You may want to know that I got warnings about leaked file descriptors in the tests of inst/montage.m with Octave-5.1, which seems to be due to code in Octave. Also, in the tests in inst/wiener.m a division by zero was reported, but the tests succeeded. Thank you for testing. These are both known issues. There's already open issues about it. Could you please provide an explicit copyright/license statement for src/configure.ac? The border to 'substantial code' seems to be passed here, I think, and...
image 2.12.0 release
You're right, the tagging commit was still on my clone only. Pushed now. Thank you for the release.
oh, I have already tagged the release, but you should be able to build it easily too. Thank you
image package 2.10.0
ping? Oliver, when you have time, could you take a look at pushing this release? Thank you
Yes, those 18 are the correct oct files that should be built.
On Sat, 27 Oct 2018 at 18:17, marco atzeri matzeri@users.sourceforge.net wrote: I do not see failure on the build, but "conndef" is absent on both 2.8.0 and 2.8.1. ~~~ octave:1> pkg load image octave:2> help conndef error: help: the 'conndef' function belongs to the image package from Octave Forge but has not yet been implemented. ~~~ For what I see on Cygwin only 14 oct shared library are built bwdist.oct bwfill.oct connectivity.oct graycomatrix.oct hough_line.oct intlut.oct nonmax_supress.oct rotate_scale.oct...
There's 197 tests failing so something odd is going on. conndef is there and should have been built. Where are you building this? Are you not seeing the same issues with version 2.8.0? Because this is a patch release, and other that the version numbers the only change is http://hg.code.sf.net/p/octave/image/rev/b56a9e8023e9 which wouldn't explain your issue.
Note that I am aware of some failing tests in impyramid and cp2tform We have no fix them for yet but they are not really expected/meant to be failing.
image package release version 2.8.0
Yes, I am aware of the test failures. They are not marked as expected failures since they are not really meant to fail, they fail because of a real bug. This is because xtest was meant not for unfixed tests but tests for non deterministic tests.
Ups. I'm sorry about that. I was trying to avoid pushing the tag and instead did not push the actual release commit. It should be fixed now.
image 2.6.2
The template at http://hg.code.sf.net/p/octave/project-web/file/dfd0d911bfc9/htdocs/templates/Makefile...
[...] cmake, which is only used (if I understand correctly) in the M4 macro to detect...
Olaf, the cmake version of debian stable has a bug that causes that. I reported it...
Makefile: use immediate set (:=) instead of laz...
Makefile: add recipe for Octave releases in zip...
Makefile: dist instead of pkg target, more chan...
I have pushed the release.
I have pushed the release.
I have pushed the release.
release pushed
Release pushed.
I have pushed the release.
A tag on mercurial is only a line on the .hgtags file. There's nothing stopping you...
Not released yet. On the process of it. Do you want me to make both releases for...
Note that the --sort option you are using for tar to get the reproducible builds...
I am a bit unsure what action you want from me. Should I just create a git repository...
I could rebuild the package just fine. However, I noticed that the list of functions...
I am closing this since gsl 2.0.0 has now been released.
I have pushed the release but I had to re-generate the html. Please update the generate...
Done.
You said it yourself, "biosig [...] contains a number of different packages". So...
Same issue as for the tsa release. Can you please provide a cset hash and a way to...
Ad mentioned on the mailing list, can you give a cset hash to reproduce the release?...
As mentioned on the mailing ist, how can one reproduce the release?
As discussed on the mailing list, can you provide a cset hash that can be used to...
I have pushed the release. You may want to confirm that the package indeed works...
I have pushed the release.
I guess your package is one of the cases with complicated releases which on the mailing...
The package html was generated with an old version of the generate_html. Can you...
I will wait for reviews. Let me know when you want the actual release done. But note...
done
Done.
This is still not fixed. The released tarball has the following files on src: $ tar...
Done.
The mex files on the package are not in the repository. We need to be able to redo...
I have pushed the release. However, you may want to keep configure.ac on the next...
I have pushed the release. Apologies for the delay. I've had to take a break due...
It is done.
It's done. I have pushed the release. A tip for a future release. After running configure...
Done. I have pushed the release.
I have pushed the releases. I removed the autoconf cached files from the optim and...
The uploaded html is for the wrong version of the package. At least the version number...
done
done
Done.
This is a very minor thing but the release date on the DESCRIPTION and NEWS file...
Could you regenerate the html with the latest version of the generate_html package?...
Can you give the checksums of the files?