Re: Test suite - commit process clarifications
On 5/19/26 7:08 PM, Eric Covener wrote: On Tue, May 19, 2026 at 12:37 PM Joe Orton wrote: On Mon, May 18, 2026 at 12:09:16PM +0200, Giannis Christodoulou wrote: Hi all, I was reading through the project guidelines [1] and the 2.4.x STATUS file [2] and had a couple of questions about the process for test suite changes. The guidelines say that changes to the active repository (trunk) are subject to lazy consensus, while changes to a prior branch (2.4.x) require consensus before commit. They also distinguish between ideas (RTC) and patches (CTR). In addition, the 2.4.x STATUS file has a list of RTC exceptions and tests are not mentioned there. So my questions are: 1. Does adding or fixing/updating a test count as a patch, hence a committer can push it to trunk under lazy consensus? Yes, definitely. 2. Should test-only changes to 2.4.x remain under full RTC? I know that a bad test can mask real issues but since they don't affect runtime behavior I think it could have a lighter process. This is fine and is covered by "CI configuration and scripts" in my view. +1 but we should make it explicit (and there is no RTC in httpd-tests) Agreed. -- Cheers Jean-Frederic
Re: Test suite - commit process clarifications
On Tue, May 19, 2026 at 12:37 PM Joe Orton wrote: > > On Mon, May 18, 2026 at 12:09:16PM +0200, Giannis Christodoulou wrote: > > Hi all, > > > > I was reading through the project guidelines [1] and the 2.4.x STATUS file > > [2] and had a couple of questions about the process for test suite changes. > > > > The guidelines say that changes to the active repository (trunk) are subject > > to lazy consensus, while changes to a prior branch (2.4.x) require consensus > > before commit. > > They also distinguish between ideas (RTC) and patches (CTR). In addition, > > the 2.4.x STATUS file has a list of RTC exceptions and tests are not > > mentioned there. > > > > So my questions are: > > > > 1. Does adding or fixing/updating a test count as a patch, hence a > >committer can push it to trunk under lazy consensus? > > Yes, definitely. > > > 2. Should test-only changes to 2.4.x remain under full RTC? I know that > >a bad test can mask real issues but since they don't affect runtime > >behavior I think it could have a lighter process. > > This is fine and is covered by "CI configuration and scripts" in my > view. +1 but we should make it explicit (and there is no RTC in httpd-tests)
Re: Test suite - commit process clarifications
On Mon, May 18, 2026 at 12:09:16PM +0200, Giannis Christodoulou wrote: > Hi all, > > I was reading through the project guidelines [1] and the 2.4.x STATUS file > [2] and had a couple of questions about the process for test suite changes. > > The guidelines say that changes to the active repository (trunk) are subject > to lazy consensus, while changes to a prior branch (2.4.x) require consensus > before commit. > They also distinguish between ideas (RTC) and patches (CTR). In addition, > the 2.4.x STATUS file has a list of RTC exceptions and tests are not > mentioned there. > > So my questions are: > > 1. Does adding or fixing/updating a test count as a patch, hence a >committer can push it to trunk under lazy consensus? Yes, definitely. > 2. Should test-only changes to 2.4.x remain under full RTC? I know that >a bad test can mask real issues but since they don't affect runtime >behavior I think it could have a lighter process. This is fine and is covered by "CI configuration and scripts" in my view. Regards, Joe
