Bug#1036263: guestfs-tools: FTBFS in testing: make[6]: *** [Makefile:1716: test-suite.log] Error 1

2023-05-27 Thread Martin Hostettler
tags -1 + unreproducible
thanks

On Thu, 25 May 2023 21:00:18 +0200 Bastian Germann  wrote:
> I cannot reproduce this.
> 

I can't reproduce this either.

I tried 2 variants and both build fine (using an up to date sid VM to
run the build using bookworm docker container or pbuilder).

docker run -it debian:bookworm /bin/bash
apt install devscripts build-essential aptitude pbuilder
dget 
https://deb.debian.org/debian/pool/main/g/guestfs-tools/guestfs-tools_1.48.2-1.dsc
/usr/lib/pbuilder/pbuilder-satisfydepends
debuild -uc -us -j4

Variant B
docker run --privileged -it debian:bookworm /bin/bash
apt install devscripts pbuilder
pbuilder create --distribution bookworm
dget 
https://deb.debian.org/debian/pool/main/g/guestfs-tools/guestfs-tools_1.48.2-1.dsc
pbuilder build --distribution bookworm guestfs-tools_1.48.2-1.dsc 



Bug#1036021: cadabra2,python3-notebook: undeclared file conflict on /usr/lib/python3/dist-packages/notebook/static/components/codemirror

2023-05-26 Thread Martin Hostettler
On Thu, May 25, 2023 at 09:52:55PM +0200, Gürkan Myczko wrote:
> On 25.05.2023 19:47, Martin Hostettler wrote:
> > On 19/05/2023 12.48, Martin Hostettler wrote:
> > > It seems what is needed is this:
> > > 
> > > mv 
> > > debian/cadabra2/usr/lib/python3.*/site-packages/notebook/static/components
> > > debian/cadabra2/usr/share/javascript
> > > rmdir
> > > debian/cadabra2/usr/lib/python3.*/site-packages/notebook/static
> > > rmdir debian/cadabra2/usr/lib/python3.*/site-packages/notebook
> > > 
> > > (The wildcards are used to avoid hard coding the current debian
> > > default
> > > python version)
> > 
> > Any chance to get this changes into an upload? The deadline for last
> > unblock requests is quickly coming.
> 
> I'm sorry, I'm busy with other things, if you feel like it please go ahead
> add yourself to
> Uploaders and upload. Or NMU, as you like. I am also fine reviewing changes,
> and uploading
> via mentors.d.n.
> 

I don't have a gpg key, so i think mentors.d.n is not an option but i
uploaded to some temporary web location. So feel free to review from there.
Also feel free to do what makes the most sense to you with the changelog
entry and version.

https://uchuujin.de/tmp/cadabra/cadabra2_2.4.3.2-0.2.dsc

I'm also attaching the intended patch, in case that is easier to review.

 - Martin
diff -ru orig/cadabra2-2.4.3.2/debian/changelog patched/cadabra2-2.4.3.2/debian/changelog
--- orig/cadabra2-2.4.3.2/debian/changelog	2023-02-06 22:21:08.0 +
+++ patched/cadabra2-2.4.3.2/debian/changelog	2023-05-26 18:17:02.306980093 +
@@ -1,3 +1,13 @@
+cadabra2 (2.4.3.2-0.2) unstable; urgency=medium
+
+  * Non-maintainer upload.
+  * Move codemirror integration to canonical path to
+avoid symlink/directory conflict on
+/usr/lib/python3/dist-packages/notebook/static/components/codemirror
+(Closes: 1036021)
+
+ -- Martin Hostettler   Fri, 26 May 2023 06:41:29 +
+
 cadabra2 (2.4.3.2-0.1) unstable; urgency=medium
 
   * Non-maintainer upload.
diff -ru orig/cadabra2-2.4.3.2/debian/rules patched/cadabra2-2.4.3.2/debian/rules
--- orig/cadabra2-2.4.3.2/debian/rules	2022-10-13 07:06:04.0 +
+++ patched/cadabra2-2.4.3.2/debian/rules	2023-05-26 06:41:08.613257764 +
@@ -27,6 +27,9 @@
 	dh_auto_install
 	sed -i s,python3.10,python3,g debian/cadabra2/usr/bin/cadabra2
 	sed -i s,site-,dist-,g debian/cadabra2/usr/bin/cadabra2
+	mv debian/cadabra2/usr/lib/python3.*/site-packages/notebook/static/components debian/cadabra2/usr/share/javascript
+	rmdir debian/cadabra2/usr/lib/python3.*/site-packages/notebook/static
+	rmdir debian/cadabra2/usr/lib/python3.*/site-packages/notebook
 
 override_dh_shlibdeps:
 	@echo test


Bug#1034824: tomcat9 should not be released with Bookworm

2023-05-25 Thread Martin Hostettler
I was asked to send a update to this bug from my notes/open tabs.

