You are viewing a plain text version of this content. The canonical link for it is here.
Posted to test-dev@httpd.apache.org by Geoffrey Young <ge...@modperlcookbook.org> on 2003/12/08 20:48:14 UTC
status of the perl-framework
hi all
currently, the perl-framework for 1.3, 2.0, and cvs fail miserably for me,
and have for some time.
1.3)
Failed Test Stat Wstat Total Fail Failed List of Failed
-------------------------------------------------------------------------------
apache/chunkinput.t 9 2 22.22% 7 9
apache/contentlength.t 255 65280 20 38 190.00% 2-20
apache/limits.t 10 1 10.00% 9
modules/autoindex2.t 3 2 66.67% 1 3
modules/include.t 9 2304 ?? ?? % ??
2.0)
Failed Test Stat Wstat Total Fail Failed List of Failed
-------------------------------------------------------------------------------
modules/autoindex.t 84 42 50.00% 43-84
modules/include.t 9 2304 ?? ?? % ??
2.1)
Failed Test Stat Wstat Total Fail Failed List of Failed
-------------------------------------------------------------------------------
modules/include.t 9 2304 ?? ?? % ??
anyway, my goal is to have the perl framework successful again for all httpd
versions by the end of the week. to that end, I'm going to make whatever
adjustments are required and commit as I go - feel free to speak up if
something doesn't look right.
--Geoff
Re: status of the perl-framework
Posted by Geoffrey Young <ge...@modperlcookbook.org>.
>>we can capture the 2.0/2.1 differences with have_min_apache_version() or
>>similar logic, so no worries there either.
>
>
> NOOOO! While we appreciate your efforts to have the *tests* run properly,
> please do *NOT* mask legitimate failures in your efforts to have 'clean'
> output from the framework. As the tests grow, we discover that some of
> our long standing expectations aren't met - and that both 1.3 and 2.0 may
> exhibit issues. Don't mask those issues - they need to be called out.
indeed. it was never my intention to just "fix" things so that the tests
pass.
with test suites there is always the question of whether the tests or the
code is wrong. right now some tests are failing, and it's been a while
since I've seen activity from the people who wrote them. so, what I was
offering to do was take the time to analyze each of the failures, track them
down in the code, see where the discrepancy is, write patches and enter bug
reports if the test expectation is reasonable but the code is not, etc.
>
> So do fix the *tests* - do not *fix* the tests :)
that's the idea :)
I'm sure we're all on the same page - virtual discourse can sometimes get in
the way. I'm a big fan of test-based development, so I'll try and do the
right thing :)
--Geoff
Re: status of the perl-framework
Posted by "William A. Rowe, Jr." <wr...@apache.org>.
At 02:02 PM 12/8/2003, Geoffrey Young wrote:
>so, this is not a big deal, and it's not just mod_include that has warning
>troubles. so, I'll tidy up the perl and that will help a few things along.
No arguement there, if the *perl-framework* is broken your fixes are
appreciated.
>> 2.1 runs fine for me. 2.0 has some known flaws which are fixed with my rewrite
>> (waiting for backport).
>
>we can capture the 2.0/2.1 differences with have_min_apache_version() or
>similar logic, so no worries there either.
NOOOO! While we appreciate your efforts to have the *tests* run properly,
please do *NOT* mask legitimate failures in your efforts to have 'clean'
output from the framework. As the tests grow, we discover that some of
our long standing expectations aren't met - and that both 1.3 and 2.0 may
exhibit issues. Don't mask those issues - they need to be called out.
So do fix the *tests* - do not *fix* the tests :)
Bill
Re: status of the perl-framework
Posted by Stas Bekman <st...@stason.org>.
Geoffrey Young wrote:
>>I'm not sure. IMHO we should leave the failures and coredumps until the module
>>is fixed :-)
>
>
> well, my idea (at least) is to have a clean test suite for 1.3, 2.0, and
> 2.1. theoretically, there may be "broken" code in 2.0 indefinitely - places
> where the answer is "yes, we know it's broken. please upgrade." in cases
> like that, it's probably more proper to skip (or todo) the test on one
> platform, rather than having the tests constantly (knowingly) fail.
I suppose for each failure you need to ask the developers whether they think
it should be fixed in 2.0 or not. If yes, then mark it as a todo item (don't
just skip it). If not, then certainly skip it and add a reason why it's going
to stay broken in 2.0.
__________________________________________________________________
Stas Bekman JAm_pH ------> Just Another mod_perl Hacker
http://stason.org/ mod_perl Guide ---> http://perl.apache.org
mailto:stas@stason.org http://use.perl.org http://apacheweek.com
http://modperlbook.org http://apache.org http://ticketmaster.com
Re: status of the perl-framework
Posted by Cliff Woolley <jw...@virginia.edu>.
On Mon, 8 Dec 2003, Geoffrey Young wrote:
> > Sounds like a cvs conflict marker.
> duh. that was it. eesh.
I only knew that because it happens to me all the time. ;) nbd.
--JC
Re: status of the perl-framework
Posted by Geoffrey Young <ge...@modperlcookbook.org>.
Cliff Woolley wrote:
> On Mon, 8 Dec 2003, Geoffrey Young wrote:
>
>
>>>>Use of bare << to mean <<"" is deprecated at modules/include.t line 120.
