varnishevent:master commitshttps://code.uplex.de/uplex-varnish/varnishevent/commits/master2023-06-09T14:07:00Zhttps://code.uplex.de/uplex-varnish/varnishevent/commit/23b1d54d0aaadd611510d5634462462020286795Use a new regression test log from varnish post 7.32023-06-09T14:07:00ZNils Gorollnils.goroll@uplex.de
varnishd now uses 64bit vxids and a new VSL format, with some tags
removed.
I did ponder to write a VSL log converter or come up with some other
deliverate way to somehow preserve the old log data in a regerssion
log in the new format.
But ultimately, it all seemed pretty much wasted effort. Let's just
start with a new log and regression test future changes.
I must admit that I only hat a bit of a superficial look at the log
files, but as we made no changes at all to varnishevent, I do not see
how they could have changed fundamentally.
varnish-post-7.3.README.rst cotains information on how the log was
created.https://code.uplex.de/uplex-varnish/varnishevent/commit/332aff6550f4dacd45869e6692c886c2b662f934Standardize LICENSE2022-12-01T15:27:26ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/6892f91c901e4e9814edf86a8f66521ac17e9c98Constify2021-10-28T15:37:05ZNils Gorollnils.goroll@uplex.de
fixes:
config.c: In function 'conf_ParseLine':
config.c:206:13: warning: assignment discards 'const' qualifier from pointer target type [enabled by default]
*op = "+=";
^
config.c:210:13: warning: assignment discards 'const' qualifier from pointer target type [enabled by default]
*op = "=";
^https://code.uplex.de/uplex-varnish/varnishevent/commit/eb95c8bda6f2e8f2aec6f8f8f30867188552c2d5Move gloabl variables to their likely home source2021-10-12T07:43:11ZNils Gorollnils.goroll@uplex.de
The previous commit broke "make check"https://code.uplex.de/uplex-varnish/varnishevent/commit/d7dc0556ab69c2fdaf1005cd3ca6ad84e84e1b04Fix multiple definition of global variables2021-10-12T07:34:22ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/5b00968be4779671e4136e74b65a92870b024215Fix use-after-free in free_incl()2021-06-29T13:44:50ZNils Gorollnils.goroll@uplex.de
Seen after linux update
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Core was generated by `../varnishevent -f varnishevent.conf -r varnish-4.1.0-doc.log -w out -v'.
Program terminated with signal SIGABRT, Aborted.
50 ../sysdeps/unix/sysv/linux/raise.c: No such file or directory.
(gdb) bthttps://code.uplex.de/uplex-varnish/varnishevent/commit/a00794a4d639d5862be0db8849d0c559ea7cdcd1simplify regression tests, update checksums2019-04-25T14:06:05ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/26c023d9d97680c6145da04f6d1803b6b6369320leave trace files from vslarg.sh for analysis2019-04-25T12:19:40ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/afa9a1857f5a5553e0b04355c0d608e28c6def7afix %h formatter2019-04-25T11:42:25ZNils Gorollnils.goroll@uplex.de
Our %h formatter was broken and we did not notice due to coincidence of
two reasons:
Before 626ddb029256513b6de36d5e642299d6708d638d in varnish-cache,
varnishncsa did not output %h for our (ancient) test file because it
does not contain BackendStart records.
varnishevent did not output %h because it was using SLT_Backend, which
had been removed from varnish-cache for long.
Because both matched, we did not notice any issue.
It is only now due to the change from
626ddb029256513b6de36d5e642299d6708d638d that we noticed the issue.
So the fix is to take %h from SLT_BackendOpen and adjust all other test
details in this regard.
