[BBC]: Commit 1381350 (Time::HiRes synch with CPAN) breaks Catalyst-Runtime and others #16474
Comments
From @jkeenanTime-HiRes is a distribution maintained under dist/, so it is supposed One such release to CPAN has broken some code in CPAN distributions Details: 1. Recent commit to blead synching Time-HiRes with CPAN: ##### Upgrade Time::HiRes from version 1.9753 to 1.9757 [DELTA] 1.9577 [2018-03-15] 1.9756 [2018-03-14] 1.9755 [2018-03-14] 1.9754 [2018-02-16] It appears that Time-HiRes version 1.9754 touched off the CPAN test 2. Time-HiRes bug ticket: https://rt.cpan.org/Ticket/Display.html?id=124450 3. Catalyst-Runtime bug ticket: 4. Devel-TimeStats issue: Thank you very much. |
From @jkeenanThis is perl 5, version 27, subversion 11 (v5.27.11 (v5.27.9-210-g18503cb)) built for x86_64-linux Copyright 1987-2018, Larry Wall Perl may be copied only under the terms of either the Artistic License or the Complete documentation for Perl, including FAQ lists, should be found on |
From @toddrWasn’t this addressed by 1.9758? |
|
The RT System itself - Status changed from 'new' to 'open' |
From @jkeenanOn Fri, 23 Mar 2018 04:39:07 GMT, TODDR wrote:
I hadn't seen that commit at the time I filed the bug report. (The bisection that confirmed the breakage point took so long that RL intervened and I got delayed in filing.) However, as correspondence earlier today on list indicates, we still have a big problem on our hands. Craig Berry argues that Time-HiRes 1.9758 merely puts a bandaid on the problem: Jarkko Hietaniemi acknowledges the problem, yields maintenance of Time-HiRes and recommends reversion to 1.9754: But that was the commit that broke the two CPAN distros, so we would still have to negotiate correction of the code in their tests with their maintainers. Since we still have work to do before 5.28.0 with respect to this problem, this ticket should remain open. Thank you very much. |
From @jkeenanOn Fri, 23 Mar 2018 04:39:07 GMT, TODDR wrote:
As of commit 3a7f897, I am still getting those test errors in both CPAN distributions. Example: ##### # Failed test 'elapsed' Test Summary Report t/unit_stats.t (Wstat: 256 Tests: 13 Failed: 1) Thank you very much. -- |
From @toddrOn Fri, 23 Mar 2018 05:39:45 -0700, jkeenan wrote:
Yeah he later corrected his suggestion to 1.9753, which is the version we were at before I broke everything. In accordance with his request, I've reverted both of my commits which caused this problem to begin with and then bumped Maintainers.pl so it'd be clear we still need to do a release to CPAN. I also updated the Changelog documenting the decision. commit e771953 Bump Time-HiRes to 1.9759 to revert changes from 1.9754..1.9758 on CPAN Porting/Maintainers.pl | 2 +- commit 727404d Revert "Upgrade Time::HiRes from version 1.9753 to 1.9757" Porting/Maintainers.pl | 2 +- commit 1eb901c Revert "Upgrade Time::HiRes from version 1.9757 to 1.9758" Porting/Maintainers.pl | 2 +- Short of releasing this to CPAN, I think we're done? |
From @craigberryOn Fri, Mar 23, 2018 at 11:59 AM, Todd Rinaldo via RT
Rolling back for 5.28 might be ok, but there were some fixes in there From dfbb2d9bc7c2d1d678db0078f17a6e9a521122e8 Mon Sep 17 00:00:00 2001 Rather than platform-specific workarounds. This is actually dist/Time-HiRes/HiRes.xs | 2 +- Inline Patchdiff --git a/dist/Time-HiRes/HiRes.xs b/dist/Time-HiRes/HiRes.xs
index 39ac3fd8f0..f91fc2f4b1 100644
--- a/dist/Time-HiRes/HiRes.xs
+++ b/dist/Time-HiRes/HiRes.xs
@@ -40,7 +40,7 @@ extern "C" {
}
#endif
-#ifdef _MSC_VER /* For Visual C */
+#ifndef TIME_HIRES_SYS_TIME_HAS_TIMEZONE
struct timezone
{
int tz_minuteswest; /* of Greenwich */
diff --git a/dist/Time-HiRes/Makefile.PL b/dist/Time-HiRes/Makefile.PL
index 8036fbf5cb..17c69eb14c 100644
--- a/dist/Time-HiRes/Makefile.PL
+++ b/dist/Time-HiRes/Makefile.PL
@@ -880,6 +880,26 @@ EOM
}
}
+ print "Trying struct timezone from sys/time.h ...";
+ my $sys_time_has_timezone = 0;
+ if (try_compile_and_link(<<EOM)) {
+#include <sys/time.h>
+
+int main(int argc, char** argv)
+{
+ struct timezone Tz;
+}
+EOM
+ $sys_time_has_timezone++;
+ DEFINE('TIME_HIRES_SYS_TIME_HAS_TIMEZONE'); #6
+ }
+
+ if ($sys_time_has_timezone) {
+ print "found.\n";
+ } else {
+ print "NOT found.\n";
+ }
+
if ($DEFINE) {
$DEFINE =~ s/^\s+//;
if (open(XDEFINE, '>', 'xdefine')) {
--
2.14.3 (Apple Git-98) |
From @craigberry0001-Make-Time-HiRes-check-for-timezone-struct-in-sys-tim.patchFrom dfbb2d9bc7c2d1d678db0078f17a6e9a521122e8 Mon Sep 17 00:00:00 2001
From: "Craig A. Berry" <craigberry@mac.com>
Date: Fri, 23 Mar 2018 13:48:42 -0500
Subject: [PATCH] Make Time::HiRes check for timezone struct in sys/time.h.
