Presumably we can convert one of the regular builds (coverage sounds good)

@bsipocz
Copy link
Member

bsipocz commented Jul 15, 2019

@astrofrog @pllim - this is one of those package-novice, good first issues that are also effort low for someone familiar with CI and compilers, but one that I am not comfortable to label as package-novice, so let's try to figure out a new label.

@pllim
Copy link
Member

pllim commented Jul 15, 2019

Re: new label -- @bsipocz , I thought we used to have a "domain expert" label but I don't see it now.

@MSeifert04
Copy link
Contributor Author

MSeifert04 commented Jul 15, 2019

Currently our coverage build does use O0 and -fn-inline-functions so maybe the issue is in the compiler (or only observable in some compilers): https://github.com/astropy/astropy/blob/master/.travis.yml#L127-L134

@astrofrog
Copy link
Member

Domain expert isn't the right one though, that was more for astronomy expertise. Maybe having something like infrastructure-expert or something would be good here. I think package-novice is the right label to have, it's just we need additional labels to indicate we need good python or other expertise.

@bsipocz
Copy link
Member

bsipocz commented Jul 15, 2019

technical-expert? not really the right one either.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

5 participants