This is the mail archive of the
mailing list for the GCC project.
Re: [patch] for PR 27735
On Sun, Aug 13, 2006 at 11:24:19PM -0700, Mark Mitchell wrote:
> Eric Botcazou wrote:
> >> However, I think that this is up to the testsuite maintainers to decide.
> >> I suggest Janice offer her opinion. Furthermore, since, surely, having
> >> the test is more important than worrying about it's name, I suggest we
> >> agree just to accept Janice's determination without comment.
> > Well, I don't think we would necessarily do Janice a favour by asking her for
> > every single testcase we add to the testsuite.
> No, but she doesn't have to -- she can just tell us once which format we
> should use, or whether it's at the discretion of the person creating the
> test case.
I was afraid someone would bring me into this! I really don't care
strongly about testsuite names. Sometimes it makes sense to use a name
that refers to the functionality being tested, some directories include
PR numbers in the name, and of course there's the traditional name with
the date the test was added, which is nice because it lets people notice
immediately that a new failure is due to a new test. In the abstract it
might be nice to have a single convention and stick with it, but I for
one, don't really care.
I also don't care whether a test includes the name of the person who
added it, although it's OK to include it. A test associated with a PR
should definitely mention the PR number in a comment, unless it's
already the name of the PR. Comments about the purpose of a test are
important, unless it's obvious from looking at the PR.