Summary: | dev-libs/newt-0.50.35-r4 does not build | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | John Nilsson <john> |
Component: | [OLD] Library | Assignee: | Robin Johnson <robbat2> |
Status: | RESOLVED DUPLICATE | ||
Severity: | normal | ||
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: | MAKEOPTS="-j1" emerge newt --oneshot &> newtbuild |
Description
John Nilsson
2005-02-25 15:06:45 UTC
Created attachment 52158 [details]
MAKEOPTS="-j1" emerge newt --oneshot &> newtbuild
The output from the build
something is badly wrong with your python: "for ver in python.3 ; do \" this is the output on my machine: "for ver in python2.3 ; do \" I reemerged python. It didn't help. could you open up newt-0.50.35-r4.ebuild, and after the call to 'python_version' please insert this line: einfo "python_version: '$PYVER_ALL' '$PYVER_MAJOR' '$PYVER_MINOR' '$PYVER_MICRO' '$PYVER'" and run the emerge again, then include the output from the above line? I added: einfo "python_version: '$PYVER_ALL' '$PYVER_MAJOR' '$PYVER_MINOR' '$PYVER_MICRO' '$PYVER'" env | grep PY to the ebuild and env | grep PY as the last statement in python_version() from python.eclass and got: PYVER_ALL=2.3.4 PYVER=.3 PYVER_MINOR=3 PYVER_MICRO=4 PYTHON_DONTCOMPILE=1 PYVER_MAJOR= * python_version: '2.3.4' '' '3' '4' '.3' PYVER_ALL=2.3.4 PYVER=.3 PYVER_MINOR=3 PYVER_MICRO=4 PYTHON_DONTCOMPILE=1 PYVER_MAJOR= however: minibox etc # python -V Python 2.3.4 minibox etc # tmpstr="$(python -V 2>&1 )"; echo ${tmpstr#Python } | cut -d. -f1 2 minibox etc # tmpstr="$(python -V 2>&1 )"; echo ${tmpstr#Python } | cut -d. -f2 3 minibox etc # tmpstr="$(python -V 2>&1 )"; echo ${tmpstr#Python } | cut -d. -f3 4 now this is strange... it seems, for some reson way beyond me, that "cut -d. -f1" does not work in that particular context it is executed in while emerging newt... emerge metakit yields similar results, which at least rules out the newt ebuild. It's completley beyond me why cut wouldn't work in that eclass... replacing cut -d. -f1 in the eclass with simply cat does print out 2.3.4 so the pipe works. If it was consistent and wouldn't work in the shell I could atleast suspect some compiler bug. ah, found the problem. it's cut being broken when compiled with -Os *** This bug has been marked as a duplicate of 81031 *** So I was close =) |