Rather than platform-specific workarounds. This is actually
pretty easy to do since Time::HiRes has great infrastructure for
testing local compiler/header capabilities.
---
dist/Time-HiRes/HiRes.xs | 2 +-
dist/Time-HiRes/Makefile.PL | 20 ++++++++++++++++++++
2 files changed, 21 insertions(+), 1 deletion(-)
diff --git a/dist/Time-HiRes/HiRes.xs b/dist/Time-HiRes/HiRes.xs
index 39ac3fd8f0..f91fc2f4b1 100644
--- a/dist/Time-HiRes/HiRes.xs
+++ b/dist/Time-HiRes/HiRes.xs
@@ -40,7 +40,7 @@ extern "C" {
}
#endif
-#ifdef _MSC_VER /* For Visual C */
+#ifndef TIME_HIRES_SYS_TIME_HAS_TIMEZONE
struct timezone
{
int tz_minuteswest; /* of Greenwich */
diff --git a/dist/Time-HiRes/Makefile.PL b/dist/Time-HiRes/Makefile.PL
index 8036fbf5cb..17c69eb14c 100644
--- a/dist/Time-HiRes/Makefile.PL
+++ b/dist/Time-HiRes/Makefile.PL
@@ -880,6 +880,26 @@ EOM
}
}
+ print "Trying struct timezone from sys/time.h ...";
+ my $sys_time_has_timezone = 0;
+ if (try_compile_and_link(<<EOM)) {
+#include <sys/time.h>
+
+int main(int argc, char** argv)
+{
+ struct timezone Tz;
+}
+EOM
+ $sys_time_has_timezone++;
+ DEFINE('TIME_HIRES_SYS_TIME_HAS_TIMEZONE'); #6
+ }
+
+ if ($sys_time_has_timezone) {
+ print "found.\n";
+ } else {
+ print "NOT found.\n";
+ }
+
if ($DEFINE) {
$DEFINE =~ s/^\s+//;
if (open(XDEFINE, '>', 'xdefine')) {
--
2.14.3 (Apple Git-98)
|
From zefram@fysh.orgJames E Keenan wrote:
The tests that got broken are at fault, and need to switch to a more -zefram |
From @dur-randirSince the last non-core maintainer of this module has stepped down, shouldn't something like this attachment be applied, so no-one would report more bug to CPAN rt? |
From @dur-randir0001-Update-rt-path-for-Time-HiRes.patchFrom 35f8345d918e73c070d463c85395a1531b0d8219 Mon Sep 17 00:00:00 2001
From: Sergey Aleynikov <sergey.aleynikov@gmail.com>
Date: Mon, 26 Mar 2018 01:49:46 +0300
Subject: [PATCH] Update rt path for Time::HiRes
---
dist/Time-HiRes/Makefile.PL | 10 ++++++++++
1 file changed, 10 insertions(+)
diff --git a/dist/Time-HiRes/Makefile.PL b/dist/Time-HiRes/Makefile.PL
index e0f7dd9a2f..7b812edc60 100644
--- a/dist/Time-HiRes/Makefile.PL
+++ b/dist/Time-HiRes/Makefile.PL
@@ -961,6 +961,16 @@ sub doMakefile {
push @makefileopts, (LICENSE => 'perl_5');
}
+ if ($ExtUtils::MakeMaker::VERSION >= 6.46) {
+ push @makefileopts, (
+ resources => {
+ repository => 'git://perl5.git.perl.org/perl.git',
+ bugtracker => 'https://rt.perl.org/rt3/',
+ homepage => "http://dev.perl.org/",
+ },
+ )
+ }
+
WriteMakefile(@makefileopts);
}
--
2.14.2
|
From @xsawyerxOn 03/26/2018 02:06 AM, Sergey Aleynikov via RT wrote:
+1 |
From @iabynOn Sun, Mar 25, 2018 at 04:06:33PM -0700, Sergey Aleynikov via RT wrote:
This sounds like a good idea, but I don't feel I'm in a position to judge Should we consider applying this or similar as a 5.28 blocker? -- |
From @xsawyerxOn 04/20/2018 12:34 PM, Dave Mitchell wrote:
I think this is correct.
