| Submitter | Richard Purdie |
|---|---|
| Date | Sept. 18, 2012, 10:32 a.m. |
| Message ID | <1347964372.13596.110.camel@ted> |
| Download | mbox | patch |
| Permalink | /patch/36729/ |
| State | Accepted |
| Commit | e21b6c04e512a3bc2339a20045b7041f1d26e859 |
| Headers | show |
Comments
Richard Purdie
<richard.purdie-hQyY1W1yCW8ekmWlsbkhG0B+6BGkLq7r@public.gmane.org>
writes:
> +STAMPCLEAN = "${TMPDIR}/stamps/${MULTIMACH_TARGET_SYS}/${PN}-*-*"
Is it intended that this pattern can match unrelated packages? E.g. for
PN='gcc-cross', the mask will apply to stampfiles for 'gcc-cross-initial'
too.
Enrico
On Thu, 2012-09-20 at 20:03 +0200, Enrico Scholz wrote: > > Richard Purdie > <richard.purdie-hQyY1W1yCW8ekmWlsbkhG0B+6BGkLq7r@public.gmane.org> > writes: > > > +STAMPCLEAN = "${TMPDIR}/stamps/${MULTIMACH_TARGET_SYS}/${PN}-*-*" > > Is it intended that this pattern can match unrelated packages? E.g. for > PN='gcc-cross', the mask will apply to stampfiles for 'gcc-cross-initial' > too. No, this is a problem :( I saw this last night and have been giving it some thought. We might have to put a separator into the stamp file format to make this easier to deal with. The downside is that is a TMPDIR layout change :/. The thing is the bug this solves is serious even if people only rarely hit it and I can't think of a better solution... Cheers, Richard
On Fri, 2012-09-21 at 10:36 +0100, Richard Purdie wrote: > On Thu, 2012-09-20 at 20:03 +0200, Enrico Scholz wrote: > > > > Richard Purdie > > <richard.purdie-hQyY1W1yCW8ekmWlsbkhG0B+6BGkLq7r@public.gmane.org> > > writes: > > > > > +STAMPCLEAN = "${TMPDIR}/stamps/${MULTIMACH_TARGET_SYS}/${PN}-*-*" > > > > Is it intended that this pattern can match unrelated packages? E.g. for > > PN='gcc-cross', the mask will apply to stampfiles for 'gcc-cross-initial' > > too. > > No, this is a problem :( > > I saw this last night and have been giving it some thought. We might > have to put a separator into the stamp file format to make this easier > to deal with. The downside is that is a TMPDIR layout change :/. > > The thing is the bug this solves is serious even if people only rarely > hit it and I can't think of a better solution... I've given this some more thought and there is an easier way out if we assume PV starts with something numeric. Since the current situation is rather problematic (I didn't see it locally due to another patch I have), I've pushed a fix: http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=3c101af5e33b21f53171c66de0c2c3e797d95247 I'm going to post some RFCs on other ideas related to this. Cheers, Richard
On Fri, Sep 21, 2012 at 6:22 AM, Richard Purdie <richard.purdie@linuxfoundation.org> wrote: > > I've given this some more thought and there is an easier way out if we > assume PV starts with something numeric is this assumption true always I wonder
Patch
diff --git a/meta/conf/bitbake.conf b/meta/conf/bitbake.conf index 9c51630..b3d3f52 100644 --- a/meta/conf/bitbake.conf +++ b/meta/conf/bitbake.conf @@ -332,7 +332,8 @@ GITDIR = "${CO_DIR}/git2" BZRDIR = "${CO_DIR}/bzr" HGDIR = "${CO_DIR}/hg" -STAMP = "${TMPDIR}/stamps/${MULTIMACH_TARGET_SYS}/${PF}" +STAMP = "${TMPDIR}/stamps/${MULTIMACH_TARGET_SYS}/${PN}-${EXTENDPE}${PV}-${PR}" +STAMPCLEAN = "${TMPDIR}/stamps/${MULTIMACH_TARGET_SYS}/${PN}-*-*" WORKDIR = "${TMPDIR}/work/${MULTIMACH_TARGET_SYS}/${PF}" T = "${WORKDIR}/temp" D = "${WORKDIR}/image" diff --git a/meta/recipes-devtools/gcc/gcc-common.inc b/meta/recipes-devtools/gcc/gcc-common.inc index 72180db..4ef2a1b 100644 --- a/meta/recipes-devtools/gcc/gcc-common.inc +++ b/meta/recipes-devtools/gcc/gcc-common.inc @@ -63,6 +63,10 @@ SS = "${TMPDIR}/stamps/work-shared/gcc-${PV}-${PR}" do_fetch[stamp-base] = "${SS}" do_unpack[stamp-base] = "${SS}" do_patch[stamp-base] = "${SS}" +SSCLEAN = "${TMPDIR}/stamps/work-shared/gcc-*-*" +do_fetch[stamp-base-clean] = "${SSCLEAN}" +do_unpack[stamp-base-clean] = "${SSCLEAN}" +do_patch[stamp-base-clean] = "${SSCLEAN}" # SW means Shared Work directory SW = "${TMPDIR}/work-shared/gcc-${PV}-${PR}"
This takes advantage of new bitbake functionality to clean up stale stamp files when creating new stamp files. [YOCTO #2961] Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> ---