The fix in varnish triggered the ncsa.sh test failing because
varnishevent would also not format %h correctly: SLT_Backend had been
retired for long, but we still took %h from it.https://code.uplex.de/uplex-varnish/varnishevent/commit/a6ac776e221b64af6365e4c4b12f8e03cd578904binary data formatting has been unified in varnish-cache2018-10-27T16:20:05ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/21e14e6915eda8e1baa7a367eaba66dd0f56a34cAdd += config file operator to avoid very long lines2018-10-04T16:06:35ZNils Gorollnils.goroll@uplex.de
... and allow for comments inbetween format definitions
Also internalize CONF_Add as conf_Add as it is not used outside config.chttps://code.uplex.de/uplex-varnish/varnishevent/commit/77e6fdc124299096a4a15405c4c6fb41c19674f6reset args after init, which now zeroes them2018-03-27T11:26:51ZNils Gorollnils.goroll@uplex.de
Fixes <a href="/uplex-varnish/varnishevent/issues/3" data-original="#3" data-link="false" data-link-reference="false" data-project="20" data-issue="22" data-reference-type="issue" data-container="body" data-placement="bottom" title="len == strlen(substr) failed in test_format_SLT at test_format.c:1659 (errno 0: Success)" class="gfm gfm-issue has-tooltip">#3</a>
was actually caused by the fix for <a href="/uplex-varnish/varnishevent/issues/2" data-original="#2" data-link="false" data-link-reference="false" data-project="20" data-issue="21" data-reference-type="issue" data-container="body" data-placement="bottom" title="test_format fails with new gcc 6 and -g -O2" class="gfm gfm-issue has-tooltip">#2</a> (and I failed to recognize this)https://code.uplex.de/uplex-varnish/varnishevent/commit/638f502d5581be43fee07ca919d7575d377c6e62init arg_t variables2018-03-27T10:56:07ZNils Gorollnils.goroll@uplex.de
Fixes <a href="/uplex-varnish/varnishevent/issues/2" data-original="#2" data-link="false" data-link-reference="false" data-project="20" data-issue="21" data-reference-type="issue" data-container="body" data-placement="bottom" title="test_format fails with new gcc 6 and -g -O2" class="gfm gfm-issue has-tooltip">#2</a>https://code.uplex.de/uplex-varnish/varnishevent/commit/9e9e39f11fc14952713702676543b6d2644270b9keep only one instance of hidx2018-03-27T10:37:34ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/90356896d49531d0e59aab1c93bd597113a53286bitmask to apply on unsigned long should be that2018-03-27T10:28:50ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/0bd618ec04e6a189f5376d168e8cc1ffdd004964avoid taking reopen_lock for every write2018-03-27T09:30:12ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/bd71d41fe5c3e221b25f7fb9bf14ff9018368340fix version check2018-03-15T10:55:35ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/e13dd0a524c13e85a8015f1c6280cb60cc6884dbaccept varnish 6.02018-03-15T10:47:11ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/16b103f4bbe9c5cc1a3ca6e8c0e622b17b99201d%{Debug}x log format is desupported in varnishncsa2018-03-07T11:58:33ZNils Gorollnils.goroll@uplex.de
See varnish-cache commits
* 2da6c5e539f8f858c446812d30f00de600ecc3e4
* 63807dd22725653e23c20071a9eb958642413f36https://code.uplex.de/uplex-varnish/varnishevent/commit/313547848e421e87871da53ceffc7a0f6e514ebfkeep output files upon failure, simplify2018-03-07T11:32:54ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/adfb6e384e23461fe1779dacd9b8baa1f4cdc957support SLT_F_UNSAFE2018-03-07T10:30:34ZNils Gorollnils.goroll@uplex.de
Ref varnish-cache master 2d0843eac8d77cfb3369591a61234d9c3ed6881e