This is metadata. It shouldn't block 5.28.
Agreed. |
From @haargOn Mon, Mar 26, 2018 at 1:06 AM, Sergey Aleynikov via RT
The resources section needs to be in a META_ADD or META_MERGE section, Also, I have fixed the Catalyst-Runtime test failure in git, and |
From @haargOn Fri, Apr 20, 2018 at 3:22 PM, Graham Knop <haarg@haarg.org> wrote:
And apparently I didn't attach the patch. |
From @haarg0001-refer-Time-HiRes-RT-queue-to-perl-core.patchFrom 4b95a46fab70873f0aab2517e4a79fc2b6de3de2 Mon Sep 17 00:00:00 2001
From: Graham Knop <haarg@haarg.org>
Date: Fri, 20 Apr 2018 15:17:58 +0200
Subject: [PATCH] refer Time::HiRes RT queue to perl core
---
dist/Time-HiRes/Makefile.PL | 12 ++++++++++++
1 file changed, 12 insertions(+)
mode change 100644 => 100755 dist/Time-HiRes/Makefile.PL
diff --git a/dist/Time-HiRes/Makefile.PL b/dist/Time-HiRes/Makefile.PL
old mode 100644
new mode 100755
index e0f7dd9a2f..7529588532
--- a/dist/Time-HiRes/Makefile.PL
+++ b/dist/Time-HiRes/Makefile.PL
@@ -961,6 +961,18 @@ sub doMakefile {
push @makefileopts, (LICENSE => 'perl_5');
}
+ if ($ExtUtils::MakeMaker::VERSION >= 6.46) {
+ push @makefileopts, (
+ META_MERGE => {
+ resources => {
+ repository => 'git://perl5.git.perl.org/perl.git',
+ bugtracker => 'https://rt.perl.org/rt3/',
+ homepage => "http://dev.perl.org/",
+ },
+ },
+ )
+ }
+
WriteMakefile(@makefileopts);
}
--
2.17.0
|
From @xsawyerxOn 04/20/2018 10:51 PM, Graham Knop wrote:
+1 to the patch. |
From @iabynOn Sat, Apr 21, 2018 at 10:22:42AM +0200, Sawyer X wrote:
Do want this applying for 5.28, or shall I remove this ticket from the -- |
From @xsawyerxOn 04/23/2018 02:43 PM, Dave Mitchell wrote:
This patch should not be blocking 5.28. I also don't think it should be applied at this moment but I'm open for |
From @khwilliamsonI added this as a 5.30 blocker, so that it doesn't slip through the cracks |
From [Unknown Contact. See original ticket]I added this as a 5.30 blocker, so that it doesn't slip through the cracks |
From @jkeenanOn Sat, 09 Mar 2019 16:49:24 GMT, khw wrote:
How are we going to deal with this ticket re perl-5.30.0? Thank you very much. |
From @iabynOn Sat, Apr 13, 2019 at 07:06:08PM -0700, James E Keenan via RT wrote:
For context, this is a patch that updates T::HiRes's Makefile.PL (And next time round we should all have a look at the blockers list when -- |
From @khwilliamsonOn 4/23/19 9:19 AM, Dave Mitchell wrote:
I don't like the idea of keep kicking the can down the road. How about |
From @iabynOn Wed, Apr 24, 2019 at 12:12:23PM -0600, Karl Williamson wrote:
I suspect we'd instead just forget the existence of that new list until So I think the trick is to just remember that the 5.32.0 list exists. -- |
From @xsawyerxOn 4/25/19 12:43 PM, Dave Mitchell wrote:
Kicking it down the line is not the best practice, but in this case, I think this shows we can improve in how we prioritize tickets. What I |
From @dur-randirHas the time finally come? |
From @khwilliamsonYes, Thanks for the reminder |
|
@khwilliamson - Status changed from 'open' to 'pending release' |
p5pRT
added
BBC
Migrated from rt.perl.org#133014 (status was 'pending release')
Searchable as RT133014$
The text was updated successfully, but these errors were encountered: