Pkg xen devel - Jan 2016 - Bug#805508: Bug#805508: Bug#805508: Bug#805508: Bug#805508: xen: Unable to create HVM domUs

On Thu, 2015-11-26 at 15:16 +0000, Ian Campbell wrote:
> anything better would need to be arranged upstream.
http://xenbits.xen.org/gitweb/?p=xen.git;a=commit;h=de858271c16851d662b2613
699401df6ecec8ef8 has now landed upstream, which allows the leaf dir to be
set at configure time and therefore allows the packaging to avoid patching
some stuff.

HVM guests have been broken in Stretch for ages now, so I'm going to look
at preparing an NMU based on that upstream fix + adding
--with-libexec-leaf-dir=xen-$(VERSION) to the configure arguments. I expect
some minor adjustments to?debian/patches/config-prefix.diff will also be
required.

An early version of that (based on the not-yet-upstream version of the
patch) is in pkg-xen.git#feature/bug805508, I don't imagine the final
result will look much different.

I'll post a debdiff here as well as updating that branch.

Ian.

On Tue, 2016-01-19 at 10:39 +0000, Ian Campbell wrote:
> On Thu, 2015-11-26 at 15:16 +0000, Ian Campbell wrote:
> > anything better would need to be arranged upstream.
> 
> http://xenbits.xen.org/gitweb/?p=xen.git;a=commit;h=de858271c16851d662b26
> 13
> 699401df6ecec8ef8 has now landed upstream, which allows the leaf dir to
> be
> set at configure time and therefore allows the packaging to avoid
> patching
> some stuff.
> 
> HVM guests have been broken in Stretch for ages now, so I'm going to
look
> at preparing an NMU based on that upstream fix + adding
> --with-libexec-leaf-dir=xen-$(VERSION) to the configure arguments. I
> expect
> some minor adjustments to?debian/patches/config-prefix.diff will also be
> required.
> 
> An early version of that (based on the not-yet-upstream version of the
> patch) is in pkg-xen.git#feature/bug805508, I don't imagine the final
> result will look much different.
> 
> I'll post a debdiff here as well as updating that branch.
Now uploaded to delayed/5 (per-devref 5.11.2 since this is an important
bug, and I'm pretty sure it is actually RC but 2 days seemed too
aggressive).

Attached are full.debdiff and sanitized.debdiff, the latter removes the
noise from the automatic files (debian/control.md5sum and?debian/.git-dpm)
as well as the noise (changed From <sha>) arising from the git-dpm
checkout-patched + git-rebase for more convenient review.

I've also pushed 2022aa90fac033bc5caa439f44b0443b6399930d to
feature/bug805508 in pkg-xen.git.

I also have a signed tag in my local repo, debian/4.6.0-1+nmu1, but I
didn't push that (seeing as how this is an NMU, I wasn't sure if I
should
or not).

Ian.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: full.debdiff
Type: text/x-patch
Size: 30845 bytes
Desc: not available
URL:
<http://lists.alioth.debian.org/pipermail/pkg-xen-devel/attachments/20160119/548f66c6/attachment-0002.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: sanitized.debdiff
Type: text/x-patch
Size: 10922 bytes
Desc: not available
URL:
<http://lists.alioth.debian.org/pipermail/pkg-xen-devel/attachments/20160119/548f66c6/attachment-0003.bin>

On Tue, 2016-01-19 at 14:51 +0000, Ian Campbell wrote:
> Now uploaded to delayed/5 (per-devref 5.11.2 since this is an important
> bug, and I'm pretty sure it is actually RC but 2 days seemed too
> aggressive).
This FTBFS because it wanted to regenerate all the control files because
debian/control.md5sum missed the "dch -r" to the changelog, this then
fails
because linux-support is not in the build deps (since it shouldn't be
needed).

I've uploaded +nmu2 with debian/control.md5sum fixed, no delay this time
since +nmu1 already at the delay.

debdiff is just:

diff -Nru xen-4.6.0/debian/changelog xen-4.6.0/debian/changelog
--- xen-4.6.0/debian/changelog	2016-01-19 14:43:55.000000000 +0000
+++ xen-4.6.0/debian/changelog	2016-02-09 16:41:18.000000000 +0000
@@ -1,3 +1,11 @@
+xen (4.6.0-1+nmu2) unstable; urgency=medium
+
+??* Ensure debian/control.md5sum is correctly updated. Fixes FTBFS of
+????4.6.0-1+nmu1 on buildds where linux-support-4.2.0-1 is not expected to be
+????installed.
+
+ -- Ian Campbell <ijc at debian.org>??Tue, 09 Feb 2016 16:41:16 +0000
+
?xen (4.6.0-1+nmu1) unstable; urgency=medium
?
???* Non-maintainer upload.
diff -Nru xen-4.6.0/debian/control.md5sum xen-4.6.0/debian/control.md5sum
--- xen-4.6.0/debian/control.md5sum	2016-01-19 14:11:33.000000000 +0000
+++ xen-4.6.0/debian/control.md5sum	2016-02-09 16:41:21.000000000 +0000
@@ -1,4 +1,4 @@
-ca3b023ce0e3d81c046ef57830e8b50f??debian/changelog
+3d446943149a2ad4e038758f77d1248d??debian/changelog
?dc7b5d9f0538e3180af4e9aff9b0bd57??debian/bin/gencontrol.py
?e8f2113c78600fc17eeae209c9fad062??debian/templates/control.main.in
?a15fa64ce6deead28d33c1581b14dba7??debian/templates/xen-hypervisor.postinst.in

[..]
> I've also pushed 2022aa90fac033bc5caa439f44b0443b6399930d to
> feature/bug805508 in pkg-xen.git.
Now added?1d629591f9e4726c587436d4b6a4c7fbfa01c568 on that.
> I also have a signed tag in my local repo, debian/4.6.0-1+nmu1, but I
> didn't push that (seeing as how this is an NMU, I wasn't sure if I
should
> or not).
Likewise +nmu2.

Ian.