Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Dec 14 22:37

    stefanseefeld on gh-pages

    Deploy to GitHub Pages: fa4e935… (compare)

  • Dec 14 18:40

    mloskot on develop

    Fix illegal narrowing conversio… (compare)

  • Dec 14 18:40
    mloskot closed #409
  • Dec 14 18:40
    mloskot closed #408
  • Dec 14 18:36
    mloskot edited #409
  • Dec 14 18:08
    mloskot edited #409
  • Dec 14 07:22
  • Dec 13 20:53
    Build #275 passed
  • Dec 13 20:03
    Travis mloskot/gil (ml/ext-io-png-fix-narrowing-conversion) passed (653)
  • Dec 13 19:45
    mloskot milestoned #408
  • Dec 13 19:45
    mloskot milestoned #409
  • Dec 13 19:41
    mloskot edited #409
  • Dec 13 19:41
    mloskot labeled #408
  • Dec 13 19:41
    mloskot labeled #408
  • Dec 13 19:40
    mloskot review_requested #409
  • Dec 13 19:40
    mloskot labeled #409
  • Dec 13 19:40
    mloskot opened #409
  • Dec 13 19:40
    mloskot labeled #409
  • Dec 13 19:30

    mloskot on develop

    Refactor CMakeSettings.json to … (compare)

  • Dec 13 19:29
    mloskot commented #408
Stefan Seefeld
@stefanseefeld
done !
Mateusz Łoskot
@mloskot
Cool!
Stefan Seefeld
@stefanseefeld
congrats to all contributors ! :-)
Mateusz Łoskot
@mloskot
Keep fingers crossed master will build CI-s green :)
Stefan Seefeld
@stefanseefeld
yeah, I already got green from a build on my own fork (where I pushed master first), so I'm not worried. At least not about the travis builds. :-)
Mateusz Łoskot
@mloskot
Good.
That made me thing the merge could go via a PR ad be CI-checked and then merged :)
By the way, we should get rid of Travis CI for its irresponsive uselessness
Stefan Seefeld
@stefanseefeld
perhaps
Rene suggested to rely on Travs outside boostorg/
https://cpplang.slack.com/archives/C27KZLB0X/p1572112578114600
Almost all PRs I merged recently, I ignored (waiting for) Travis boostorg/gil and I just added pre-merge comment and linked successful buld to my Travis for mloskot/gil. This is annoying though.
On cpplang.slack.com, on #boost-ci channel, there was discussion about paying $$$ to Travis to solve the problem but it's been quiet about it
Anyhow, it is something to consider how to improve the situation in near future.
Stefan Seefeld
@stefanseefeld
yes, absolutely.
Mateusz Łoskot
@mloskot
After the release, I'm also going to update our CI-s simplifying the config and unifying them a bit (majority work finished inhttps://github.com/mloskot/gil/tree/pr/ml/unify-ci-and-switch-to-boost-ci). So, then we can recall the issue and decide.
Stefan Seefeld
@stefanseefeld
OK, sounds good.
Mateusz Łoskot
@mloskot

Back to the release, here is full change log for 1.72 (chanes since 1.71)

https://gist.github.com/mloskot/993e3d3b471a7e3c02e5ac8199d2b4ad

Those are subjects of commits linked to GitHub but, thanks to our disciplined integration of major changes via PRs :wink: , each includes PR number.

So, I think fishing out for major changes from the ones with PR numbers should be enough for release notes.

@stefanseefeld What do you think?
Stefan Seefeld
@stefanseefeld
Yeah, that's a great idea. However (now you are going to disagree with me :-) ), I'd rather add a section into our own docs about release notes, than doing this dance with boost's website repo. You know my opinion...
Mateusz Łoskot
@mloskot
Surprise surprise :-)
I really think having a release-notes.md page in GIL docs is a great idea!
Stefan Seefeld
@stefanseefeld
ok, let's start with that, then we can still debate whether (and how) to push this upstream.
Mateusz Łoskot
@mloskot
I wonder if we could have NEWS.md file next to README.md where all release notes are collected.
Then, feed Sphinx with this file
Stefan Seefeld
@stefanseefeld
yeah, something like that would be ideal. I'm not sure whether sphinx can consume markdown. If not, whether github would serve restructured text. Either way, we'll figure something out... :-)
Mateusz Łoskot
@mloskot
And, Sphinx would generate boostorg.github.io/gil/html/news.html
This message was deleted
Stefan Seefeld
@stefanseefeld
yes, we are on the same page there. :-)
Mateusz Łoskot
@mloskot
I told ya ;)

One of projects I contribute to does this:
maintains
https://github.com/OSGeo/PROJ/blob/master/NEWS
and generates
https://proj.org/news.html

This is all Sphinx-based, but I will have to look how it is done

Mateusz Łoskot
@mloskot
Sounds like it may do the job
Stefan Seefeld
@stefanseefeld
anyhow, gotta go. Thanks for all your help with the merge to master !
Mateusz Łoskot
@mloskot
The PROJ project maintains release notes in NEWS and manually copies it to news.rst during release
https://github.com/OSGeo/PROJ/blob/master/HOWTO-RELEASE#L178
I will look into the Sphinx and Markdown integration so we can have it written ones in NEWS.md :)
I'm glad the merge is done
Mateusz Łoskot
@mloskot

Alternative to single NEWS.md - which is quite nice for quick inspection, search and display in top-level directory -
could be CMake's approach https://github.com/Kitware/CMake/tree/master/Help/release

So, we coud maintain release notes in similar way in doc/release/X.Y.rst.

I'm kind of more keen to have single NEWS.md
Our notes are note huge and single file is easier to maintain, read, search, etc.
now I go, bye :-)
Mateusz Łoskot
@mloskot

As part of release notes management, looking what people do these days and watching this talk about reno

https://www.youtube.com/watch?v=tEOGJ_h0Lx0

@stefanseefeld :point_up:
Mateusz Łoskot
@mloskot
FYI, today Packt is giving away "Mastering OpenCV 4" 3rd edition by Roy Shilkrot, David Millán Escrivá at https://www.packtpub.com/free-learning
Olzhas Zhumabek
@simmplecoder
wow, thanks. Was needing something like that
Mateusz Łoskot
@mloskot
@stefanseefeld I'm a bit busy, kids sick, did not manage to work out the release notes yet. I think for this release we need to go the old way.
I also did not have time to look at reports by @misos1 e.g. #401
I will try to do it ASAP, but not seeing chances for the release now
Pranam Lashkari
@lpranam
https://github.com/boostorg/gil/blob/c24e8b2518145fef26f27adbf1e3f9abe3a14015/include/boost/gil/extension/numeric/kernel.hpp#L161
Here Core is expected to be some container like std::array or std::vector but these containers do not have virtual destructors. How is that handled?
Mateusz Łoskot
@mloskot
AFAIK, none of those are used in situation where instances are deleted via pointer to base class.