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
NetBSD 7.1 Smoke Test Failures #16078
Comments
From carlos@carlosguevara.comAttached file netbsd.failures.txt contains the results of the following on Attached file perl-Ilib-V.txt contains the results of './perl -Ilib -V' in Please let me know if I can be of any further assistance. |
From carlos@carlosguevara.com../cpan/Sys-Syslog/t/syslog.t .. # Failed (TODO) test 'setlogsock() should return true: '0'' Test Summary Report ../lib/locale.t (Wstat: 0 Tests: 699 Failed: 1) |
From carlos@carlosguevara.comSummary of my perl5 (revision 5 version 27 subversion 2) configuration: Characteristics of this binary (from libperl): |
From @jkeenanOn Tue, 11 Jul 2017 06:31:33 GMT, carlos@carlosguevara.com wrote:
Thanks for preparing that report. In gratitude, we ask you to prepare more reports! The failures in ../cpan/Sys-Syslog/t/syslog.t are not consistent from one smoke run to the next. However, they are like those being newly reported for DragonflyBSD as well. Let's set those aside for now. The output from lib/locale.t suggests how to proceed. In the same directory, could you run the following command and include the output as an attachment in responding to this message? ##### Thank you very much. -- |
The RT System itself - Status changed from 'new' to 'open' |
From carlos@carlosguevara.comNo problem. Report attached. On Wed, Jul 12, 2017 at 7:59 AM, James E Keenan via RT <
|
From @jkeenanOn Thu, 13 Jul 2017 13:27:35 GMT, carlos@carlosguevara.com wrote:
Thanks. Last year Karl Williamson used output like this to diagnose and correct failures in lib/locale.t on FreeBSD. I'm hoping he'll be able to do likewise on NetBSD. If you can continue to submit smoke test reports for Perl 5 blead on NetBSD 7 -- say, weekly or semi-weekly -- that would be helpful. Thank you very much. |
From @khwilliamsonOn 07/12/2017 11:54 PM, Carlos Guevara wrote:
This output indicates problems with the NetBSD locale definitions, Other platforms also have various defective locales, which we can tell lib/locale.t is written to look for anomalies in the locale definitions. You can see the failures by grepping for 'failed' in the output. They # failed 431 (Verify that isn't both [:graph:] and [:space:]) with 0xA0 is the No-break space in ISO 8859-2. Posix stipulates that # failed 398 (Verify that /[[:upper:]]/ matches all alpha X for which BD and BF are upper and lowercase "Ŋ" (Eng) characters in 8859-4. The # failed 407 (Verify that [:lower:] is a subset of [:alpha:]) with DF is the German "ß" alphabetic letter. This is a defect in the locale # failed 426 (Verify that the space character is in [:print:]) with This is the one I've never seen before. I was unfamiliar with this |
From @khwilliamsonOn 07/13/2017 12:59 PM, Karl Williamson wrote:
But there are indications in the output you gave us that there are bugs |
From @khwilliamson |
From carlos@carlosguevara.comNo problem. Reports attached. On Fri, Jul 14, 2017 at 12:19 PM, Karl Williamson <public@khwilliamson.com>
|
From carlos@carlosguevara.comSummary of my perl5 (revision 5 version 27 subversion 2) configuration: Characteristics of this binary (from libperl): |
From @khwilliamsonOn 07/16/2017 10:58 AM, Carlos Guevara wrote:
Thanks. That showed me that the previous version of lib/locale.t had a bug and Fixing that also correctly showed why netbsd is failing this test. 8% Because many systems have similar defective locales, an error isn't (Note that your 8859-1 locales don't have the same problem, and yet they Some of your other locales also are being found to have defects by this I'll think about the appropriate way to change the test to not cause But these are real defects that I believe should be reported to the OS
|
From @khwilliamsonOn 07/17/2017 02:43 PM, Karl Williamson wrote:
After a couple more minutes of thought (less than I thought would be |
From @khwilliamson |
From @jkeenanOn Mon, 17 Jul 2017 20:46:07 GMT, public@khwilliamson.com wrote:
