Summary: | x11-wm/ctwm-4.0.3 fails tests | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Agostino Sarubbo <ago> |
Component: | Current packages | Assignee: | Jake Lilly <lillyjsm> |
Status: | CONFIRMED --- | ||
Severity: | normal | CC: | proxy-maint, sam |
Priority: | Normal | Keywords: | TESTFAILURE |
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 784059 | ||
Attachments: |
build.log
LastTest.log |
Description
Agostino Sarubbo
![]() Created attachment 707742 [details]
build.log
build log and emerge --info
Not reproducible, this bug is useless without the LastTest.log. Created attachment 709197 [details]
LastTest.log
Hit this on ppc.
2/4 Testing: cfgchk
2/4 Test: cfgchk
Command: "/bin/sh" "-c" "[ -z "$DISPLAY" ] && exit 99 ; /var/tmp/portage/x11-wm/ctwm-4.0.3/work/ctwm-4.0.3_build/ctwm --cfgchk -f /var/tmp/portage/x11-wm/ctwm-4.0.3/work/ctwm-4.0.3/system.ctwmrc"
Directory: /var/tmp/portage/x11-wm/ctwm-4.0.3/work/ctwm-4.0.3_build/tests
"cfgchk" start time: May 16 10:22 -00
Output:
----------------------------------------------------------
XPM support disabled, ignoring image xpm:background8.xpm
Failed to get fontset -adobe-helvetica-bold-r-normal--*-120-*-*-*-*-*-*,*
/var/tmp/portage/x11-wm/ctwm-4.0.3/work/ctwm-4.0.3_build/ctwm: unable to open fonts "-adobe-helvetica-bold-r-normal--*-120-*-*-*-*-*-*" or "-misc-fixed-medium-r-semicondensed--13-120-75-75-c-60-*"
<end of output>
Test time = 0.46 sec
----------------------------------------------------------
Test Failed.
"cfgchk" end time: May 16 10:22 -00
"cfgchk" time elapsed: 00:00:00
----------------------------------------------------------
Looks like REQUIRED_USE="test? ( xpm )" then, right? After some tinkering, I was able to reproduce this on amd64. I think it will fail to open the fontset depending on what locale is declared. For example, amd64 with LANG set to C.UTF8 will fail, but en_US.utf8 will pass. It seems, at the minimum, LC_CTYPE has to be configured to something other than C. Perhaps I should look into disabling this check? Could you rather look/write a patch for it, or try to export it in the ebuild? Just do grep -r LC_CTYPE in our repository to see examples for both ways. |