You are viewing a plain text version of this content. The canonical link for it is here.
Posted to sysadmins@spamassassin.apache.org by Paul Stead <pa...@gmail.com> on 2019/05/01 09:59:07 UTC
/tags nightly commits
Howdy!
As per -
https://lists.apache.org/thread.html/984ee3c26163e0281727256eabf0f29079a3fc053664416f58581ba0@%3Csysadmins.spamassassin.apache.org%3E
It looks as though the /tags commit is broken too -
http://svn.apache.org/viewvc/spamassassin/tags/?view=log
---8<---
t/basic_lint_without_sandbox.t ..
okhttps://lists.apache.org/thread.html/984ee3c26163e0281727256eabf0f29079a3fc053664416f58581ba0@%3Csysadmins.spamassassin.apache.org%3E
depends on __DKIMWL_WL_HI which is nonexistent
FSL_BULK_SIG depends on __DKIMWL_WL_HI which is nonexistent
# Failed test at t/basic_meta.t line 93.
# Looks like you failed 1 test of 2.
t/basic_meta.t ..................
Dubious, test returned 1 (wstat 256, 0x100)
Failed 1/2 subtests
---8<---
The sub rule does exist, but FSL_BULK_SIG doesn't depend on
__DKIMWL_WL_HI any more.
It looks as though the SVN that is being worked with isn't up to date
for some reason.
Paul
Re: /tags nightly commits
Posted by David Jones <dj...@ena.com>.
On 5/2/19 3:44 AM, Paul Stead wrote:
> On Wed, 1 May 2019 at 17:43, Bill Cole
> <sa-bugz-20080315@billmail.scconsult.com
> <ma...@billmail.scconsult.com>> wrote:
>
> ...
>
> The resulting generate-new-scores.sh was a mangled mess, with r1815390
> being the committed version in trunk. I've fixed this up.
>
> I've also committed a change made locally on sa-vm1 to
> build/automc/apache2-le-ssl.conf.
>
> Apparently part of the problem causing this diversity of svn breakage
> was work done by root in ~automc/svn/*/, causing various checked-out
> files and SVN nodes to become owned by root, so that the various 'svn
> update' jobs run as automc (from /etc/cron.d/automc) had random
> breakage.
>
> All should be well now.
>
>
> Looks like there might still be an issue today :(
>
> https://lists.apache.org/thread.html/0c694eda1a9c73cf77b13e69130cc79aa4c1606cc2a5f8437d0fe892@%3Csysadmins.spamassassin.apache.org%3E
>
> Looks like build/mkrules is still working with an old version
>
I recall that depending on timing, this can take 30 to 48 hours to
happen for sure. When I was debugging, it was very slow and painful to
work through issues.
> ---8<---
> Test Summary Report
> -------------------
> t/basic_meta.t (Wstat: 256 Tests: 2 Failed: 1)
> Failed test: 2
> Non-zero exit status: 1
> Files=3, Tests=6, 11 wallclock secs ( 0.06 usr 0.00 sys + 10.16 cusr
> 0.71 csys = 10.93 CPU)
> Result: FAIL
> make: *** [test_dynamic] Error 1
> Makefile:1588: recipe for target 'test_dynamic' failed
> + exit 2
> ---8<---
--
David Jones
Re: /tags nightly commits
Posted by "Kevin A. McGrail" <km...@apache.org>.
On 5/5/2019 10:34 AM, Mark Thomas wrote:
> E_WRONG_LIST
>
Thanks Mark and apologies to comdev for the noise!
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
For additional commands, e-mail: dev-help@community.apache.org
Re: /tags nightly commits
Posted by Mark Thomas <ma...@apache.org>.
E_WRONG_LIST
Mark
On 05/05/2019 15:18, Kevin A. McGrail wrote:
> On 5/5/2019 6:10 AM, Paul Stead wrote:
>> Promotions and tags still seem to be a little problematic at the moment. A
>> run of build/mkupdates/listpromotable locally shows +300 lines difference
>> to active.list over the one being submitted.
>
> Before I deal with this issue:
>
> +Dev.
>
> TL;DR: REMINDER: PLEASE use a make test on your local checkout before
> committing!
>
> We had two issues. It appears an svn commit failed on a release and
> that left cruft that was colliding for commit. That issue was caused by
> a manual edit to active.list:
>
> Revision *1854477*
> <http://svn.apache.org/viewvc?view=revision&revision=1854477> - (view
> <http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=markup>)
> (download
> <http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=co>)
> (annotate
> <http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?annotate=1854477>)
> - [select for diffs]
> <http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log&r1=1854477>
>
> Modified /Wed Feb 27 18:16:20 2019 UTC/ (2 months ago) by /gbechis/
> File length: 15052 byte(s)
> Diff to previous 1854447
> <http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477>
> (colored
> <http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477&diff_format=h>)
>
>
> Switch to https and fix some 404 errors
> bz #7652
>
> The second issue is a continuing one and that is committers submitting
> without 'make test' first. This occurred in January when I tried to
> build the last release. If the svn for trunk including rules is broken,
> it will break masscheck, etc. We are currently passing again so just a
> friendly reminder. It takes a village to stop a spammer :-)
>
>
> So now, what I see in the logs is that the commit didn't occur but
> there is no error, nothing. Checking
> http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log
> concurs:
>
> 5/5 (note that Committing promotions is NOT followed by an svn commit)
>
> + echo 'Committing promotions in rules/active.list...'
> + svn commit -m 'promotions validated' rules/active.list
> Committing promotions in rules/active.list...
> + /usr/bin/perl masses/rule-qa/list-bad-rules
>
> 5/4 (example where the commit worked)
> + echo 'Committing promotions in rules/active.list...'
> Committing promotions in rules/active.list...
> + svn commit -m 'promotions validated' rules/active.list
> Sending rules/active.list
> Transmitting file data .done
> Committing transaction...
> Committed revision 1858595.
>
>
> I'm out of time to keep pursuing this but hoping someone else can look
> at it.
>
> Regards,
>
> KAM
>
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@community.apache.org
For additional commands, e-mail: dev-help@community.apache.org
Re: /tags nightly commits
Posted by "Kevin A. McGrail" <km...@apache.org>.
