Enable the usage of selftests

[CohenArthur]

From what I can understand, self tests seem to only be used in the C family of languages and in no other gcc frontend. And for some reason one of the self test for the gccrs frontend seem to be the parsing of C code and C types, which fails, which is why this PR is a draft. Once this is fixed, selftests should be enabled for the frontend and we should be able to add some to the project.

[CohenArthur]

@philberty @tschwinge I ended up returning early from the "C tests" that are in gcc files common to all languages in 606250b. This is definitely not a good solution and I will also ask for advice on the mailing list as to how to handle the situation. I think these tests could be moved, or maybe we might need to add some extra flags to not run them if we are running the rust selftests.

[CohenArthur]

But we can now have selftests and their failure is caught by the CI :)

[tschwinge]

@CohenArthur, given the replies to your http://mid.mail-archive.com/20211105093847.urtlztspddvlmahz@oneshade "C-family selftests in language-independant source files", do you know how to proceed or would like me to look into it at this point? (No worries, of course, if you've just not yet gotten around to continuing that.)

[CohenArthur]

Hi @tschwinge ! Thanks for offering your help :D I think I know how to proceed, or at least where to start, but I haven't gotten around to it. Thanks for the review, I'll fix up that commit

[CohenArthur]

I think it should be good now @tschwinge @philberty :) I'll add some unit tests for the base 62 code as a starting point

[philberty]

nice work @CohenArthur this is something I am really excited about! :D

[philberty]

As far as i can tell this looks good but I am not sure about the final commit 777d1d2 do we need to touch this or can we drop it?

[tschwinge]

Great; I'll have a look.

[CohenArthur]

As far as i can tell this looks good but I am not sure about the final commit 777d1d2 do we need to touch this or can we drop it?

The last commit is "extremely important" as it prevents our selftests from running some C specific selftests which we obviously fail. This is the commit that fixes what David Malcolm was suggesting over on the gcc mailing list

[dkm]

As far as i can tell this looks good but I am not sure about the final commit 777d1d2 do we need to touch this or can we drop it?

The last commit is "extremely important" as it prevents our selftests from running some C specific selftests which we obviously fail. This is the commit that fixes what David Malcolm was suggesting over on the gcc mailing list

Should this be sent upstream instead ? If that was suggested by David, that should be easy to have it merged ?

[tschwinge]

We're getting there! :-)

[CohenArthur]

@tschwinge Thanks a lot for the review and offer to help getting one commit upstream. @dkm I think this commit should definitely be sent upstream and I'll start looking into it!

[CohenArthur]

Commits have been fixed-up as talked about during the jitsi call :)

[tschwinge]

... and my response simply is:

bors: merge

[bors]

Build succeeded:

• build-and-check

[tschwinge]

As far as i can tell this looks good but I am not sure about the final commit 777d1d2 do we need to touch this or can we drop it?

The last commit is "extremely important" as it prevents our selftests from running some C specific selftests which we obviously fail. This is the commit that fixes what David Malcolm was suggesting over on the gcc mailing list

Should this be sent upstream instead ? If that was suggested by David, that should be easy to have it merged ?

🎉 This has recently been pushed (commit 9815244 "selftest: Move C-specific tests to c_family" -- @CohenArthur's first in GCC upstream!), and I'm working on another #247 merge, and then also resolve any local divergence we may have in these selftest files. (..., but first need to resolve some other GCC/Rust local breakage re *.opt files; stay tuned...)