Re: Optimizing libpng encode time

On Sun, Aug 30, 2020 at 7:40 AM Bob Friesenhahn
<[email protected]> wrote:
>
> On Sun, 30 Aug 2020, Akira Kakuto wrote:
[...]
> > There is a github repository by current maintainer, Cosmin Truta:
> > https://github.com/pnggroup/libpng-EXPERIMENTAL
>
> Unfortunately, the last post by Cosmin Truta to this list appears to
> be on November 25th last year.
>
> Libpng does need continual maintainership, although it might be just a
> slow reliable "tick" of activity.

Hi, Akira and Bob.

I have looked into that repository a bit more.  The sourceforge code
repository appears to have all of the changes in the github
libpng-EXPERIMENTALrepository.  I pulled both repositories
into a git tree, confirmed that they have the same branches and
confirmed for each branch name that if I check out the sourceforge
branch and then attempt to merge the github branch, I get the
message "Already up to date."  Going the opposite way,from
github to sourceforge, does merge changes, at least for master.
Nevertheless, I have just now submitted a one line pull request
to the github repository for png_do_check_palette_indexes
apparently ignoring the last byte of each line, just in case
that turns out to be the right way to submit a contribution.

Bob, one good sign that we can take is that the last commit in
the sourceforge master branch was as recently as May 24, 2020,
around the time of some discussion on the mailing list.  That
commit seemed to be more relevant to the development
process rather than delivering any user functionality, so I
think I can see why it would not be the impetus for a release.
So, perhaps contributions are being handled in a reasonably
timely manner and there just have not been many
contributions recently.

With that said, as a potential contributor, I was a bit discouraged
to see that a couple of the merge requests at
https://sourceforge.net/p/libpng/code/merge-requests/ have been
open there since February, 2012 with no comment.  The
submitters have not posted any nag messages there, so perhaps
the issues were resolved in some other way, such as by other
patches being integrated, but it would be helpful to see some
response posted to those messages indicating that they are being
left open intentionally for now and perhaps mentioning why.
However, compared to the normal loose ends one finds in any
active project, that's really small.  So, I remain optimistic that
everything is on track.  We'll see.

Also, Bob, thanks for your other email.  I am looking into
some of the things you mentioned and will try to respond
in the next few days.

Adam