>>>>
>>>>which is probably new to 5.8.2.
>>>
>>>
>>>Interesting. << doesn't occur within include.t. So it happens within the
>>>framework?
>>
>>hmm, perhaps. I was actually planning on getting down and dirty tomorrow :)
>> but yes, it could be the framework or an interaction between them.
>
>
> Sounds like a cvs conflict marker.
duh. that was it. eesh.
--Geoff
Re: status of the perl-framework
Posted by Cliff Woolley <jw...@virginia.edu>.
On Mon, 8 Dec 2003, Geoffrey Young wrote:
> >>Use of bare << to mean <<"" is deprecated at modules/include.t line 120.
> >>
> >>which is probably new to 5.8.2.
> >
> >
> > Interesting. << doesn't occur within include.t. So it happens within the
> > framework?
>
> hmm, perhaps. I was actually planning on getting down and dirty tomorrow :)
> but yes, it could be the framework or an interaction between them.
Sounds like a cvs conflict marker.
--Cliff
Re: status of the perl-framework
Posted by Geoffrey Young <ge...@modperlcookbook.org>.
>>Use of bare << to mean <<"" is deprecated at modules/include.t line 120.
>>
>>which is probably new to 5.8.2.
>
>
> Interesting. << doesn't occur within include.t. So it happens within the
> framework?
hmm, perhaps. I was actually planning on getting down and dirty tomorrow :)
but yes, it could be the framework or an interaction between them.
>>sound ok?
>
>
> I'm not sure. IMHO we should leave the failures and coredumps until the module
> is fixed :-)
well, my idea (at least) is to have a clean test suite for 1.3, 2.0, and
2.1. theoretically, there may be "broken" code in 2.0 indefinitely - places
where the answer is "yes, we know it's broken. please upgrade." in cases
like that, it's probably more proper to skip (or todo) the test on one
platform, rather than having the tests constantly (knowingly) fail.
where I'm going with all of this is that if we have a clean test suite for
all versions, then maybe it would help to relieve the burden from core folks
if patchers could say "patch attached, all tests pass." not that it would
replace other due diligence factors, mind you, but it might make the
difference to some core developer in deciding whether to shepherd a patch in
if they knew the didn't need to mess around with even getting it to compile
or the tests to pass.
now, you can't change a community, and I'm not trying to (well, not directly
anyway :) however, none of the above can happen without a clean test suite,
so I'm doing what I can in case it makes a difference. if you build it,
they will come.
but contrary opinions welcome :)
--Geoff
Re: status of the perl-framework
Posted by André Malo <nd...@perlig.de>.
* Geoffrey Young <ge...@modperlcookbook.org> wrote:
> André Malo wrote:
> > * Geoffrey Young <ge...@modperlcookbook.org> wrote:
> >
> >
> >>modules/include.t 9 2304 ?? ?? % ??
> >
> >
> > what does this null information mean?
>
> pretty much that the script died before it could complete. for 2.1, for
> instance, it's a result of this warning
>
> Use of bare << to mean <<"" is deprecated at modules/include.t line 120.
>
> which is probably new to 5.8.2.
Interesting. << doesn't occur within include.t. So it happens within the
framework?
> we can capture the 2.0/2.1 differences with have_min_apache_version() or
> similar logic, so no worries there either.
>
> basically, this kind of tidying was what I had in mind - I really doubt that
> there will be much in the logic of the tests that will need to change, if
> anything.
>
> sound ok?
I'm not sure. IMHO we should leave the failures and coredumps until the module
is fixed :-)
nd
Re: status of the perl-framework
Posted by Geoffrey Young <ge...@modperlcookbook.org>.
André Malo wrote:
> * Geoffrey Young <ge...@modperlcookbook.org> wrote:
>
>
>>modules/include.t 9 2304 ?? ?? % ??
>
>
> what does this null information mean?
pretty much that the script died before it could complete. for 2.1, for
instance, it's a result of this warning
Use of bare << to mean <<"" is deprecated at modules/include.t line 120.
which is probably new to 5.8.2.
so, this is not a big deal, and it's not just mod_include that has warning
troubles. so, I'll tidy up the perl and that will help a few things along.
>
>
>>2.1)
>>Failed Test Stat Wstat Total Fail Failed List of Failed
>>---------------------------------------------------------------------------
>>---- modules/include.t 9 2304 ?? ?? % ??
>
>
> 2.1 runs fine for me. 2.0 has some known flaws which are fixed with my rewrite
> (waiting for backport).
we can capture the 2.0/2.1 differences with have_min_apache_version() or
similar logic, so no worries there either.
basically, this kind of tidying was what I had in mind - I really doubt that
there will be much in the logic of the tests that will need to change, if
anything.
sound ok?
--Geoff
Re: status of the perl-framework
Posted by André Malo <nd...@perlig.de>.
* Geoffrey Young <ge...@modperlcookbook.org> wrote:
> modules/include.t 9 2304 ?? ?? % ??
what does this null information mean?
> 2.1)
> Failed Test Stat Wstat Total Fail Failed List of Failed
> ---------------------------------------------------------------------------
> ---- modules/include.t 9 2304 ?? ?? % ??
2.1 runs fine for me. 2.0 has some known flaws which are fixed with my rewrite
(waiting for backport). What happens at your checkout?
nd