Resending with correct dev list :-)
On 5/5/2019 6:10 AM, Paul Stead wrote:
> Promotions and tags still seem to be a little problematic at the moment. A
> run of build/mkupdates/listpromotable locally shows +300 lines difference
> to active.list over the one being submitted.
Before I deal with this issue:
+Dev.
TL;DR: REMINDER: PLEASE use a make test on your local checkout before
committing!
We had two issues. It appears an svn commit failed on a release and
that left cruft that was colliding for commit. That issue was caused by
a manual edit to active.list:
Revision *1854477*
<http://svn.apache.org/viewvc?view=revision&revision=1854477> - (view
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=markup>)
(download
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=co>)
(annotate
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?annotate=1854477>)
- [select for diffs]
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log&r1=1854477>
Modified /Wed Feb 27 18:16:20 2019 UTC/ (2 months ago) by /gbechis/
File length: 15052 byte(s)
Diff to previous 1854447
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477>
(colored
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477&diff_format=h>)
Switch to https and fix some 404 errors
bz #7652
The second issue is a continuing one and that is committers submitting
without 'make test' first. This occurred in January when I tried to
build the last release. If the svn for trunk including rules is broken,
it will break masscheck, etc. We are currently passing again so just a
friendly reminder. It takes a village to stop a spammer :-)
So now, what I see in the logs is that the commit didn't occur but
there is no error, nothing. Checking
http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log
concurs:
5/5 (note that Committing promotions is NOT followed by an svn commit)
+ echo 'Committing promotions in rules/active.list...'
+ svn commit -m 'promotions validated' rules/active.list
Committing promotions in rules/active.list...
+ /usr/bin/perl masses/rule-qa/list-bad-rules
5/4 (example where the commit worked)
+ echo 'Committing promotions in rules/active.list...'
Committing promotions in rules/active.list...
+ svn commit -m 'promotions validated' rules/active.list
Sending rules/active.list
Transmitting file data .done
Committing transaction...
Committed revision 1858595.
I'm out of time to keep pursuing this but hoping someone else can look
at it.
Regards,
KAM
Re: /tags nightly commits
Posted by "Kevin A. McGrail" <km...@apache.org>.
Resending with correct dev list :-)
On 5/5/2019 6:10 AM, Paul Stead wrote:
> Promotions and tags still seem to be a little problematic at the moment. A
> run of build/mkupdates/listpromotable locally shows +300 lines difference
> to active.list over the one being submitted.
Before I deal with this issue:
+Dev.
TL;DR: REMINDER: PLEASE use a make test on your local checkout before
committing!
We had two issues. It appears an svn commit failed on a release and
that left cruft that was colliding for commit. That issue was caused by
a manual edit to active.list:
Revision *1854477*
<http://svn.apache.org/viewvc?view=revision&revision=1854477> - (view
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=markup>)
(download
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=co>)
(annotate
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?annotate=1854477>)
- [select for diffs]
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log&r1=1854477>
Modified /Wed Feb 27 18:16:20 2019 UTC/ (2 months ago) by /gbechis/
File length: 15052 byte(s)
Diff to previous 1854447
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477>
(colored
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477&diff_format=h>)
Switch to https and fix some 404 errors
bz #7652
The second issue is a continuing one and that is committers submitting
without 'make test' first. This occurred in January when I tried to
build the last release. If the svn for trunk including rules is broken,
it will break masscheck, etc. We are currently passing again so just a
friendly reminder. It takes a village to stop a spammer :-)
So now, what I see in the logs is that the commit didn't occur but
there is no error, nothing. Checking
http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log
concurs:
5/5 (note that Committing promotions is NOT followed by an svn commit)
+ echo 'Committing promotions in rules/active.list...'
+ svn commit -m 'promotions validated' rules/active.list
Committing promotions in rules/active.list...
+ /usr/bin/perl masses/rule-qa/list-bad-rules
5/4 (example where the commit worked)
+ echo 'Committing promotions in rules/active.list...'
Committing promotions in rules/active.list...
+ svn commit -m 'promotions validated' rules/active.list
Sending rules/active.list
Transmitting file data .done
Committing transaction...
Committed revision 1858595.
I'm out of time to keep pursuing this but hoping someone else can look
at it.
Regards,
KAM
Re: /tags nightly commits
Posted by "Kevin A. McGrail" <km...@apache.org>.
Resending with correct dev list :-)
On 5/5/2019 6:10 AM, Paul Stead wrote:
> Promotions and tags still seem to be a little problematic at the moment. A
> run of build/mkupdates/listpromotable locally shows +300 lines difference
> to active.list over the one being submitted.
Before I deal with this issue:
+Dev.
TL;DR: REMINDER: PLEASE use a make test on your local checkout before
committing!
We had two issues. It appears an svn commit failed on a release and
that left cruft that was colliding for commit. That issue was caused by
a manual edit to active.list:
Revision *1854477*
<http://svn.apache.org/viewvc?view=revision&revision=1854477> - (view
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=markup>)
(download
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=co>)
(annotate
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?annotate=1854477>)
- [select for diffs]
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log&r1=1854477>
Modified /Wed Feb 27 18:16:20 2019 UTC/ (2 months ago) by /gbechis/
File length: 15052 byte(s)
Diff to previous 1854447
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477>
(colored
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477&diff_format=h>)
Switch to https and fix some 404 errors
bz #7652
The second issue is a continuing one and that is committers submitting
without 'make test' first. This occurred in January when I tried to
build the last release. If the svn for trunk including rules is broken,
it will break masscheck, etc. We are currently passing again so just a
friendly reminder. It takes a village to stop a spammer :-)
So now, what I see in the logs is that the commit didn't occur but
there is no error, nothing. Checking
http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log
concurs:
5/5 (note that Committing promotions is NOT followed by an svn commit)
+ echo 'Committing promotions in rules/active.list...'
+ svn commit -m 'promotions validated' rules/active.list
Committing promotions in rules/active.list...
+ /usr/bin/perl masses/rule-qa/list-bad-rules
5/4 (example where the commit worked)
+ echo 'Committing promotions in rules/active.list...'
Committing promotions in rules/active.list...
+ svn commit -m 'promotions validated' rules/active.list
Sending rules/active.list
Transmitting file data .done
Committing transaction...
