Colm MacCarthaigh wrote:
> On Thu, Jan 26, 2006 at 09:57:28PM -0500, Geoffrey Young wrote:
> 
>>hmph.  I'm sorry you had to go through all that, but next time just speak up
> 
> 
> I didn't know where to :) 

you're probably not alone - the lists have recently changed.

> I was actually surprised I could commit to it,

here's the history...

Apache-Test started as part of the mp2 project.  then the httpd-test project
was started (owned by httpd) and Apache-Test spun off as the engine for the
perl-framework under httpd-test.  then more things got added to httpd-test
and only mod_perl people were working on Apache-Test proper anyway, so last
year we took it back over under our project.  but all the original
committers to httpd-test (all of httpd) were allowed to keep their commit
privs to Apache-Test, kind of as a courtesy.

note that A-T is actually a separate project using svn externals on checkout
of httpd-test:

  http://svn.apache.org/repos/asf/perl/Apache-Test/

> and then I thought [EMAIL PROTECTED] was the place for discussion and assumed 
> it
> was dead. It was only later that I found this list, so my bad.

no, that's ok.  Apache-Test was given its own list (this one) when we took
it back over last year, making two lists - one for A-T and one for the rest
of httpd-test (the now dead [EMAIL PROTECTED]).  then maybe last month justin
retired [EMAIL PROTECTED] and moved discussion over to httpd.

so, the way things stand now, A-T probems/issues/discussion should happen
here, while the rest of httpd-test (including perl-framework/t) should
happen on httpd.  the idea was to get the real httpd tests discussed in the
right place, while still keeping the perl engine (A-T) visible and
maintained by the ASF people that know perl the best (and who happen to be
the nearly sole creators and maintaners of it, since we use it for the mp2
test suite and maintain a separate distribtion of just A-T).

> 
> 
>>- I would have happily worked on fixing it rather than have you go through a
>>few different iterations trying to make it work then backing the whole thing
>>out.  sorry I don't follow new-httpd as much as I should...
> 
> 
> I'll update our docs later to be more clear on where things are tracked,
> here is a fine place, I don't think you should have to track  the httpd
> list, it's pretty noisy!

:)

if you document that perl-framework/Apache-Test is really a separate project
in and of itself, under the direction of the perl pmc, and has its own
mailing list it should be a bit more clear.

really, I'd like to restrict commit access to just mod_perl folks (extended
to httpd people that have contributed and understand the project), or at
least ask that non-familiar-with-the-project folks follow RTC, since nobody
seems to enjoy coding perl quite as much as we do :)  but I don't want to
offend anyone either.

> 
> 
>>anyway, the problem was mostly a variable stomping issue.  but with php in
>>particular it was a cut and paste error: need_php4 was checking for php4 or
>>php5 instead of php4 and sapi_apache.c.  and that was even carried over to
>>the current version as well ;)
>>
>>so, should be all fixed now, so give it another whirl.
> 
> 
> Yep, no more errors or dubious results.

excellent.

> 
> 
>>oh, I tweaked need_imagemap a bit for you and added a brief doc while I was
>>at it ;)
> 
> 
> Thank you very much!

sure.  just for your own knowledge there exists two sets of testing methods
- need_* and have_* - such that things like need_imagemap() ought to have a
have_imagemap() counterpart.  the difference is that need_imagemap()
populates the skip message so can be used with plan(), while have_imagemap()
performs the exact same check without populating the skip message.  the net
result is that you can use have_imagemap() anywhere in your test code to
skip individual tests without affecting the overall skip message.

anyway, so every time you implement a need_* function you ought to do have_*
as well.  luckily the have_* counterparts are automagically created - all
you need to do is populate the @need array at the top of Test.pm with the
name of your new method.

just FYI if you're interested :)

--Geoff

Reply via email to