looks like events.o lacks depend on conio_cmds.h, so building in parallel fails: root@G5[ppc] 0:0 src # make events.o powerpc-unknown-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I. -I../include -I../libfiretalk -I../libltdl -Wall -O2 -mcpu=970 -pipe -g -c events.c events.c:14:24: error: conio_cmds.h: No such file or directory
Created attachment 155285 [details, diff] Parallel build patch Ebuild patch coming...
Created attachment 155287 [details, diff] Ebuild patch Works with MAKEOPTS="-j15", whereas without these patches it dies.
shouldnt it also declare ${srcdir}/events.c ? and i dont think conio_cmds.h lives in srcdir, so you'll need to drop the prefix on that
Created attachment 155367 [details, diff] Updated patch With regards to your first question, my bad. Fixed in the new patch. As to your concern that conio_cmds.h doesn't live in ${srcdir}, negative. naim-xxx/src/Makefile/ defines srcdir to equal '.'
that isnt how srcdir works. you configured in the source tree, so srcdir and builddir will be the same ("."). if the header is a generated file, then a srcdir prefix is incorrect. you could use ${builddir}/, but i'm not aware of any setup where builddir != pwd and so most code ive seen assumes this.
*** Bug 240758 has been marked as a duplicate of this bug. ***
*** Bug 273716 has been marked as a duplicate of this bug. ***
Could the pending removal flag be removed from this package? Also is simply disabling the -j switch on make for this package an option, especially since it is rather small.
(In reply to comment #8) > Could the pending removal flag be removed from this package? > Also is simply disabling the -j switch on make for this package an option, > especially since it is rather small. > Maybe, but what about bugs 176276 and 205492 (The depending ones.) ? They've been open since ages and noone seems to care.
Created attachment 204990 [details, diff] overrides -j in MAKEOPTS with -j 1
I just posted a patch to override the -j switch in MAKEOPTS. The two bugs you mentioned are for the older 11.7.2 version of naim, which was released in 2004. I realize this is the version that both of the 11.8 versions are keyword masked, but 11.7.2 could be dropped keeping the other two. > > Maybe, but what about bugs 176276 and 205492 (The depending ones.) ? They've > been open since ages and noone seems to care. >
Comment on attachment 204990 [details, diff] overrides -j in MAKEOPTS with -j 1 Unmasked naim in tree and forced -j1. Leaving this bug open, because it's not a real solution. There's also version bump available.
Removed from tree.