Petr Spacek wrote:
On 2.8.2016 17:12, Rob Crittenden wrote:
Pavel Vomacka wrote:
Hello,

please review attached patches which Split make lint to more targets and
add jslint

What's the driver to split the checks out into separate targets?

Most importantly, makeapi and makeaci do not need to be called 10 times during
single build (and can be called in parallel, once our build system can manage
that).

No argument there, but this is what tickets are for: to outline the problem, let the group scope it, target a release, THEN do the work.

Not picking on Pavel here, but having to ask the purpose of a patch is a bad sign.

I jumped in only to slow the process down so these other things could be considered. Something I might normally have done during triage.

You are moving the makeapi and makeaci from version-update to lint. They were
in version-update for a reason: downstream builds do not call lint. Downstream
may patch code. API cannot break.

This is very true, I definitely agree. Even better, downstream should call as
much checks as possible.

lint is another matter. IIRC it was deemed to be less important because it is working against released code and giving the one doing the release the benefit of the doubt that they know what they are doing.

It may also be that RHEL 5 or 6 didn't have pylint at the time (git may say).

This split allows the downstream spec to call all tools which are available in
downstream as part of %check section of the spec, which was not possible
before the split.

%check is not appropriate for these.

If you want a proper %check then get IPA installable using socket wrappers and execute the in-tree unit tests.

rob

--
Manage your subscription for the Freeipa-devel mailing list:
https://www.redhat.com/mailman/listinfo/freeipa-devel
Contribute to FreeIPA: http://www.freeipa.org/page/Contribute/Code

Reply via email to