Re: FreeMedForms new upstream
Hi Eric,
On Mon, Jul 08, 2013 at 06:24:02PM +0200, Eric Maeker wrote:
> > ???? There is no need at all to install a testing machine at all! The
> > only reason why I wrote that I was building under testing was the hint
> > that it was not done under pbuilder as *any* build should be done. Can
> > you get the package builded with pdebuild?
>
> Yes, but pbuilder is toooooo slow on my VBox (> 6 hours for *one* build).
Ups, that's *really* slow. You might gain some spead with cowbuilder
(that uses copy on write to create the chroot system).
> So I did try to install a dual boot with debian testing/OsX but I went through big problems... Now I only succeeded to have a debian testing with xfce running without crashing.
I can not comment on this because it is to rare information.
> But I did not succeed with the SVN+SSH configuration. Why??? I miss time to explore.
As I wasked in my previous mail, please try to ssh into svn.d.o. It is
really hard to tall why it fails on your side without doing these simple
tests.
> > I currently can't update my pbuilder environment (from Sao Paolo airport)
>
> Yes, your Qt4 packages was the latests. May be we need to add
>
> qt4-default
>
> to build-dep to get around this issue? Not sure.
Well, your latest change was to add libopencv-objdetect-dev as
Build-Depends which solved the problem I was observing on my side in the
first place. So it somehow seems settled.
> > What *exactly* is your issue? I can only recommend
> >
> > http://wiki.debian.org/Alioth/SSH
>
> I read this. May be I doing something wrong... But what?
I can only comment on this if you tried the step by step advise and
report accordingly.
> > r14082 | ericmaeker-guest | 2013-07-05 00:57:42 +0200 (Fr, 05. Jul 2013) | 1 Zeile
> >
> > has solved the first stumbling stone and the compilation was finished
> > successfully.
>
> Ok you can test my very latest commit, it should work.
I checked your last commit and also noticed the enhanced d/watch file.
However, the current download page does not contain the version you
injected in d/changelog. Please note that I reverted the upstream
version to what is downloadable for public. I also have set the target
distribution to UNRELEASED - please also do it this way until the
package is actually uploaded.
I did another change that uses xz compression in the resulting binaries
- seems to be a good idea considering their size.
What is confusing me a bit is the fact that the build keeps on working
under my 'testing' system but fails again in pbuilder with the problem
I reported initially. I will try to track down this on my side first
because Thorsten was obviously able to build the package.
> >> - understand the subtleties of the rpath/dh_shlibs as freemedforms
> >> does not provide any public libraries
> >
> > Well, in one of your previous mails you wrote that you have done some
> > experiments with RPATH which failed. I have the feeling that your last
> > commit was somewhere in the middle of your experiments. IMHO it would
> > be sufficient to revert your change regarding RPATH and simply keep on
> > delivering the packages including the RPATH warning. So my test build
> > was just intended to give you an idea about the problem your latest
> > commit might have caused.
>
> Yes I commit a draft code that I've reverted later.
So this explains why it failed on my side.
> > Ahh, it is about the LSM presentation, right? How many time is left?
> > As I said I can not do anything the next 24 hours but I might have a
> > look into what has changed on Tuesday once I might be back to normal
> > operation.
>
> I'm presenting FreeMedForms on the 11th july conference + workshop.
OK. You should be fine with the package that was uploaded to
experimental by Thorsten. Any other polishing should be done without
time pressure to finally polish everything that deserves beeing solved.
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: