On 10/4/2012 10:27 AM, Axb wrote:
There were not freemail rule changes, they were changes I made in
maddoc's sandbox (with his permission) - the rules where named with
FREEMAIL
Roger that. They are rules dealing with Freemail is all I'm saying.
But that change removing those few rules has been rippling a bit since.
Either way, yeah, this stupid change to the freemail rules has broken
the code a couple of times so I blame you. You may either roll 8+ on
2d6 or sacrifice an intern to avoid wrath.
Is it possible that Jenkins coughs if he sees changes during him doing
some foo? (butlers ain't what they used to be anymore)
I duplicated the error on my devel system first, then added the rule to
10_has_base then did make test again and committed
Jenkins error from test points to a rule in John Hardin's sandbox:
https://builds.apache.org/job/SpamAssassin-trunk/8070/testReport/make_test/t_basic_meta_t/test______2/
Error Message
not ok 2
Stacktrace
1..2
# Running under perl version 5.008006 for solaris
# Current time local: Thu Oct 4 10:15:17 2012
# Current time GMT: Thu Oct 4 10:15:17 2012
# Using Test.pm version 1.25
TO_NO_BRKTS_NORDNS_HTML depends on __FROM_FREEMAIL which is nonexistent
TO_NO_BRKTS_NORDNS_HTML depends on __FROM_FREEMAIL which is nonexistent
TO_NO_BRKTS_NORDNS_HTML depends on __FROM_FREEMAIL which is nonexistent
TO_NO_BRKTS_NORDNS_HTML depends on __FROM_FREEMAIL which is nonexistent
ok 1
not ok 2
# Failed test 2 in t/basic_meta.t at line 91
Failed 1/2 subtests
Standard Output
1..2
# Running under perl version 5.008006 for solaris
# Current time local: Thu Oct 4 10:15:17 2012
# Current time GMT: Thu Oct 4 10:15:17 2012
# Using Test.pm version 1.25
TO_NO_BRKTS_NORDNS_HTML depends on __FROM_FREEMAIL which is nonexistent
TO_NO_BRKTS_NORDNS_HTML depends on __FROM_FREEMAIL which is nonexistent
TO_NO_BRKTS_NORDNS_HTML depends on __FROM_FREEMAIL which is nonexistent
TO_NO_BRKTS_NORDNS_HTML depends on __FROM_FREEMAIL which is nonexistent
ok 1
not ok 2
# Failed test 2 in t/basic_meta.t at line 91
Failed 1/2 subtests
John's files look like they are from June so this is likely something where the
rules you disabled in maddoc's sandbox are in use but it takes a while for the
rules testing to push the errors to the top.
Code change? Procedure change? I always have hated the way sandbox is "play" and
"real" code intertwined.
Regards,
KAM