Introducedhttps://code.uplex.de/uplex-varnish/varnishevent/commit/76741251d60c125b4215cfdcd341bb55a9ab43aeonly include vdef.h once in tests2017-11-30T18:34:49ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/f659452d44fe048e4b45eb352aa9f173b4137c5cAlways include vdef.h first - directly or via varnishevent.h2017-11-30T17:45:04ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/8845dd32d26235e6b9a1fbf87e04194c12b4e647new include world order2017-11-09T14:43:41ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/120e2a4dce8e1b3b26acf9cd5a0fa1659f63762dnew include world order2017-11-09T14:05:46ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/2ed420ad9ebd2812c2c4afb5fb241b2f95ded747more include cleanup required for varnish-cache master2017-09-28T14:00:59ZNils Gorollnils.goroll@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/16fcff4286276d07e91da1999da56f8b332f5761The global VAS fail function pointer is now called VAS_Fail_Func2017-09-15T06:27:14ZNils Gorollnils.goroll@uplex.de
Ref: varnish-cache 9ad80c88f26d9a90d21b59f265d31be61ae545a0https://code.uplex.de/uplex-varnish/varnishevent/commit/d96667ebe5623c096831c827c0bf2b87c8004643Bugfix use of VSM_Dup() to obtain the Varnish instance name.2017-09-07T13:43:43ZGeoff Simmonsgeoff@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/99cc0b1ed603ab39c25bee7eaf394d74088c4ec4Remove the -N arg, no longer available in the VSM/VSL API.2017-09-07T11:51:51ZGeoff Simmonsgeoff@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/cb411094cf39e796a9b0da972404ba8584849ee8Remove the -N arg, no longer available in the VSM/VSL API.2017-09-07T11:12:17ZGeoff Simmonsgeoff@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/8e2eb99cfd80e68fae782c5150c1e782097dfd4fUpdate for compatibility with Varnish 5.2.2017-09-07T11:05:34ZGeoff Simmonsgeoff@uplex.deThe VSM API has changed.https://code.uplex.de/uplex-varnish/varnishevent/commit/cb052cc3d2bb7ee3ff88af2c4923945afc76e5c8the -N option has been retired in varnish-cache master2017-05-10T14:37:18ZNils Gorollnils.goroll@uplex.de
Ref: e9c70c7306eb94c14768648dca03fc5b6c3252f9https://code.uplex.de/uplex-varnish/varnishevent/commit/d8d880b63b0498f1d38d96eb4ef51fdc5da43c2bUpdate the compatibility statement (compatible with trunk since 5.1).2017-03-15T18:11:51ZGeoff Simmonsgeoff@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/e6f959dfe48c8191ff20215de09c6bea39cd1e1dEnsure that the installed Varnish bin dir is on PATH before make check2017-02-21T18:23:04ZGeoff Simmonsgeoff@uplex.dein particular to find varnishncsahttps://code.uplex.de/uplex-varnish/varnishevent/commit/552d95675ea4b2233dd8bcefe319ce9ebef92377Explicitly double-quote data for binary tags, since VSB_quote doesn't2017-02-21T17:48:09ZGeoff Simmonsgeoff@uplex.dedo that any more. But will have to revisit the formatting, because
Varnish VSL and its tools apparently don't use VSB_quote for binary
data any more.https://code.uplex.de/uplex-varnish/varnishevent/commit/7981187521b5299e3c61ec6b011f97391b60cc5dmake the regression tests independent of local time zone in which they are run2016-12-15T17:36:31ZGeoff Simmonsgeoff@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/add7c8556de8a5c60b6a76269c35ef49f921ae78make the unit tests for strfTIM independent of (my) local time zone2016-12-14T17:54:48ZGeoff Simmonsgeoff@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/eda06ca72129747829f586d60009401085fbf596proper use of AC_SUBSTituted variables in a Makefile template2016-12-14T13:11:47ZGeoff Simmonsgeoff@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/fc6c2133d3084aa7945becdc9807b2a9f53a0775add the path for Varnish's m4 scripts to aclocal calls invoked from make2016-12-14T13:08:52ZGeoff Simmonsgeoff@uplex.dehttps://code.uplex.de/uplex-varnish/varnishevent/commit/d3c22eef4ce28671a80cff95067ef993784a3326include the path for installed Varnish m4 scripts to the aclocal call2016-12-13T07:46:36ZGeoff Simmonsgeoff@uplex.de