e95ff8baba
Changelog entries for 3.002, too late for release. oops. and what exactly where the 3.001 changes? oh well.
Scott Walters2014-09-16 17:49:01 -04:00
cad782dbf2
bump version to 3.002, and also change each file from having their own version number to referring to $SOAP::WSDL::VERSION since they aren't really individually versions anyway but just copies of that number.
Scott Walters2014-09-16 17:40:18 -04:00
99c5eb4dd7
remove META.yml (I still don't know how that's supposed to work) and Makefile (which I do know wasn't supposed to be in git)
Scott Walters2014-09-16 03:29:56 -04:00
ed715a4c88
smoke testing looks pretty good, only failing on amd64-netbsd machines and one random Linux machine (I should look why but I really don't want to know), so, bump version to 3.001 and do a release. FAIL 5.10.1 NetBSD 6.1.4 amd64-netbsd FAIL 5.14.2 GNU/Linux 3.11.0-18-generic x86_64-linux-thread-multi FAIL 5.10.1 NetBSD 6.1.4 amd64-netbsd-thread-multi
Scott Walters2014-08-29 15:05:06 -04:00
f5787f6057
version bump to 3.00.0_2 and Changes entry
Scott Walters2014-08-28 21:16:50 -04:00
374793ef68
get rid of Build.PL; we use ExtUtils::MakeMaker now.
Scott Walters2014-08-28 21:08:13 -04:00
7f6af60ed8
oops, global version search and replace ruined a comment. leave it ruined
Scott Walters2014-08-28 21:05:13 -04:00
37740b870a
fix a hash randomization bug that was causing random failures in t/SOAP/WSDL/05_simpleType-list.t
Scott Walters2014-08-28 21:03:39 -04:00
31f5deacf5
2624a9 for #86142, adding the XML declaration to the output, broke another part of the system, so work around that. I accidentally stopped doing "recursive" tests (tests in subdirs) at some point and missed some of this fun breakage.
Scott Walters2014-08-28 21:01:51 -04:00
3197383a7d
update POD; thought I commit this already... ?
Scott Walters2014-08-28 20:38:31 -04:00
285ffa463c
okay, dam@cpan.org's fix for using not using ->usa('UNIVERSAL') to test if something loaded apparently works in a situation where mine doesn't, that was causing tests to fail.
Scott Walters2014-08-28 18:26:41 -04:00
35321ce5d0
run tests recursively under ExtUtils::MakeMaker.
Scott Walters2014-08-28 16:25:55 -04:00
08b156a546
take Module::Build out of inc. it doesn't work. don't use it.
Scott Walters2014-08-28 16:25:29 -04:00
1994b36d8b
switch from Module::Build to Make::Maker based Makefile.PL after Module::Build vomited on trying to 'make dist', and then when that was fixed, generated a completely busted dist without error. good riddance.
Scott Walters2014-08-27 17:53:50 -04:00
f3ae85aed5
default directory for t/098_pod.t. not sure why it was doing @dir = (); if(ENV var) { @dir = 'some dir that exists' }. this would make a path analysis tool cry.
Scott Walters2014-08-27 17:52:41 -04:00
7d81cfce87
bump version to 3.00.00_1; update POD a bit; include contact info for myself; declare this module deprecated and in maintenance mode; include Module::Build in the 'inc' dir to avoid a warning about it being deprecated from core inclusion; add mention of the UNIVERSAL->isa() bug created by Class::Std::Fast to the POD for BUGS AND LIMITATIONS.
Scott Walters2014-08-27 17:39:52 -04:00
45982ff330
Huh. Between 5.12 and 5.16 somewhere, the behavior of ClassName->isa('UNIVERSAL') changed. It used to be that that would return true if and only if that namespace existed. Now it always returns true for any random non-existant made up name. Changed this to do ->can() on a known existing method modules with this API have (serialize) instead after floundering around for a bit. exists ${"main::"}{$type.'::'} worked a little bit but made it barf for some reason I didn't investigate. This is the problem apparently behind the previous "haunted house level shit" fixes and reversions. This thing does automatically load these modules on the fly. Would be nice if the unit tests had some comments in them.
Scott Walters2014-08-21 08:37:49 -04:00
cfd5892b18
Revert "was barfing with "Can't locate object method "new" via package "MyElement" (perhaps you forgot to load "MyElement"?) at lib/SOAP/WSDL/XSD/Typelib/Comple" ref haunted house level shit going weird message in previous revert.
Scott Walters2014-08-21 07:48:30 -04:00
6e95d01610
Revert "maybe I'm doing the wrong thing by adding use lines for test modules" Wow. Even with the "fix" in, it runs correctly once, and then starts vomiting with a similar error again as before the fix. it'll fail 20 times in a row, then pass 20 times in a row. there's some haunted house level shit going wrong here.
Scott Walters2014-08-21 07:47:02 -04:00
b121a061d8
maybe I'm doing the wrong thing by adding use lines for test modules that unit tests use. maybe there's some magic that's supposed to automatically use those things, even though they're hidden in deeply buried paths with no machine predictable naming scheme. I can't imagine that these unit tests passed before unless there is/was some magic. I hobbled this in to working with a use lib and a use for the module, but after doing this second one of these, I'm having serious doubts about whether I'm just spackling over the problem.
Scott Walters2014-08-21 07:37:22 -04:00
d2f324a7f2
oops, changing error messages breaks tests. make them match again. don't change the tests. change the error messages. that way the error message's awkwardness hints at the code's awkwardness.
Scott Walters2014-08-21 07:34:19 -04:00
f2ce9511e0
I don't get it. There were two calls to get_port() right next to each other, in the very same ? :, and one of them had a ->[0] tacked on the end and the other one didn't. Well guess what... the one without was returning an arrayref, which caused these failures in the unit tests: Can't call method "get_binding" on unblessed reference at /home/scott/projects/SOAP-WSDL/blib/lib/SOAP/WSDL.pm line 186.
Scott Walters2014-08-21 07:27:58 -04:00
50c8d5b31f
oh, look, ComplexType.pm has cut and paste hash reversing code from SimpleType.pm. that means that we can cut and paste the fix from SimpleType.pm into ComplexType.pm. how many wrongs make a right? I forget.
Scott Walters2014-08-21 07:26:30 -04:00
47dc61bd9e
protip: reverse %hash with '#default' => 'urn:myNamespace', 'tns' => 'urn:myNamespace' results in misery. random hash ordering made a pile of tests in t/003_wsdl_based_serializer.t randomly pass and fail in unison. I smelled a smell.
Scott Walters2014-08-21 06:32:11 -04:00
3670df3e69
"vectors" (totally not the same thing as a vector in other languages or on a Cray 1) have been deprecated and removed. pretend like they never existed.
Scott Walters2014-08-21 05:20:56 -04:00
d221e87b29
was barfing with "Can't locate object method "new" via package "MyElement" (perhaps you forgot to load "MyElement"?) at lib/SOAP/WSDL/XSD/Typelib/Comple xType.pm line 213.". Okay, then, MyElement.pm exists in t/lib. Let's try using it. And it worked. Incidentally, the code at lib/SOAP/WSDL/XSD/Typelib/ComplexType.pm is heinous.
Scott Walters2014-08-21 05:17:05 -04:00
001c36f8e8
fix two instances of the "Can't modify non-lvalue subroutine call at lib/SOAP/WSDL.pm line 167." error. it looks like the author wrote
Scott Walters2014-08-20 08:37:33 -04:00
dlage
Scott Walters