On 19 July 2011 at 16:34, Lucas Nussbaum wrote:
| On 19/07/11 at 09:16 -0500, Dirk Eddelbuettel wrote:
| >
| > On 19 July 2011 at 15:52, Lucas Nussbaum wrote:
| > | On 19/07/11 at 05:23 -0500, Dirk Eddelbuettel wrote:
| > | >
| > | > Hi Lucas,
| > | >
| > | > Appreciate your rebuild efforts.
| > | >
| > | > Question, though:
| > | >
| > | > On 19 July 2011 at 08:33, Lucas Nussbaum wrote:
| > | > | Source: cairodevice
| > | > | Version: 2.15-1
| > | > | Severity: serious
| > | > | Tags: wheezy sid
| > | > | User: debian...@lists.debian.org
| > | > | Usertags: qa-ftbfs-20110718 qa-ftbfs
| > | > | Justification: FTBFS on amd64
| > | > |
| > | > | Hi,
| > | > |
| > | > | During a rebuild of all packages in sid, your package failed to build
on
| > | > | amd64.
| > | > |
| > | > | Relevant part:
| > | > | > make[1]: Entering directory
`/build/cairodevice-pw6pXr/cairodevice-2.15/src'
| > | > | > make[1]: Leaving directory
`/build/cairodevice-pw6pXr/cairodevice-2.15/src'
| > | > | > installing to
/build/cairodevice-pw6pXr/cairodevice-2.15/debian/r-cran-cairodevice/usr/lib/R/site-library/cairoDevice/libs
| > | > | > ** R
| > | > | > ** demo
| > | > | > ** preparing package for lazy loading
| > | > | > ** help
| > | > | > *** installing help indices
| > | > | > ** building package indices ...
| > | > | > ** testing if installed package can be loaded
| > | > | > Error : .onLoad failed in loadNamespace() for 'cairoDevice',
details:
| > | > | > call: fun(...)
| > | > | > error: GDK display not found - please make sure X11 is running
| > | >
| > | > This has worked for many years by a Build-Depends: on xvfb, and then
using
| > | > xvfb-run.
| > | >
| > | > Could this be something local to your setup?
| > |
| > | It's possible.
| > |
| > | > | > Error: loading failed
| > | > | > Execution halted
| > | > | > ERROR: loading failed
| > | > | > * removing
'/build/cairodevice-pw6pXr/cairodevice-2.15/debian/r-cran-cairodevice/usr/lib/R/site-library/cairoDevice'
| > | > | > kill: 186: No such process
| > | >
| > | > That 'kill: 186' is very bizarre, and identical to three other bug
reports
| > | > you just sent me. Is there a chance something went wrong at your end?
| > |
| > | It's possible, again. I don't have much time to investigate. I can only
| > | say that many R packages were failing, and I didn't file bugs for all of
| > | them. Also, I run builds twice to make sure that a temporary glitch
| > | doesn't cause false positives, but it could still be the case.
| >
| > Ok, here is a little colour from me:
| >
| > -- there are now well over 100 r-cran-*, a few dozen are 'mine'
| >
| > -- I have quite some experience which "batch building" as we ran a service
| > called cran2deb once which took all of CRAN and (with some tools)
created
| > r-cran-* for each of these; at the same around 2100 or so packages (and
| > CRAN now has more than 3100)
| >
| > -- pbuilder did NOT reliably work in parallel there, we always serialized
|
| I use sbuild, not pbuilder, and usually it works fine in parallel (for
| all packages). I don't think that the problem is there, and your patch
| described below should solve the remaining problem.
|
| > -- I got a recent patch to the r-cran.mk called from debian/rules which
lets
| > xvfb-run 'rotate' over a set of ports; that should help with the
previous
| > builds
|
| Indeed
That patch was applied a few weeks ago:
r-base (2.13.1-1) unstable; urgency=low
* New upstream release
* R/m4, configure: Override bzip2 test for 1.0.6 to let our patched
1.0.5 version pass as lintian does not like the embedded library
-- Dirk Eddelbuettel <e...@debian.org> Fri, 08 Jul 2011 06:17:17 -0500
r-base (2.13.0-4) unstable; urgency=low
* debian/r-cran.mk: Make call to xvfb-run more robust by adding '-a'
option to allow automatic switching to free virtual server; with thanks
to Salvatore Bonaccorso for the suggestions (Closes: #630869)
-- Dirk Eddelbuettel <e...@debian.org> Wed, 22 Jun 2011 08:59:55 -0500
So I assume you already used it?
| > -- I cannot offer other help besides saying ... "in my chroot it all works"
| >
| > What can we do to improve things?
|
| I have no idea. Could you diff your log with mine to see if something
| comes up?
There is nothing there. In a sense, R copied Debian and standardised: For
any given package foo, you just do
R CMD INSTALL foo
pointing at directory foo/ or a tarball. It then runs configure etc if
needed, builds a Makefile from stanza, ... So your logs always fail at the
same step: that one R CMD INSTALL call I make from r-cran.mk (after setting a
few vars on our Debian end).
So nuttin' to go by I fear.
Dirk
|
| Lucas
--
Gauss once played himself in a zero-sum game and won $50.
-- #11 at http://www.gaussfacts.com
--
To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
