g2k22 - OpenBSD hackathon Burg Liebenzell
# 23 Sep 2022Since I have been living near Stuttgart for over a year I decided to travel to g2k22 by bike.
On the way to #g2k22 #OpenBSD pic.twitter.com/Hdolr8x9k0
— Rafael Sadowski (@sizeofvoid) August 29, 2022
The days before g2k22 I was a little unmotivated because I didn’t know what to work on. My yearly huge KDE Gear update to 22.08 was already committed. Even CMake 2.24.1 has already made it in.
When I arrived at the castle, I met sdk@ first. I asked him if he could build the Qt6 for me and at the and of the first day I got an OK from him and we have almost Qt6 imported in OpenBSD now. Of course without x11/qt6/qtwebengine I’ll hack on it if we see consumers otherwise wasting several hours of work makes no sense.
Almost done, #OpenBSD Qt6 porting Sunday morning pic.twitter.com/xzbF9mVXfe
— Rafael Sadowski (@sizeofvoid) June 19, 2022
Next “big” thing on my list was cmake.port.mk
. I thought we could certainly
improve that. I end up with 3 changes:
-
Use cmake(1) and ctest(1) instead of ninja(1)
cmake(1) controls the build and install task/jobs. This is the way CMake prefers and it works very well for us. Full bulk build passed! Victims already fixed.
-
Remove USE_NINJA from consumers Is no longer necessary but still optional (requested by sthen@). This diff removes all usage. Expect www/h2o (see at ports@) This is possible because of the use of cmake(1).
-
Improve Verbose mode
Before:
===> Regression tests for cmark-0.30.2p0
Test project /usr/ports/pobj/cmark-0.30.2/build-amd64
Start 1: api_test
1/9 Test #1: api_test ......................... Passed 0.01 sec
Start 2: html_normalization
2/9 Test #2: html_normalization ............... Passed 0.16 sec
Start 3: spectest_library
With the new option MODCMAKE_VERBOSE
(default to yes):
===> Regression tests for cmark-0.30.2p0
UpdateCTestConfiguration from :/usr/ports/pobj/cmark-0.30.2/build-amd64/DartConfiguration.tcl 21:09:28 [17/151]
Test project /usr/ports/pobj/cmark-0.30.2/build-amd64
Constructing a list of tests
Done constructing a list of tests
Updating test list for fixtures
Added 0 tests to meet fixture requirements
Checking test dependency graph...
Checking test dependency graph end
test 1
Start 1: api_test
1: Test command: /usr/ports/pobj/cmark-0.30.2/build-amd64/api_test/api_test
1: Working Directory: /usr/ports/pobj/cmark-0.30.2/build-amd64/testdir
1: Test timeout computed to be: 10000000
1: 539 tests passed, 0 failed, 0 skipped
1: PASS
1/9 Test #1: api_test ......................... Passed 0.01 sec
test 2
Start 2: html_normalization
2: Test command: /usr/local/bin/python3.9 "-m" "doctest" "/usr/ports/pobj/cmark-0.30.2/cmark-0.30.2/test/normalize.py"
2: Working Directory: /usr/ports/pobj/cmark-0.30.2/build-amd64/testdir
2: Test timeout computed to be: 10000000
2/9 Test #2: html_normalization ............... Passed 0.16 sec
That was easier than I thought and tb@ was kind enough to start an bulk build for me. Expect from 2-3 broken ports not much had been broken. I was able to fix the two victims quickly. I’m starting to be really happy with our CMake port and module.
Apart from that, here’s a short list of what I’ve been working on.
- devel/boost 1.80
- Update to 1.80 and backport some patches from upstram:
- Fix: Boost.Unordered containers are not in a valid state after moving
- https://www.boost.org/users/history/version_1_80_0.html
- https://github.com/boostorg/unordered/issues/139
- graphics/mapnik
- cad/qcad
- qcad to 3.27.6.7
- gpgme to 1.18.0
- Remove mlt6 and webvfx
- Update quazip to 1.3
- chessx to 1.5.6
- krita to 5.1.0
- doxygen to 1.9.5
- cmark to 0.30.2
- textproc/codespell
- misc/open62541
- meta/qt5
- fonts/font-awesome
- devel/jenkins/devel
- sysutils/krename
- www/h2o
- meta/qt6
- multimedia/assimp
- net/neochat
- devel/cmocka
All in all I can look back on a really successful hackathon. Since I mainly operate only in ports, it was nice to meet people from the kernel space. On Monday morning I left the castle in the direction of Nürnberg where I worked for a customer for the next two days.
Thank you very much to Genua, the team from Burg Liebenzell and jan@ for organizing a great hackathon.