Re: Any progress with FIS GT.M?
Hi Yaroslav,
thanks for the very interesting summary. It seems we are closer than
ever to GT.M packages and whatever might come first - debian freeze or
fis-gtm upload - it is a very important step which brings Debian Med
forward.
Many thanks for your and all the other people working on this effort
Andreas.
On Wed, Jun 13, 2012 at 08:26:17PM -0400, Yaroslav Halchenko wrote:
> so -- here is a brief summary/status while kitware+GT.M folks are
> finishing cmakification of GT.M's build system. I have pushed my
> changes for debian/ (summarized below) to
> https://github.com/yarikoptic/fis-gtm/tree/debian
>
> My main concern was that upstream's way to deploy GT.M requires
> installation time building of some .m files, which they suggested
> to do at .postinst time... which obviously I hoped/argued to avoid.
>
> With the manual tweaking 'stage2' installation I have tested, GM.T seems
> to work even if we move the installation directory into the 'deployment
> path' -- it just gives full 'build path' if a corresponding *.m file is
> not found which is imho a minor issue. So it is a good sign although it
> would require to run at least some reasonable test battery to say that
> it works with some assurance. After cmakification gets ironed out so
> built binaries work without segfaulting ;) then we can simply do 'stage
> 2' building/installation (gmtinstall call) at package build time thus
> achieving the desired behavior (there is commented out draft of
> such call in debian/rules atm... it just needs also version
> specification for ICU library... yeah).
>
> changes:
> [ Yaroslav Halchenko ]
> * debian/control:
> - boosted policy and added few -dev packages we know to be needed
> - removed unused ATM adduser, ucf, debconf from *Depends
> - provide both meta and versioned fis-gtm packages
> - boosted policy to 3.9.3
> - boosted dh compat 9 to get hardening flags automagically
> * debian/copyright:
> - adjusted for DEP5
> * debian/rules:
> - disable call to gtminstall at package build time:
> upstream relies on the installdir path internally so "$test" command
> used in various tools would not work rendering system broken according
> to upstream
> - generate lintian-overrides for shlib-with-executable-stack
> - provide dumped from web changelog for the most recent release
>
> Cheers and I am hitting the road back now
>
>
> On Wed, 13 Jun 2012, Luis Ibanez wrote:
>
> > Hi Andreas,
>
> > Good news,
> > the plan for the hackathon today is unfolding nicely:
>
> > [1]https://github.com/luisibanez/fis-gtm/wiki/Debian-Packaging-Hackathon
>
> > We will be starting at 9am EST, probably focusing first
> > on the fisgtm+cmake portion of the agenda, and we
> > anticipate that the afternoon will be focused on the
> > Debian packaging.
>
> > Yaroslav will be joining us in the afternoon,
> > so that should work as a nice timing.
>
> > Should we run into any problems we will be looking
> > for help in this list, or the mentors list, or IRC.
>
> > Thanks
>
> > Luis
>
> > References
>
> > Visible links
> > 1. https://github.com/luisibanez/fis-gtm/wiki/Debian-Packaging-Hackathon
>
> --
> Yaroslav O. Halchenko
> Postdoctoral Fellow, Department of Psychological and Brain Sciences
> Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
> Phone: +1 (603) 646-9834 Fax: +1 (603) 646-1419
> WWW: http://www.linkedin.com/in/yarik
>
>
> --
> To UNSUBSCRIBE, email to debian-med-REQUEST@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
> Archive: [🔎] 20120614002617.GD15236@onerussian.com">http://lists.debian.org/[🔎] 20120614002617.GD15236@onerussian.com
>
>
--
http://fam-tille.de
Reply to: