On Tue, 2013-10-22 at 14:34 +0200, Éloi Rivard wrote: > https://gist.github.com/azmeuk/7099818 > > The full log tells approximately the same thing. CONFIG_SHELL var > seems to be lacking.
Not a barf I have seen on gub. But gub is notoriously difficult to get working, the LilyPond folk themselves created a distro for running gub in as I understand it. Jeremiah has it working in Debian stable, though when I tried with my version of Debian (nominally also Debian stable) it failed, and when I tried creating a new Debian stable in a virtual machine - so that there would be no extraneous packages installed - it also failed. What O/S is the automated build using? Richard > > > > 2013/10/22 Richard Shann <[email protected]> > On Tue, 2013-10-22 at 11:53 +0200, Éloi Rivard wrote: > > I have another issue running gub > > > > Command barfed: > > > cd /home/eloi/jjbenhamgub/target/mingw/src/lilypondcairo-2.13.54 > > && ./smart-autogen.sh --noconfigure > > > > Tail of target/mingw/log/lilypondcairo.log >>>>>>>> > > + unset CDPATH > > > /home/eloi/jjbenhamgub/target/tools/root/usr/bin/autoconf: > line > > 188: CONFIG_SHELL: unbound variable > > + exit 1 > > Command barfed: > > > cd /home/eloi/jjbenhamgub/target/mingw/src/lilypondcairo-2.13.54 > > && ./smart-autogen.sh --noconfigure > > <<<<<<<< Tail of target/mingw/log/lilypondcairo.log > > > > *** Failed target: mingw::lilypondcairo > > > > Do you have an idea ? > > when you get these barfs in GUB the tail of the log as printed > out is > rarely enough to find the problem, you have to open the log > file > referred to and look back from the bottom for the error. > > Richard > > > > > > > > > 2013/10/21 Jeremiah Benham <[email protected]> > > Gub compiles its own mingw I don't think it is > required. > > > > Jeremiah > > > > On Oct 21, 2013 11:18 AM, "Richard Shann" > > <[email protected]> wrote: > > On Mon, 2013-10-21 at 11:12 -0500, Jeremiah > Benham > > wrote: > > > You need to install texlive-metapost. > > > > That seems to be what Eloi has > > > > > > mingw_dependencies() > > { > > sudo apt-get install intltool > > binutils-mingw-w64-i686 > > gcc-mingw-w64-i686 g++-mingw-w64-i686 wine > > texlive-metapost > > } > > > > in the reference > > > https://github.com/azmeuk/denemo/blob/travis/tools/travis/install.sh > he > > sent > > > > Richard > > > > > > > > Jeremiah > > > > > > On Oct 21, 2013 9:56 AM, "Éloi Rivard" > > <[email protected]> wrote: > > > mf and mpost are not present and > this makes > > my build fail. > > > > > > http://denemo.org/~rshann/gub/downloads/mpost/ > > > > http://denemo.org/~rshann/gub/downloads/mf/ > > > > > > https://travis-ci.org/azmeuk/denemo/jobs/12722937#L1576 > > > > > > > > > > > > 2013/10/20 Jeremiah Benham > > <[email protected]> > > > > > > > > > > > > On Thu, Oct 17, 2013 at > 5:22 PM, > > Éloi Rivard > > > <[email protected]> wrote: > > > Hi, > > > > > > I recently > discovered > > continuous integration > > > tools, and this is > great! > > The principle is to > > > build a project > and run some > > tests in > > > different > environments at > > each commit > > > (continuously). > Each build > > is made from a > > > clean new distro > image. You > > prepare some > > > scripts to install > the > > dependencies, and some > > > other scripts to > build the > > project and run > > > some other tests. > Then you > > can check the > > > result somewhere, > on a > > webpage for example. > > > > > > A lot of services > exist: > > Travis, Jenkins, > > > Wercker etc. > > > > > > > > > This is great to > easily spot > > regressions on > > > projects that can > run on > > several environments, > > > such as denemo. > > > > > > > > > > > > I played a bit > with > > Travis-ci and denemo: > > > > > https://travis-ci.org/azmeuk/denemo . It is > > > synchronized with > my github > > account: > > > > > https://github.com/azmeuk/denemo > > > > > > For the moment the > only test > > done is > > > compilation, but > it is very > > flexible and it is > > > possible to run > some unit > > tests. > > > Unfortunately, > there is no > > test suite in > > > denemo :). > However, a good > > non-regression test > > > could be to launch > the > > program, check that > > > commands are > correctly > > launched for instance, > > > and close it just > after. > > > > > > It works great > with gcc and > > clang, and fails > > > with mingw as I > did not > > prepared a good script > > > to run. Jeremiah, > if you > > want to help me > > > there, you are > welcome :) > > Your gub fork fail > > > to download > ImageMagick. > > > > > > > > > > > > > > > > > > This sounds interesting. I > am not > > sure how much I can > > > help because I have very > little time > > to work on this > > > stuff. I could potentially > install > > jenkins on > > > denemo.org but I am not > sure what > > the size > > > requirements would be. I > will give > > it a thought after > > > checking out their website > and > > perhaps some docs. That > > > is unless the travis-ci > thing is > > better suited to our > > > needs. > > > > > > > > > Jeremiah > > > > > > > > > > > > > > > > > > The main drawback > here is > > that Travis seem to > > > only works with > Github. > > Other repositories are > > > available if you > pay. To use > > continuous > > > integration with > denemo, it > > is either needed > > > to use another > service or > > copy each revision > > > into a github > repository. > > Create a > > > synchronized > github clone of > > the savannah > > > repository can > easily be > > done. Jenkins for > > > instance is a free > ci > > system, but to use > > > Jenkins, a > dedicated host is > > needed. > > > > > > > > > Well, I would like > to know > > your thoughts about > > > this. Do you like > the idea ? > > Which service do > > > you prefer etc. > Should I > > commit my Travis > > > scripts to master > if someone > > is interested ? > > > > > > > > > > > > > > > _______________________________________________ > > > Denemo-devel > mailing list > > > > [email protected] > > > > > > https://lists.gnu.org/mailman/listinfo/denemo-devel > > > > > > > > > > > > > > > > > > > > > > > > -- > > > Éloi Rivard - [email protected] > > > > > > « On perd plus à être indécis qu'à > se > > tromper. » > > > > > > > _______________________________________________ > > > Denemo-devel mailing list > > > [email protected] > > > > https://lists.gnu.org/mailman/listinfo/denemo-devel > > > > > > > > > > > > -- > > Éloi Rivard - [email protected] > > > > « On perd plus à être indécis qu'à se tromper. » > > > > > > > > > -- > Éloi Rivard - [email protected] > > « On perd plus à être indécis qu'à se tromper. » > _______________________________________________ Denemo-devel mailing list [email protected] https://lists.gnu.org/mailman/listinfo/denemo-devel