Committed revision 1858595.
I'm out of time to keep pursuing this but hoping someone else can look
at it.
Regards,
KAM
Re: /tags nightly commits
Posted by "Kevin A. McGrail" <km...@apache.org>.
On 5/5/2019 6:10 AM, Paul Stead wrote:
> Promotions and tags still seem to be a little problematic at the moment. A
> run of build/mkupdates/listpromotable locally shows +300 lines difference
> to active.list over the one being submitted.
Before I deal with this issue:
+Dev.
TL;DR: REMINDER: PLEASE use a make test on your local checkout before
committing!
We had two issues. It appears an svn commit failed on a release and
that left cruft that was colliding for commit. That issue was caused by
a manual edit to active.list:
Revision *1854477*
<http://svn.apache.org/viewvc?view=revision&revision=1854477> - (view
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=markup>)
(download
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=co>)
(annotate
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?annotate=1854477>)
- [select for diffs]
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log&r1=1854477>
Modified /Wed Feb 27 18:16:20 2019 UTC/ (2 months ago) by /gbechis/
File length: 15052 byte(s)
Diff to previous 1854447
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477>
(colored
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477&diff_format=h>)
Switch to https and fix some 404 errors
bz #7652
The second issue is a continuing one and that is committers submitting
without 'make test' first. This occurred in January when I tried to
build the last release. If the svn for trunk including rules is broken,
it will break masscheck, etc. We are currently passing again so just a
friendly reminder. It takes a village to stop a spammer :-)
So now, what I see in the logs is that the commit didn't occur but
there is no error, nothing. Checking
http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log
concurs:
5/5 (note that Committing promotions is NOT followed by an svn commit)
+ echo 'Committing promotions in rules/active.list...'
+ svn commit -m 'promotions validated' rules/active.list
Committing promotions in rules/active.list...
+ /usr/bin/perl masses/rule-qa/list-bad-rules
5/4 (example where the commit worked)
+ echo 'Committing promotions in rules/active.list...'
Committing promotions in rules/active.list...
+ svn commit -m 'promotions validated' rules/active.list
Sending rules/active.list
Transmitting file data .done
Committing transaction...
Committed revision 1858595.
I'm out of time to keep pursuing this but hoping someone else can look
at it.
Regards,
KAM
Re: /tags nightly commits
Posted by "Kevin A. McGrail" <km...@apache.org>.
On 5/5/2019 6:10 AM, Paul Stead wrote:
> Promotions and tags still seem to be a little problematic at the moment. A
> run of build/mkupdates/listpromotable locally shows +300 lines difference
> to active.list over the one being submitted.
Before I deal with this issue:
+Dev.
TL;DR: REMINDER: PLEASE use a make test on your local checkout before
committing!
We had two issues. It appears an svn commit failed on a release and
that left cruft that was colliding for commit. That issue was caused by
a manual edit to active.list:
Revision *1854477*
<http://svn.apache.org/viewvc?view=revision&revision=1854477> - (view
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=markup>)
(download
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=co>)
(annotate
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?annotate=1854477>)
- [select for diffs]
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log&r1=1854477>
Modified /Wed Feb 27 18:16:20 2019 UTC/ (2 months ago) by /gbechis/
File length: 15052 byte(s)
Diff to previous 1854447
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477>
(colored
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477&diff_format=h>)
Switch to https and fix some 404 errors
bz #7652
The second issue is a continuing one and that is committers submitting
without 'make test' first. This occurred in January when I tried to
build the last release. If the svn for trunk including rules is broken,
it will break masscheck, etc. We are currently passing again so just a
friendly reminder. It takes a village to stop a spammer :-)
So now, what I see in the logs is that the commit didn't occur but
there is no error, nothing. Checking
http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log
concurs:
5/5 (note that Committing promotions is NOT followed by an svn commit)
+ echo 'Committing promotions in rules/active.list...'
+ svn commit -m 'promotions validated' rules/active.list
Committing promotions in rules/active.list...
+ /usr/bin/perl masses/rule-qa/list-bad-rules
5/4 (example where the commit worked)
+ echo 'Committing promotions in rules/active.list...'
Committing promotions in rules/active.list...
+ svn commit -m 'promotions validated' rules/active.list
Sending rules/active.list
Transmitting file data .done
Committing transaction...
Committed revision 1858595.
I'm out of time to keep pursuing this but hoping someone else can look
at it.
Regards,
KAM
Re: /tags nightly commits
Posted by "Kevin A. McGrail" <km...@apache.org>.
On 5/5/2019 6:10 AM, Paul Stead wrote:
> Promotions and tags still seem to be a little problematic at the moment. A
> run of build/mkupdates/listpromotable locally shows +300 lines difference
> to active.list over the one being submitted.
Before I deal with this issue:
+Dev.
TL;DR: REMINDER: PLEASE use a make test on your local checkout before
committing!
We had two issues. It appears an svn commit failed on a release and
that left cruft that was colliding for commit. That issue was caused by
a manual edit to active.list:
Revision *1854477*
<http://svn.apache.org/viewvc?view=revision&revision=1854477> - (view
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=markup>)
(download
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?revision=1854477&view=co>)
(annotate
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?annotate=1854477>)
- [select for diffs]
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log&r1=1854477>
Modified /Wed Feb 27 18:16:20 2019 UTC/ (2 months ago) by /gbechis/
File length: 15052 byte(s)
Diff to previous 1854447
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477>
(colored
<http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?r1=1854447&r2=1854477&diff_format=h>)
Switch to https and fix some 404 errors
bz #7652
The second issue is a continuing one and that is committers submitting
without 'make test' first. This occurred in January when I tried to
build the last release. If the svn for trunk including rules is broken,
it will break masscheck, etc. We are currently passing again so just a
friendly reminder. It takes a village to stop a spammer :-)
So now, what I see in the logs is that the commit didn't occur but
there is no error, nothing. Checking
http://svn.apache.org/viewvc/spamassassin/trunk/rules/active.list?view=log
concurs:
5/5 (note that Committing promotions is NOT followed by an svn commit)
+ echo 'Committing promotions in rules/active.list...'
+ svn commit -m 'promotions validated' rules/active.list
Committing promotions in rules/active.list...
+ /usr/bin/perl masses/rule-qa/list-bad-rules
5/4 (example where the commit worked)
+ echo 'Committing promotions in rules/active.list...'
Committing promotions in rules/active.list...
+ svn commit -m 'promotions validated' rules/active.list
Sending rules/active.list
Transmitting file data .done
Committing transaction...
Committed revision 1858595.
I'm out of time to keep pursuing this but hoping someone else can look
at it.
Regards,
KAM
Re: /tags nightly commits
Posted by Paul Stead <pa...@gmail.com>.
Promotions and tags still seem to be a little problematic at the moment. A
run of build/mkupdates/listpromotable locally shows +300 lines difference
to active.list over the one being submitted.
https://lists.apache.org/thread.html/ceb563e7c5a6b83dda106d24e0d234e7c76cd5286ae898661b7a49c7@%3Csysadmins.spamassassin.apache.org%3E
Looks like parts of the build are still failing. Seems quite a painful part
of the setup - it looks that both build and ruleqa are working from the
same svn - is this causing conflicts? I'm commenting from a hidden view as
I don't have any actual access to the sa-vm1, sorry if assumptions are
wrong.
I guess you're also looking at these, I want to make sure people know :)
Paul
Re: /tags nightly commits
Posted by Bill Cole <sa...@billmail.scconsult.com>.
On 3 May 2019, at 2:05, Paul Stead wrote:
> The interesting part which was showing yesterday was
>
> ---8<---
> t/basic_lint.t .................. ok
> t/basic_lint_without_sandbox.t .. ok
> FSL_BULK_SIG depends on __DKIMWL_WL_HI which is nonexistent
> FSL_BULK_SIG depends on __DKIMWL_WL_HI which is nonexistent
>
> # Failed test at t/basic_meta.t line 93.
> # Looks like you failed 1 test of 2.
> ---8<---
>
> https://lists.apache.org/thread.html/0c694eda1a9c73cf77b13e69130cc79aa4c1606cc2a5f8437d0fe892@%3Csysadmins.spamassassin.apache.org%3E
>
> __DKIM_WL_HI exists and FSL_BULK_SIG hasn't depended on it since Sat
> March
> 23 - hoping it is just the delay that David refers to.
It was something more subtle. Apparently ~automc/svn/trunk hadn't been
updated in full since March 21. It's not clear to me what exactly is
supposed to take care of that, but I've fixed it manually.
--
Bill Cole
bill@scconsult.com or billcole@apache.org
(AKA @grumpybozo and many *@billmail.scconsult.com addresses)
Available For Hire: https://linkedin.com/in/billcole
Re: /tags nightly commits
Posted by Paul Stead <pa...@gmail.com>.
Promotions and tags still seem to be a little problematic at the moment. A
run of build/mkupdates/listpromotable locally shows +300 lines difference
to active.list over the one being submitted.
https://lists.apache.org/thread.html/ceb563e7c5a6b83dda106d24e0d234e7c76cd5286ae898661b7a49c7@%3Csysadmins.spamassassin.apache.org%3E
Looks like parts of the build are still failing. Seems quite a painful part
of the setup - it looks that both build and ruleqa are working from the
same svn - is this causing conflicts? I'm commenting from a hidden view as
I don't have any actual access to the sa-vm1, sorry if assumptions are
wrong.
I guess you're also looking at these, I want to make sure people know :)
Paul
Re: /tags nightly commits
Posted by Paul Stead <pa...@gmail.com>.
The interesting part which was showing yesterday was
---8<---
t/basic_lint.t .................. ok
t/basic_lint_without_sandbox.t .. ok
FSL_BULK_SIG depends on __DKIMWL_WL_HI which is nonexistent
FSL_BULK_SIG depends on __DKIMWL_WL_HI which is nonexistent
# Failed test at t/basic_meta.t line 93.
# Looks like you failed 1 test of 2.
---8<---
https://lists.apache.org/thread.html/0c694eda1a9c73cf77b13e69130cc79aa4c1606cc2a5f8437d0fe892@%3Csysadmins.spamassassin.apache.org%3E
__DKIM_WL_HI exists and FSL_BULK_SIG hasn't depended on it since Sat March
23 - hoping it is just the delay that David refers to.
Thanks for pointing out the lint failures on my rules, I've been adjusting
my sandbox a bit to try and figure this out - committed fixes for this just
now.
On Fri, 3 May 2019 at 04:20, Bill Cole <
sa-bugz-20080315@billmail.scconsult.com> wrote:
> On 2 May 2019, at 4:44, Paul Stead wrote:
>
> > On Wed, 1 May 2019 at 17:43, Bill Cole <
> > sa-bugz-20080315@billmail.scconsult.com> wrote:
> >
> >> ...
> >>
> >> The resulting generate-new-scores.sh was a mangled mess, with
> >> r1815390
> >> being the committed version in trunk. I've fixed this up.
> >>
> >> I've also committed a change made locally on sa-vm1 to
> >> build/automc/apache2-le-ssl.conf.
> >>
> >> Apparently part of the problem causing this diversity of svn breakage
> >> was work done by root in ~automc/svn/*/, causing various checked-out
> >> files and SVN nodes to become owned by root, so that the various 'svn
> >> update' jobs run as automc (from /etc/cron.d/automc) had random
> >> breakage.
> >>
> >> All should be well now.
> >>
> >
> > Looks like there might still be an issue today :(
> >
> >
> https://lists.apache.org/thread.html/0c694eda1a9c73cf77b13e69130cc79aa4c1606cc2a5f8437d0fe892@%3Csysadmins.spamassassin.apache.org%3E
> >
> > Looks like build/mkrules is still working with an old version
> >
> > ---8<---
> > Test Summary Report
> > -------------------
> > t/basic_meta.t (Wstat: 256 Tests: 2 Failed: 1)
> > Failed test: 2
> > Non-zero exit status: 1
> > Files=3, Tests=6, 11 wallclock secs ( 0.06 usr 0.00 sys + 10.16 cusr
> > 0.71
> > csys = 10.93 CPU)
> > Result: FAIL
> > make: *** [test_dynamic] Error 1
> > Makefile:1588: recipe for target 'test_dynamic' failed
> > + exit 2
> > ---8<---
>
> Or that it is working with the current rules...
>
> Here's the interesting part of what I get from 'make test' with an
> updated trunk:
>
> t/basic_lint_without_sandbox.t .... ok
> t/basic_meta.t .................... __PDS_NEWDOMAIN depends on
> AUTHOR_FMBLA_NEWDOM which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
> T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
> nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
> T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
> nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
> T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
> nonexistent
> T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
> nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
> t/basic_meta.t .................... 1/2
> # Failed test at t/basic_meta.t line 93.
> # Looks like you failed 1 test of 2.
> t/basic_meta.t .................... Dubious, test returned 1 (wstat 256,
> 0x100)
> Failed 1/2 subtests
>
>
>
> --
> Bill Cole
>
Re: /tags nightly commits
Posted by Paul Stead <pa...@gmail.com>.
The interesting part which was showing yesterday was
---8<---
t/basic_lint.t .................. ok
t/basic_lint_without_sandbox.t .. ok
FSL_BULK_SIG depends on __DKIMWL_WL_HI which is nonexistent
FSL_BULK_SIG depends on __DKIMWL_WL_HI which is nonexistent
# Failed test at t/basic_meta.t line 93.
# Looks like you failed 1 test of 2.
---8<---
https://lists.apache.org/thread.html/0c694eda1a9c73cf77b13e69130cc79aa4c1606cc2a5f8437d0fe892@%3Csysadmins.spamassassin.apache.org%3E
__DKIM_WL_HI exists and FSL_BULK_SIG hasn't depended on it since Sat March
23 - hoping it is just the delay that David refers to.
Thanks for pointing out the lint failures on my rules, I've been adjusting
my sandbox a bit to try and figure this out - committed fixes for this just
now.
On Fri, 3 May 2019 at 04:20, Bill Cole <
sa-bugz-20080315@billmail.scconsult.com> wrote:
> On 2 May 2019, at 4:44, Paul Stead wrote:
>
> > On Wed, 1 May 2019 at 17:43, Bill Cole <
> > sa-bugz-20080315@billmail.scconsult.com> wrote:
> >
> >> ...
> >>
> >> The resulting generate-new-scores.sh was a mangled mess, with
> >> r1815390
> >> being the committed version in trunk. I've fixed this up.
> >>
> >> I've also committed a change made locally on sa-vm1 to
> >> build/automc/apache2-le-ssl.conf.
> >>
> >> Apparently part of the problem causing this diversity of svn breakage
> >> was work done by root in ~automc/svn/*/, causing various checked-out
> >> files and SVN nodes to become owned by root, so that the various 'svn
> >> update' jobs run as automc (from /etc/cron.d/automc) had random
> >> breakage.
> >>
> >> All should be well now.
> >>
> >
> > Looks like there might still be an issue today :(
> >
> >
> https://lists.apache.org/thread.html/0c694eda1a9c73cf77b13e69130cc79aa4c1606cc2a5f8437d0fe892@%3Csysadmins.spamassassin.apache.org%3E
> >
> > Looks like build/mkrules is still working with an old version
> >
> > ---8<---
> > Test Summary Report
> > -------------------
> > t/basic_meta.t (Wstat: 256 Tests: 2 Failed: 1)
> > Failed test: 2
> > Non-zero exit status: 1
> > Files=3, Tests=6, 11 wallclock secs ( 0.06 usr 0.00 sys + 10.16 cusr
> > 0.71
> > csys = 10.93 CPU)
> > Result: FAIL
> > make: *** [test_dynamic] Error 1
> > Makefile:1588: recipe for target 'test_dynamic' failed
> > + exit 2
> > ---8<---
>
> Or that it is working with the current rules...
>
> Here's the interesting part of what I get from 'make test' with an
> updated trunk:
>
> t/basic_lint_without_sandbox.t .... ok
> t/basic_meta.t .................... __PDS_NEWDOMAIN depends on
> AUTHOR_FMBLA_NEWDOM which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
> T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
> nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
> T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
> nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
> T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
> nonexistent
> T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
> nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
> __PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
> t/basic_meta.t .................... 1/2
> # Failed test at t/basic_meta.t line 93.
> # Looks like you failed 1 test of 2.
> t/basic_meta.t .................... Dubious, test returned 1 (wstat 256,
> 0x100)
> Failed 1/2 subtests
>
>
>
> --
> Bill Cole
>
Re: /tags nightly commits
Posted by Bill Cole <sa...@billmail.scconsult.com>.
On 2 May 2019, at 4:44, Paul Stead wrote:
> On Wed, 1 May 2019 at 17:43, Bill Cole <
> sa-bugz-20080315@billmail.scconsult.com> wrote:
>
>> ...
>>
>> The resulting generate-new-scores.sh was a mangled mess, with
>> r1815390
>> being the committed version in trunk. I've fixed this up.
>>
>> I've also committed a change made locally on sa-vm1 to
>> build/automc/apache2-le-ssl.conf.
>>
>> Apparently part of the problem causing this diversity of svn breakage
>> was work done by root in ~automc/svn/*/, causing various checked-out
>> files and SVN nodes to become owned by root, so that the various 'svn
>> update' jobs run as automc (from /etc/cron.d/automc) had random
>> breakage.
>>
>> All should be well now.
>>
>
> Looks like there might still be an issue today :(
>
> https://lists.apache.org/thread.html/0c694eda1a9c73cf77b13e69130cc79aa4c1606cc2a5f8437d0fe892@%3Csysadmins.spamassassin.apache.org%3E
>
> Looks like build/mkrules is still working with an old version
>
> ---8<---
> Test Summary Report
> -------------------
> t/basic_meta.t (Wstat: 256 Tests: 2 Failed: 1)
> Failed test: 2
> Non-zero exit status: 1
> Files=3, Tests=6, 11 wallclock secs ( 0.06 usr 0.00 sys + 10.16 cusr
> 0.71
> csys = 10.93 CPU)
> Result: FAIL
> make: *** [test_dynamic] Error 1
> Makefile:1588: recipe for target 'test_dynamic' failed
> + exit 2
> ---8<---
Or that it is working with the current rules...
