www/firefox: Python(?) issues building under poudriere (stable/14)
Date: Sat, 11 May 2024 14:29:29 UTC
TL;DR: As the Subject says; log & info at
https://www.catwhisker.org/~david/FreeBSD/ports/firefox/
Additional detail:
I have 2 broad categories of FreeBSD machines here at home (all amd64):
* Development machines (a build machine & some laptops)
* Production machines (a multi-homed "traffic cop"/gateway & a main
machine)
The development machines get updated daily, tracking stable/14 & head
(on dfferent slices); installed ports are updated daily on stable/14
using portmaster (except for www/chromium). The laptops have a
slightly-customized kernel; the build machine runs GENERIC.
The production machines get updated weekly (Sunday mornings), with the
just-built snapshot of stable/14 (courtesy of the build machine). The
build machine also runs poudriere-devel to build packages for the
production machines (and for the laptops, in the case of chromium).
Package-building is done on a weekly cycle (augmented by as-needed
chromium-specific runs); in order to try to save a bit of time on
"update day" (Sunday), I make an initial package-building run on
Saturday (which, not so coincidentally, is today).
Usually, this process is ... well, boring. (Which is a Very Good Thing,
as I remind my colleagues at work -- especially when I'm shepherding a
new release.) It also goes a fairly long way toward ensuring that the
machines I use have "consistent" environments: there is no "piecemeal
updating" going on here.
Last week, poudriere failed to build firefox (on both Saturday and
Sunday, IIRC). While that was moderately annoying, it didn't seem
fatal: the previously-built firefox had already been installed on the
one production machine that has it. And the laptops (that build firefox
using portmaster) didn't seem to have a problem building & using it.
But the port has since been updated (to 126.0_1,2. And poudriere
has complained again. So this time, I have copied the poudriere
log of the failed build to
https://www.catwhisker.org/~david/FreeBSD/ports/firefox/ (and also
made a compressed copy; my Internet connectivity leaves a fair bit
to be desired).
Specific information for this failure:
The build machine was (and is) running:
FreeBSD freebeast.catwhisker.org 14.1-STABLE FreeBSD 14.1-STABLE #135 stable/14-n267652-91df7d335dd4: Sat May 11 10:47:37 UTC 2024 root@freebeast.catwhisker.org:/common/S1/obj/usr/src/amd64.amd64/sys/GENERIC amd64 1401500 1401500
The ports tree is at: main-n662668-8e136401b5b9
poudriere is poudriere-git-3.4.99.20240424
The fatal error in the log appears to be a Python ImportError about
Undefined symbol "sqlite3_enable_load_extension"
...
/wrkdirs/usr/ports/www/firefox/work/.mozbuild/srcdirs/firefox-126.0-10171689b53e/_virtualenvs/build/bin/python -m mozbuild.action.file_generate /wrkdirs/usr/ports/www/firefox/work/firefox-126.0/toolkit/components/telemetry/build_scripts/gen_scalar_enum.py main toolkit/components/telemetry/TelemetryScalarEnums.h toolkit/components/telemetry/.deps/TelemetryScalarEnums.h.pp toolkit/components/telemetry/.deps/TelemetryScalarEnums.h.stub /wrkdirs/usr/ports/www/firefox/work/firefox-126.0/toolkit/components/telemetry/Scalars.yaml
./TelemetryEventData.h.stub
/wrkdirs/usr/ports/www/firefox/work/.mozbuild/srcdirs/firefox-126.0-10171689b53e/_virtualenvs/build/bin/python -m mozbuild.action.file_generate /wrkdirs/usr/ports/www/firefox/work/firefox-126.0/toolkit/components/telemetry/build_scripts/gen_event_data.py main toolkit/components/telemetry/TelemetryEventData.h toolkit/components/telemetry/.deps/TelemetryEventData.h.pp toolkit/components/telemetry/.deps/TelemetryEventData.h.stub /wrkdirs/usr/ports/www/firefox/work/firefox-126.0/toolkit/components/telemetry/Events.yaml
Traceback (most recent call last):
File "/usr/local/lib/python3.9/runpy.py", line 197, in _run_module_as_main
return _run_code(code, main_globals, None,
File "/usr/local/lib/python3.9/runpy.py", line 87, in _run_code
exec(code, run_globals)
File "/wrkdirs/usr/ports/www/firefox/work/firefox-126.0/python/mozbuild/mozbuild/action/file_generate.py", line 154, in <module>
sys.exit(main(sys.argv[1:]))
File "/wrkdirs/usr/ports/www/firefox/work/firefox-126.0/python/mozbuild/mozbuild/action/file_generate.py", line 85, in main
spec.loader.exec_module(module)
File "<frozen importlib._bootstrap_external>", line 850, in exec_module
File "<frozen importlib._bootstrap>", line 228, in _call_with_frames_removed
File "/wrkdirs/usr/ports/www/firefox/work/firefox-126.0/toolkit/components/glean/build_scripts/glean_parser_ext/run_glean_parser.py", line 11, in <module>
import cpp
File "/wrkdirs/usr/ports/www/firefox/work/firefox-126.0/toolkit/components/glean/build_scripts/glean_parser_ext/cpp.py", line 14, in <module>
from glean_parser import metrics, util
File "/wrkdirs/usr/ports/www/firefox/work/firefox-126.0/third_party/python/glean_parser/glean_parser/metrics.py", line 15, in <module>
from . import pings
File "/wrkdirs/usr/ports/www/firefox/work/firefox-126.0/third_party/python/glean_parser/glean_parser/pings.py", line 14, in <module>
from . import util
File "/wrkdirs/usr/ports/www/firefox/work/firefox-126.0/third_party/python/glean_parser/glean_parser/util.py", line 17, in <module>
import diskcache # type: ignore
File "/wrkdirs/usr/ports/www/firefox/work/firefox-126.0/third_party/python/diskcache/diskcache/__init__.py", line 9, in <module>
from .core import Cache, Disk, EmptyDirWarning, JSONDisk, UnknownFileWarning, Timeout
File "/wrkdirs/usr/ports/www/firefox/work/firefox-126.0/third_party/python/diskcache/diskcache/core.py", line 14, in <module>
import sqlite3
File "/usr/local/lib/python3.9/sqlite3/__init__.py", line 57, in <module>
from sqlite3.dbapi2 import *
File "/usr/local/lib/python3.9/sqlite3/dbapi2.py", line 27, in <module>
from _sqlite3 import *
ImportError: /usr/local/lib/python3.9/lib-dynload/_sqlite3.cpython-39.so: Undefined symbol "sqlite3_enable_load_extension"
gmake[2]: *** [backend.mk:684: toolkit/components/glean/.deps/HistogramGIFFTMap.h.stub] Error 1
gmake[2]: *** Waiting for unfinished jobs....
....
The only difference in port options I find between the laptops &
poudriere is that the laptops have:
SNDIO=off: Sndio audio support
while poudriere shows:
OPTIONS_FILE_SET+=SNDIO
Finally (for this post), I admit that I've never really picked up Python
(and that SQL-related things are things I tend to muddle through, then
write scripts to handle so I don't need to try to think about them any
more).
So: Do I need to fix my process? Is this actually symptomatic of
something that warrants a bug report?
Thanks!
Peace,
david
--
David H. Wolfskill david@catwhisker.org
Please do not mistake "authoritarian" for "conservative" -- or vice versa.
See https://www.catwhisker.org/~david/publickey.gpg for my public key.