New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
make test tries to use a different perl than the one being built, /t/lib/croak.t t/lib/common.pl #12821
Comments
From @bulk88Created by @bulk88I did a make test on a blead and I got a bunch of fails, I have done The "../../lib" in @INC I think is coming from The simple answer is dont keep a build perl dir in a perl install dir, _______________________________________________________________________________ C:\p517\g\win32> __________________________________________________________________________________ __________________________________________________________________________________ C:\p517\g\t>..\perl.exe -I..\lib harness -v ./lib/croak.t Perl API version v5.17.9 of Win32 does not match v5.17.10 at Test Summary Report __________________________________________________________________________________ CWD.pm was modified to C:\p517>dir Directory of C:\p517 02/27/2013 04:37 AM <DIR> . C:\p517> Perl Info
|
From @nwc10On Wed, Feb 27, 2013 at 11:29:56AM -0800, bulk88 wrote:
I wasn't aware of what Win32CORE does - ie that it autoloads Win32 That has the potential for a lot of surprises in the test suite - calling Thanks for the very thorough diagnosis of this. I've pushed a fix for this
Yes, it's wrong that it's doing this, and the offending tests should be
Of the failures you see, I think that this will fix all but To try to replicate this, I made a booby-trapped lib directory one up from $ for file in `find lib -type d`; do mkdir -p ../$file; done I see the following failures: Test Summary Report porting/utils.t (Wstat: 0 Tests: 87 Failed: 79) So I don't see your ../dist/IO/t/cachepropagate-tcp.t failure, I can It's a shame that the NYC hackathon was last weekend, not next weekend, as I Nicholas Clark |
The RT System itself - Status changed from 'new' to 'open' |
From @bulk88Nicholas Clark wrote:
Will try that soon. I attached a crude *do not apply* fix that I did to
I have working files in the wrong directory, not poisoned files so more
cachepropagate-tcp.t always failing is unique to that machine, I dont |
From @bulk880001-crude-fix-for-116971-do-not-apply.patchFrom beb4dde25f4fd7ed3672934c5768e0e62a898451 Mon Sep 17 00:00:00 2001
From: bulk88 <bulk88@hotmail.com>
Date: Wed, 6 Mar 2013 15:09:34 -0500
Subject: [PATCH] crude fix for 116971 do not apply
---
t/lib/common.pl | 1 +
1 file changed, 1 insertion(+)
diff --git a/t/lib/common.pl b/t/lib/common.pl
index e6a33b2..82ab799 100644
--- a/t/lib/common.pl
+++ b/t/lib/common.pl
@@ -64,6 +64,7 @@ my $cleanup = 1;
END {
if ($cleanup) {
chdir '..' or die "Couldn't chdir .. for cleanup: $!";
+ shift @INC eq '../../lib' or die '@INC mangled';
rmtree($tempdir);
}
}
--
1.8.0.msysgit.0
|
From @bulk88On Wed Mar 06 04:01:12 2013, nicholas wrote:
I compiled _______________________________________________________________________________ op/time.t (Wstat: C:\p517\g\win32> C:\p517\g\t> Test Summary Report ../dist/IO/t/io_sock.t (Wstat: 0 Tests: 25 Failed: 0) C:\p517\g\t> C:\p517\g\t> End of (I think) irrelavent test fails. Now for %INC test related bugs. CWD C:\p517\g\t\ which calls CWD C:\p517\g\ which calls which is waiting with C callstack NtRaiseHardError does a call into the kernel to send an IPC message to
which I *think* (optimized build, watch window is often garbage) is my $libref = dl_load_file($file, 0) or do { A Dependency Walker log. Starting profile on 3/6/2013 at 7:23:19 PM Operating System: Microsoft Windows 2003 Server R2, Enterprise (64-bit), Options Selected: Started "PERL.EXE" (process 0x8A8) at address 0x0000000140000000. Since you reproduced (from -- |
From @bulk88[116971 re.dll.png](https://rt-archive.perl.org/perl5/Ticket/Attachment/1198897/618475/116971 re.dll.png) |
From @bulk88If it wasn't obvious from my last post, -- |
From @bulk88On Wed Mar 06 16:54:13 2013, bulk88 wrote:
http://perl5.git.perl.org/perl.git/commit/9c5b8d56084afc9da630a00b5bab6b257f62d1aa
While debugging, I found some code Anyways, There is a chunk of (modified by me) code in porting/pod_rules.t BEGIN { use strict; _________________________________________________________________ __________________________________________________________________ C:\p517\g\t> The chdir call in pod_rules.t blames to I can't say anything further since IDK what code is correct and what 1. this line in /t/TEST adds "-I.." to the child process's args . 2. the OOP nightmare called TAP:: should be but isn't chdiring from 3. pod_rules.t shouldn't be chdiring. Superficially, it looks like the podcheck.t # Verify that all files generated by perl scripts are up to date. BEGIN { use strict; require './test.pl'; plan('no_plan'); # Fail for every PERL_ARGS_ASSERT* macro that was declared but not used. my %declared; my $prefix = ''; unless (-d 't' && -f 'MANIFEST') { # Test that there are no missing Maintainers in Maintainers.pl BEGIN { chdir '..' unless -d 't'; -- |
From @bulk88This got a commit in blead at |
From @bulk88On Fri Jun 28 02:18:36 2013, bulk88 wrote:
http://perl5.git.perl.org/perl.git/commit/96d268e2f48e69b4cb65326df6690ffc21120f3c
using a blead 5.19.6 this is still a problem but only porting/utils.t is Summary of my perl5 (revision 5 version 19 subversion 6) configuration: Characteristics of this binary (from libperl): attaching console log of failures -- |
From @bulk88# Porting/bisect-example.sh isn't a Perl script |
From @bulk88On Tue Oct 22 14:31:33 2013, bulk88 wrote:
Still a problem. Unhandled exception at 0x10004d4f (Util.dll) in perl.exe: 0xC0000005: Access violation reading location 0x0a0d0a0d.
perl.exe 00400000-00408000 C:\perl519\src\t\perl.exe 1 [2796] perl.exe: Native 1/8/2014 9:37 PM Symbols loaded. line 1064 is http://perl5.git.perl.org/perl.git/blob/78add018a01221839c23f0cfecc3110bb28d038f:/cpan/Scalar-List-Utils/ListUtil.xs#l1064 " rmcsv = GvSVn(rmcgv);" but its not really important where it crashed, notice C:\perl519\lib\auto\List\Util\Util.dll is a system perl, but C:\perl519\src\t\perl.exe is a "make test" perl. Cmd line of the process that crashed was "C:\perl519\src\t\perl.exe "-I../lib" -c "Porting/acknowledgements.pl"" and CWD at time of crash is "C:\perl519\src\"., the crashed process was not called by harness, but from "./perl -I.. -MTestInit porting/utils.t" that was called from harness. -- |
From @tonycozOn Tue Oct 22 14:31:33 2013, bulk88 wrote:
I believe this is now fixed in blead. I used a poisoned ../lib similar to what Nicholas described, this caused failures Fixed in b6e1fb8 (utils.t) and 7f1eeb9 (manifest.t). I also modified pending-author.t to use File::Spec->devnull instead of a literal /dev/null, which allows this test to work on Win32. Tony |
@tonycoz - Status changed from 'open' to 'resolved' |
Migrated from rt.perl.org#116971 (status was 'resolved')
Searchable as RT116971$
The text was updated successfully, but these errors were encountered: