Hi, > More general question: Can the package release cycle be a little more iterative?
I give up releasing beta and rc. Literally nobody was testing them. And release a IUP version is very time consuming and stressing. So I stick with the SVN availability, so anybody can build it and test it anytime. What happen with that module was an exception that last a day or two. We are working to produce a new CMake based build configuration, that I hope will make things easier for Windows too. Sub-numbers I use only for emergencial bugfix releases when the main release has a critical flaw, for the same reason described. About 3, Yes we should use branches. Best, Scuri Em dom., 22 de mar. de 2020 às 07:55, sur-behoffski < sur_behoff...@grouse.com.au> escreveu: > G'day, > > Scuri moved quickly to address the errors I reported a week ago, > regarding iup's "iup_flattree.c". The report was against iup-r5646; > the latest revision (9 hours ago, according to SourceForge) of iup-r5670 > now compiles and links (e.g. dynamic library "required") successfully. > > I've only done my usual "hello, world" sanity check on the builds, > for im-r754, cd-r849 and iup-r5670, and this trivial check has passed > on the following OSes (using virtual machines, but with updates applied, > so that my installation tools try to be tested against a "clean room" > distribution environment): > > Distro GCC Kernel > > Ubuntu-18.04.3 7.5.0-3ubuntu1... 5.3.0-32-generic > Ubuntu-19.10 9.2.1-9ubuntu2 5.3.0-42-generic > GNU/Linux Mint 19.1 7.5.0-3ubuntu1... 4.15.0-91-generic > GNU/Linux Mint 19.3 7.5.0-3ubuntu1... 5.3.0-42-generic > > So, compilation errors from the sources are gone (although I have an > issue of my own with GNU/Linux Mint 19.1 and libwebkitgtk... I have > to look at both libwbkitgtk and libwebkit2gtk across multiple > distributions... may need an extra package installed). > > I hope to return to meaningfully addressing warnings at some point. > > ---- > > More general question: Can the package release cycle be a little more > iterative?: > > Proposal 1: Use e.g. -beta1, -beta2, -rc1, -rc2, <release>; > > Proposal 2: Use sub-numbering more freely (e.g. iup-3.28, > iup-3.28.1, iup-3.28.2...) with the sub-numbers devoted > to bugfixes only? This would ease the pressure that's > currently being applied to the Subversion repository > head for each package; and/or > > Proposal 3: (Perhaps in combination with Prop. 2): Use Subversion's > branch/merge facilities to confine new features to > branches, merge bugfixes from the head to the branch(es), > and combine the branches with the mainline as a precursor > to announcing a "beta" release? > > I acknowledge that each of these proposals places more strain on project > management, which is an issue as it's not trivial to clone Antonio, but > not just the "iup_flattree.c" case, but also the previous > "EXCLUDE_TARGETS=iupvled" demonstrates the fragility of the current > regime. > > ---- > > In any case, I am complaining about problems in a rich and valuable > resource. I certainly wish to acknowledge the continuing value of these > packages; my hope is that my contribution(s) can be helpful. > > ---- > > cheers, > > sur-behoffski (Brenton Hoff) > programmer, Grouse Software > > > _______________________________________________ > Iup-users mailing list > Iup-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/iup-users >
_______________________________________________ Iup-users mailing list Iup-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/iup-users
