Summary: | dev-lang/R-2.14.0 - Error in textConnection("rval", "w", local = TRUE) : unsupported mode | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Jean-Francois Ostiguy <jfostiguy> |
Component: | [OLD] Development | Assignee: | Gentoo Science Mathematics related packages <sci-mathematics> |
Status: | RESOLVED INVALID | ||
Severity: | normal | ||
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | x86 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
emerge log
emerge --info output |
Description
Jean-Francois Ostiguy
2011-11-24 01:16:42 UTC
1) Please post your `emerge --info' output. 2) Please attach the entire build log to this bug report. Created attachment 293933 [details]
emerge log
Created attachment 293935 [details]
emerge --info output
Please try with just committed R-2.14.1. I can not reproduce it. Try other locale as well. Thanks (In reply to comment #4) > Please try with just committed R-2.14.1. I can not reproduce it. Try other > locale as well. > > Thanks Unfortunately, R-2.14.1 fails in the same manner. My locale is defined by LANG="en_US.utf8"; surely this is a fairly common setting. I am not sure what other locale I should be trying. BTW: I also have LINGUAS defined in make.conf as follows: LINGUAS="en en_US fr" I tried removing "fr". This did not help. I also tried USE=-doc Same story. The strange thing is that R-2.13.1-r1 emerged just fine. I would try emerging it again, just as a test, but the ebuild has apparently been removed. Resolution: I was about to give up on this "bug". My home machine is an x86 box; I have also access to another box, running gentoo / amd64. Since I did not have any issue with R-2.14.1 on the amd64 box, I decided, in desperation, to compare the build log files between the 2 machines. It turns out that I had an extra CFLAG on the x86 box : "-fno-builtin-strlen", leftover inadvertently from some testing I did a few weeks ago. In principle, all this should do is inhibit inlining optimization of the strlen libc function i.e. it should just use a regular call for strlen. Why this should make the R textConnection function fail is beyond me, but it does ;-( Bottom line: the problem "solved". Sebastien: Thank you for your help. |