Here's the interesting part of what I get from 'make test' with an
updated trunk:
t/basic_lint_without_sandbox.t .... ok
t/basic_meta.t .................... __PDS_NEWDOMAIN depends on
AUTHOR_FMBLA_NEWDOM which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
nonexistent
T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
t/basic_meta.t .................... 1/2
# Failed test at t/basic_meta.t line 93.
# Looks like you failed 1 test of 2.
t/basic_meta.t .................... Dubious, test returned 1 (wstat 256,
0x100)
Failed 1/2 subtests
--
Bill Cole
Re: /tags nightly commits
Posted by Bill Cole <sa...@billmail.scconsult.com>.
On 2 May 2019, at 4:44, Paul Stead wrote:
> On Wed, 1 May 2019 at 17:43, Bill Cole <
> sa-bugz-20080315@billmail.scconsult.com> wrote:
>
>> ...
>>
>> The resulting generate-new-scores.sh was a mangled mess, with
>> r1815390
>> being the committed version in trunk. I've fixed this up.
>>
>> I've also committed a change made locally on sa-vm1 to
>> build/automc/apache2-le-ssl.conf.
>>
>> Apparently part of the problem causing this diversity of svn breakage
>> was work done by root in ~automc/svn/*/, causing various checked-out
>> files and SVN nodes to become owned by root, so that the various 'svn
>> update' jobs run as automc (from /etc/cron.d/automc) had random
>> breakage.
>>
>> All should be well now.
>>
>
> Looks like there might still be an issue today :(
>
> https://lists.apache.org/thread.html/0c694eda1a9c73cf77b13e69130cc79aa4c1606cc2a5f8437d0fe892@%3Csysadmins.spamassassin.apache.org%3E
>
> Looks like build/mkrules is still working with an old version
>
> ---8<---
> Test Summary Report
> -------------------
> t/basic_meta.t (Wstat: 256 Tests: 2 Failed: 1)
> Failed test: 2
> Non-zero exit status: 1
> Files=3, Tests=6, 11 wallclock secs ( 0.06 usr 0.00 sys + 10.16 cusr
> 0.71
> csys = 10.93 CPU)
> Result: FAIL
> make: *** [test_dynamic] Error 1
> Makefile:1588: recipe for target 'test_dynamic' failed
> + exit 2
> ---8<---
Or that it is working with the current rules...
Here's the interesting part of what I get from 'make test' with an
updated trunk:
t/basic_lint_without_sandbox.t .... ok
t/basic_meta.t .................... __PDS_NEWDOMAIN depends on
AUTHOR_FMBLA_NEWDOM which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
nonexistent
T_GOOGLE_DRIVE_REPLY_BAD_NTLD depends on __GOOGLE_DRIVE_SHARE which is
nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM14 which is nonexistent
__PDS_NEWDOMAIN depends on AUTHOR_FMBLA_NEWDOM28 which is nonexistent
t/basic_meta.t .................... 1/2
# Failed test at t/basic_meta.t line 93.
# Looks like you failed 1 test of 2.
t/basic_meta.t .................... Dubious, test returned 1 (wstat 256,
0x100)
Failed 1/2 subtests
--
Bill Cole
Re: /tags nightly commits
Posted by Paul Stead <pa...@gmail.com>.
On Wed, 1 May 2019 at 17:43, Bill Cole <
sa-bugz-20080315@billmail.scconsult.com> wrote:
> ...
>
> The resulting generate-new-scores.sh was a mangled mess, with r1815390
> being the committed version in trunk. I've fixed this up.
>
> I've also committed a change made locally on sa-vm1 to
> build/automc/apache2-le-ssl.conf.
>
> Apparently part of the problem causing this diversity of svn breakage
> was work done by root in ~automc/svn/*/, causing various checked-out
> files and SVN nodes to become owned by root, so that the various 'svn
> update' jobs run as automc (from /etc/cron.d/automc) had random
> breakage.
>
> All should be well now.
>
Looks like there might still be an issue today :(
https://lists.apache.org/thread.html/0c694eda1a9c73cf77b13e69130cc79aa4c1606cc2a5f8437d0fe892@%3Csysadmins.spamassassin.apache.org%3E
Looks like build/mkrules is still working with an old version
---8<---
Test Summary Report
-------------------
t/basic_meta.t (Wstat: 256 Tests: 2 Failed: 1)
Failed test: 2
Non-zero exit status: 1
Files=3, Tests=6, 11 wallclock secs ( 0.06 usr 0.00 sys + 10.16 cusr 0.71
csys = 10.93 CPU)
Result: FAIL
make: *** [test_dynamic] Error 1
Makefile:1588: recipe for target 'test_dynamic' failed
+ exit 2
---8<---
Re: /tags nightly commits
Posted by Paul Stead <pa...@gmail.com>.
On Wed, 1 May 2019 at 17:43, Bill Cole <
sa-bugz-20080315@billmail.scconsult.com> wrote:
> ...
>
> The resulting generate-new-scores.sh was a mangled mess, with r1815390
> being the committed version in trunk. I've fixed this up.
>
> I've also committed a change made locally on sa-vm1 to
> build/automc/apache2-le-ssl.conf.
>
> Apparently part of the problem causing this diversity of svn breakage
> was work done by root in ~automc/svn/*/, causing various checked-out
> files and SVN nodes to become owned by root, so that the various 'svn
> update' jobs run as automc (from /etc/cron.d/automc) had random
> breakage.
>
> All should be well now.
>
Looks like there might still be an issue today :(
https://lists.apache.org/thread.html/0c694eda1a9c73cf77b13e69130cc79aa4c1606cc2a5f8437d0fe892@%3Csysadmins.spamassassin.apache.org%3E
Looks like build/mkrules is still working with an old version
---8<---
Test Summary Report
-------------------
t/basic_meta.t (Wstat: 256 Tests: 2 Failed: 1)
Failed test: 2
Non-zero exit status: 1
Files=3, Tests=6, 11 wallclock secs ( 0.06 usr 0.00 sys + 10.16 cusr 0.71
csys = 10.93 CPU)
Result: FAIL
make: *** [test_dynamic] Error 1
Makefile:1588: recipe for target 'test_dynamic' failed
+ exit 2
---8<---
Re: /tags nightly commits
Posted by Bill Cole <sa...@billmail.scconsult.com>.