Karl, thanks for that explanation as to how lib/locale.t works. We -- P5P collectively -- need to come up with a Standard Operating Procedure for informing OSes of locale-related problems -- or, more generally, of problems which some OSes appear to have solved better than others. A smoke-tester who reports failures like these -- as Carlos Guevara has done at my request -- is not necessarily an expert in either locales or the operating system being reported on. So we as an open-source software project ought to figure out a way to communicate these problems with other such projects. Thank you very much. |
From carlos@carlosguevara.comNo problem. Reports attached. On Mon, Jul 17, 2017 at 4:40 PM, Karl Williamson <public@khwilliamson.com>
|
From carlos@carlosguevara.comSummary of my perl5 (revision 5 version 27 subversion 2) configuration: Characteristics of this binary (from libperl): |
From @jkeenanOn Mon, 17 Jul 2017 20:46:07 GMT, public@khwilliamson.com wrote:
This page describes how to submit a bug report to the NetBSD project: ##### I would be willing to be the submitter of such a bug report, but I would need: 1. From Carlos: a. The NetBSD release where you encountered the problem; b. Output of 'uname -a' on the smoke-testing rig. 2. From khw: a. Text for the "Full Description" box on that bug reporting page.* b. Text for the "How to repeat the problem" and "Fix to the problem if known" boxes on that page, if appropriate. *Karl, is there is some boilerplate text we could use in submitting such locale-related bug reports? If so, that would facilitate preparation of such reports in the future. So, if you all can supply the requested information, I will submit a report to NetBSD. Thank you very much. -- |
From carlos@carlosguevara.comOn Thu, Jul 20, 2017 at 8:13 AM, James E Keenan via RT <
NetBSD Release: uname -a: |
From @khwilliamsonOn 07/18/2017 09:37 PM, Carlos Guevara wrote:
That didn't fix it. I think it's a casing problem, where I was
|
From @khwilliamson |
From carlos@carlosguevara.comOn Sun, Jul 23, 2017 at 12:00 AM, karl williamson via RT <
Reports attached. |
From carlos@carlosguevara.comSummary of my perl5 (revision 5 version 27 subversion 3) configuration: Characteristics of this binary (from libperl): |
From @khwilliamsonOn 07/23/2017 09:41 PM, Carlos Guevara wrote:
When I finally got around to looking at why my file didn't fix the |
From @khwilliamson |
From carlos@carlosguevara.comOn Sun, Jul 30, 2017 at 11:41 PM, karl williamson via RT <
Reports attached. |
From carlos@carlosguevara.comSummary of my perl5 (revision 5 version 27 subversion 3) configuration: Characteristics of this binary (from libperl): |
From @khwilliamsonOn 07/22/2017 03:49 PM, Carlos Guevara wrote:
The changes to locale.t are now in blead. This should fix this portion
|
From @jkeenanOn Mon, 07 Aug 2017 17:45:42 GMT, public@khwilliamson.com wrote:
The only other smoke testing failure being reported on this platform (http://perl.develop-help.com/raw/?id=200535) looks like this: ##### I confess that I never quite know how to react to cases where t/TEST and t/harness come up with different results. I only smoke with t/harness because the VMs in which I smoke are situated on my everyday laptop. And I only have the energy to focus on tests which are failing very consistently. That's even more true on platforms like this one where we're lucky to have one or two people submitting smoke reports -- much less someone who can actually diagnose and correct problems. So, I'm going to mark this ticket Resolved because we (P5P contributors) have accomplished as much as we reasonably can. (Send me some appropriate language and I'll open up a bug ticket with NetBSD about the locales.) Thank you very much. -- |
@jkeenan - Status changed from 'open' to 'resolved' |
From @iabynOn Mon, Aug 07, 2017 at 07:27:32PM -0700, James E Keenan via RT wrote:
There are two common reasons for a difference. First the failure may be -- |
Migrated from rt.perl.org#131735 (status was 'resolved')
Searchable as RT131735$
The text was updated successfully, but these errors were encountered: