On Wed, Mar 06, 2013 at 06:16:25AM -0600, Dirk Eddelbuettel wrote: > > On 6 March 2013 at 15:52, Boris Pek wrote: > | Package: r-base-core > | Version: 2.15.2-1 > | Severity: normal > | > | Hi, > | > | elmerfem is FTBFS on arm* again because of recent changes in r-base. > | > | After small investigation [0] I found that elmerfem: > | * was built successfully with r-base-core: 2.15.1-4, 2.15.1-5 > | * was FTBFS with r-base-core: 2.15.1-2, 2.15.3~20130327-1, 2.15.3-1 > | > | After reading change log [1] and bug report #679180 I think that this change > | in r-base 2.15.2-1 caused a regression: > | * src/library/tools/R/install.R: Pass "perl = TRUE" parameter to sub() > | function when defining SHLIB_* macros to work-around an erroneus > | string replacement which causes build failures on arm* architectures > | but make the change conditonal on actually being on arm*. > | > | Could you fix it? > > No, not really. It is an arm-specific failure that does not happen anywhere > else. > > I think this is a side-effect of #695411. CCing Johannes who worked on this.
Yes, the build failure appears to occur during the R package installation, as sub() is misbehaving on ARM. In the R bug tracking system [2] there was a comment that this is due to some change in the tre version shipped with R. > Is there a willing and able ARM* porter within Debian who can help? Yes, this would be great. Although it seems to be a problem with the tre version shipped with R in combination wih the build environment of wheezy. sub() behaved correctly when R 2.15.2 was compiled on squeeze. Should R not be linked to the Debian tre package [3] anyway? No idea how this would be done though. Johannes [2] https://bugs.r-project.org/bugzilla3/show_bug.cgi?id=15087 [3] http://packages.debian.org/source/sid/tre > > Dirk > > | > | [0] https://buildd.debian.org/status/logs.php?pkg=elmerfem&arch=armel > | [1] > http://packages.debian.org/changelogs/pool/main/r/r-base/current/changelog > | > | elmerfem 6.1.0.svn.5396.dfsg2-2 was prepared for Wheezy and this FTBFS is > only > | one stopper now. > | > | Best regards, > | Boris > > -- > Dirk Eddelbuettel | e...@debian.org | http://dirk.eddelbuettel.com -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