On 1 May 2019, at 8:59, Kevin A. McGrail wrote:
> Adding sysadmins@
>
> Dave, I am guessing something got left in a previous run when svn
> commit's failed. Can you or I delete a check out dir that is
> cruft?
> I'm thinking a run went bad in february and couldn't commit. I
> figure
> it should re-export, etc.?
There apparently was a conflict in place for over a year:
automc@sa-vm1:~/svn/masses/rule-update-score-gen$ ls -l
generate-new-scores.sh*
-rwxrwxr-x 1 automc automc 11530 Nov 15 2017 generate-new-scores.sh
-rw-rw-r-- 1 automc automc 11009 Nov 15 2017
generate-new-scores.sh.mine
-rw-rw-r-- 1 automc automc 10827 Nov 15 2017
generate-new-scores.sh.r1815375
-rw-rw-r-- 1 automc automc 10888 Nov 15 2017
generate-new-scores.sh.r1815390
The resulting generate-new-scores.sh was a mangled mess, with r1815390
being the committed version in trunk. I've fixed this up.
I've also committed a change made locally on sa-vm1 to
build/automc/apache2-le-ssl.conf.
Apparently part of the problem causing this diversity of svn breakage
was work done by root in ~automc/svn/*/, causing various checked-out
files and SVN nodes to become owned by root, so that the various 'svn
update' jobs run as automc (from /etc/cron.d/automc) had random
breakage.
All should be well now.
> Regards,KAM
>
> On 5/1/2019 5:59 AM, Paul Stead wrote:
>> Howdy!
>>
>> As per -
>> https://lists.apache.org/thread.html/984ee3c26163e0281727256eabf0f29079a3fc053664416f58581ba0@%3Csysadmins.spamassassin.apache.org%3E
>>
>> It looks as though the /tags commit is broken too -
>> http://svn.apache.org/viewvc/spamassassin/tags/?view=log
>>
>> ---8<---
>>
>> t/basic_lint_without_sandbox.t ..
>> okhttps://lists.apache.org/thread.html/984ee3c26163e0281727256eabf0f29079a3fc053664416f58581ba0@%3Csysadmins.spamassassin.apache.org%3E
>> depends on __DKIMWL_WL_HI which is nonexistent
>> FSL_BULK_SIG depends on __DKIMWL_WL_HI which is nonexistent
>>
>> # Failed test at t/basic_meta.t line 93.
>> # Looks like you failed 1 test of 2.
>> t/basic_meta.t ..................
>> Dubious, test returned 1 (wstat 256, 0x100)
>> Failed 1/2 subtests
>>
>> ---8<---
>>
>> The sub rule does exist, but FSL_BULK_SIG doesn't depend on
>> __DKIMWL_WL_HI any more.
>>
>> It looks as though the SVN that is being worked with isn't up to date
>> for some reason.
>>
>> Paul
>>
>
> --
> Kevin A. McGrail
> Member, Apache Software Foundation
> Chair Emeritus Apache SpamAssassin Project
> https://www.linkedin.com/in/kmcgrail - 703.798.0171
--
Bill Cole
bill@scconsult.com or billcole@apache.org
(AKA @grumpybozo and many *@billmail.scconsult.com addresses)
Available For Hire: https://linkedin.com/in/billcole
Re: /tags nightly commits
Posted by Dave Jones <da...@apache.org>.
On 5/1/19 7:59 AM, Kevin A. McGrail wrote:
> Adding sysadmins@
>
> Dave, I am guessing something got left in a previous run when svn
> commit's failed. Can you or I delete a check out dir that is cruft?
> I'm thinking a run went bad in february and couldn't commit. I figure
> it should re-export, etc.?
I was probably rushed to fix some thing and did it as root without
fixing perms before checking it in. Sorry about that. Been working too
many hours at my primary job which will cause these kinds of accidents.
Looks like Bill Cole has sorted this out. Thanks Bill.
>
> Regards,KAM
>
> On 5/1/2019 5:59 AM, Paul Stead wrote:
>> Howdy!
>>
>> As per -
>> https://lists.apache.org/thread.html/984ee3c26163e0281727256eabf0f29079a3fc053664416f58581ba0@%3Csysadmins.spamassassin.apache.org%3E
>>
>> It looks as though the /tags commit is broken too -
>> http://svn.apache.org/viewvc/spamassassin/tags/?view=log
>>
>> ---8<---
>>
>> t/basic_lint_without_sandbox.t ..
>> okhttps://lists.apache.org/thread.html/984ee3c26163e0281727256eabf0f29079a3fc053664416f58581ba0@%3Csysadmins.spamassassin.apache.org%3E
>> depends on __DKIMWL_WL_HI which is nonexistent
>> FSL_BULK_SIG depends on __DKIMWL_WL_HI which is nonexistent
>>
>> # Failed test at t/basic_meta.t line 93.
>> # Looks like you failed 1 test of 2.
>> t/basic_meta.t ..................
>> Dubious, test returned 1 (wstat 256, 0x100)
>> Failed 1/2 subtests
>>
>> ---8<---
>>
>> The sub rule does exist, but FSL_BULK_SIG doesn't depend on
>> __DKIMWL_WL_HI any more.
>>
>> It looks as though the SVN that is being worked with isn't up to date
>> for some reason.
>>
>> Paul
>>
>
Re: /tags nightly commits
Posted by Bill Cole <sa...@billmail.scconsult.com>.
On 1 May 2019, at 8:59, Kevin A. McGrail wrote:
> Adding sysadmins@
>
> Dave, I am guessing something got left in a previous run when svn
> commit's failed. Can you or I delete a check out dir that is
> cruft?
> I'm thinking a run went bad in february and couldn't commit. I
> figure
> it should re-export, etc.?
There apparently was a conflict in place for over a year:
automc@sa-vm1:~/svn/masses/rule-update-score-gen$ ls -l
generate-new-scores.sh*
-rwxrwxr-x 1 automc automc 11530 Nov 15 2017 generate-new-scores.sh
-rw-rw-r-- 1 automc automc 11009 Nov 15 2017
generate-new-scores.sh.mine
-rw-rw-r-- 1 automc automc 10827 Nov 15 2017
generate-new-scores.sh.r1815375
-rw-rw-r-- 1 automc automc 10888 Nov 15 2017
generate-new-scores.sh.r1815390
The resulting generate-new-scores.sh was a mangled mess, with r1815390
being the committed version in trunk. I've fixed this up.
I've also committed a change made locally on sa-vm1 to
build/automc/apache2-le-ssl.conf.
Apparently part of the problem causing this diversity of svn breakage
was work done by root in ~automc/svn/*/, causing various checked-out
files and SVN nodes to become owned by root, so that the various 'svn
update' jobs run as automc (from /etc/cron.d/automc) had random
breakage.
All should be well now.
> Regards,KAM
>
> On 5/1/2019 5:59 AM, Paul Stead wrote:
>> Howdy!
>>
>> As per -
>> https://lists.apache.org/thread.html/984ee3c26163e0281727256eabf0f29079a3fc053664416f58581ba0@%3Csysadmins.spamassassin.apache.org%3E
>>
>> It looks as though the /tags commit is broken too -
>> http://svn.apache.org/viewvc/spamassassin/tags/?view=log
>>
>> ---8<---
>>
>> t/basic_lint_without_sandbox.t ..
>> okhttps://lists.apache.org/thread.html/984ee3c26163e0281727256eabf0f29079a3fc053664416f58581ba0@%3Csysadmins.spamassassin.apache.org%3E
>> depends on __DKIMWL_WL_HI which is nonexistent
>> FSL_BULK_SIG depends on __DKIMWL_WL_HI which is nonexistent
>>
>> # Failed test at t/basic_meta.t line 93.
>> # Looks like you failed 1 test of 2.
>> t/basic_meta.t ..................
>> Dubious, test returned 1 (wstat 256, 0x100)
>> Failed 1/2 subtests
>>
>> ---8<---
>>
>> The sub rule does exist, but FSL_BULK_SIG doesn't depend on
>> __DKIMWL_WL_HI any more.
>>
>> It looks as though the SVN that is being worked with isn't up to date
>> for some reason.
>>
>> Paul
>>
>
> --
> Kevin A. McGrail
> Member, Apache Software Foundation
> Chair Emeritus Apache SpamAssassin Project
> https://www.linkedin.com/in/kmcgrail - 703.798.0171
--
Bill Cole
bill@scconsult.com or billcole@apache.org
(AKA @grumpybozo and many *@billmail.scconsult.com addresses)
Available For Hire: https://linkedin.com/in/billcole
Re: /tags nightly commits
Posted by "Kevin A. McGrail" <km...@apache.org>.
Adding sysadmins@
Dave, I am guessing something got left in a previous run when svn
commit's failed. Can you or I delete a check out dir that is cruft?
I'm thinking a run went bad in february and couldn't commit. I figure
it should re-export, etc.?
Regards,KAM
On 5/1/2019 5:59 AM, Paul Stead wrote:
> Howdy!
>
> As per -
> https://lists.apache.org/thread.html/984ee3c26163e0281727256eabf0f29079a3fc053664416f58581ba0@%3Csysadmins.spamassassin.apache.org%3E
>
> It looks as though the /tags commit is broken too -
> http://svn.apache.org/viewvc/spamassassin/tags/?view=log
>
> ---8<---
>
> t/basic_lint_without_sandbox.t ..
> okhttps://lists.apache.org/thread.html/984ee3c26163e0281727256eabf0f29079a3fc053664416f58581ba0@%3Csysadmins.spamassassin.apache.org%3E
> depends on __DKIMWL_WL_HI which is nonexistent
> FSL_BULK_SIG depends on __DKIMWL_WL_HI which is nonexistent
>
> # Failed test at t/basic_meta.t line 93.
> # Looks like you failed 1 test of 2.
> t/basic_meta.t ..................
> Dubious, test returned 1 (wstat 256, 0x100)
> Failed 1/2 subtests
>
> ---8<---
>
> The sub rule does exist, but FSL_BULK_SIG doesn't depend on
> __DKIMWL_WL_HI any more.
>
> It looks as though the SVN that is being worked with isn't up to date
> for some reason.
>
> Paul
>
--
Kevin A. McGrail
Member, Apache Software Foundation
Chair Emeritus Apache SpamAssassin Project
https://www.linkedin.com/in/kmcgrail - 703.798.0171
Re: /tags nightly commits
Posted by "Kevin A. McGrail" <km...@apache.org>.
Adding sysadmins@
Dave, I am guessing something got left in a previous run when svn
commit's failed. Can you or I delete a check out dir that is cruft?
I'm thinking a run went bad in february and couldn't commit. I figure
it should re-export, etc.?
Regards,KAM
On 5/1/2019 5:59 AM, Paul Stead wrote:
> Howdy!
>
> As per -
> https://lists.apache.org/thread.html/984ee3c26163e0281727256eabf0f29079a3fc053664416f58581ba0@%3Csysadmins.spamassassin.apache.org%3E
>
> It looks as though the /tags commit is broken too -
> http://svn.apache.org/viewvc/spamassassin/tags/?view=log
>
> ---8<---
>
> t/basic_lint_without_sandbox.t ..
> okhttps://lists.apache.org/thread.html/984ee3c26163e0281727256eabf0f29079a3fc053664416f58581ba0@%3Csysadmins.spamassassin.apache.org%3E
> depends on __DKIMWL_WL_HI which is nonexistent
> FSL_BULK_SIG depends on __DKIMWL_WL_HI which is nonexistent
>
> # Failed test at t/basic_meta.t line 93.
> # Looks like you failed 1 test of 2.
> t/basic_meta.t ..................
> Dubious, test returned 1 (wstat 256, 0x100)
> Failed 1/2 subtests
>
> ---8<---
>
> The sub rule does exist, but FSL_BULK_SIG doesn't depend on
> __DKIMWL_WL_HI any more.
>
> It looks as though the SVN that is being worked with isn't up to date
> for some reason.
>
> Paul
>
--
Kevin A. McGrail
Member, Apache Software Foundation
Chair Emeritus Apache SpamAssassin Project
https://www.linkedin.com/in/kmcgrail - 703.798.0171