>From what i can see this is still a problem and it is getting very late
to fix all the fallout.

There are still 2 packages that are not fixed for this.

src:trapperkeeper-webserver-jetty9-clojure (#1036250) which is as far as
I understand a dependency of puppet, which is used by a lot of admins
including Debian's own DSA.

Which even after trying to fix the build problem left the package in a
state where it the whole logging is non functional: 

Quoting from J�r�me Charaoui in (#1036250):
> I did further tests with puppetserver, which is a downstream dependency 
> of trapperkeeper-webserver-jetty9-clojure and unfortunately, the web 
> requests (access) logging remains broken. There are no warnings or error 
> messages anywhere: as you can imagine, the logging events are simply 
> lost in the ether.

I'm not sure if the latest patches from 2023-05-22 do fix those, but there
was no follow up on the bug with details.

Then there is src:tomcatjss (1031816) which seems to have zero progress
since the bug was filed.

This is a dependency of at least dogtag-pki, pki-ca, pki-kra, pki-ocsp,
pki-server, pki-tks and pki-tps

I'm not sure what the actual state of src:logback is. It seems the problems
in trapperkeeper-webserver-jetty9-clojure are partially related to the
state of logback.
Do we know that it properly works with the tomcat10 migration patchset?
Logback seems to have quite a few reverse dependencies as well.


Some bugs have according to the bts been fixed and migrated meanwile:

#1035995: bazel-bootstrap
#1011597: tiles
#1033366: resteasy3.0

What is the plan here to get this in shape for in time before last unblock
requests for bookworm on the 28th?

 - Martin



Bug#1036021: cadabra2,python3-notebook: undeclared file conflict on /usr/lib/python3/dist-packages/notebook/static/components/codemirror

2023-05-25 Thread Martin Hostettler
On 19/05/2023 12.48, Martin Hostettler wrote:
> It seems what is needed is this:
> 
> mv 
> debian/cadabra2/usr/lib/python3.*/site-packages/notebook/static/components 
> debian/cadabra2/usr/share/javascript
> rmdir debian/cadabra2/usr/lib/python3.*/site-packages/notebook/static
> rmdir debian/cadabra2/usr/lib/python3.*/site-packages/notebook
> 
> (The wildcards are used to avoid hard coding the current debian default
> python version)

Any chance to get this changes into an upload? The deadline for last
unblock requests is quickly coming.

 - Martin



Bug#1036021: cadabra2,python3-notebook: undeclared file conflict on /usr/lib/python3/dist-packages/notebook/static/components/codemirror

2023-05-22 Thread Martin Hostettler
On Mon, May 22, 2023 at 11:59:37AM +0200, Andreas Beckmann wrote:
> On 19/05/2023 12.48, Martin Hostettler wrote:
> > It seems what is needed is this:
> > 
> > mv 
> > debian/cadabra2/usr/lib/python3.*/site-packages/notebook/static/components 
> > debian/cadabra2/usr/share/javascript
> > rmdir debian/cadabra2/usr/lib/python3.*/site-packages/notebook/static
> > rmdir debian/cadabra2/usr/lib/python3.*/site-packages/notebook
> > 
> > (The wildcards are used to avoid hard coding the current debian default
> > python version)
> 
> Does this still work if there is more than one supported python version
> (during python transitions)? It's probably OK to to break support for
> multiple python versions now (as we currently only have one and stable
> releases only have one), but this may need to be revisited for bookworm+1.
> 

As this is in the temporary directory where the build code is installed i
would suspect this will work with more supported python versions as well.
I think this is just a case of the upstream build system including the
python version of the python version used to build the package.
(i think usually debhelper takes care of removing that, but this is before
any of that runs)

At least as long as the code there is compatible with all supported python
versions in debian.

But i don't know anything for certain, as i don't know cadabre but just
want to help solve RC bugs against packages i use (like python3-notebook)

 - Martin



Bug#1036272: youtube-dl - Should this be released with Bookworm?

2023-05-21 Thread Martin Hostettler
> Do we really need to release this package in a not so usable state?

It seems that currently mkchromecast still depends on this package. So just
removing it would remove mkchromecast as well. But i'm not sure how much
of mkchromecast works without a binary called youtube-dl.

Also some users might benefit from automatically transitioning to
youtube-dlp, on the other hand yt-dlp does not seem to provide youtube-dl
as a binary, so maybe this is unexpected as well.

 - Martin



Bug#1036021: cadabra2,python3-notebook: undeclared file conflict on /usr/lib/python3/dist-packages/notebook/static/components/codemirror

2023-05-19 Thread Martin Hostettler
On Fri, May 19, 2023 at 11:09:41AM +0200, Helmut Grohne wrote:
> On Fri, May 19, 2023 at 10:27:46AM +0200, Martin Hostettler wrote:
> 
> Do I understand correctly that all we have to do here is ship the file
> in its "canonical" path after resolving the symbolic link from
> python3-notebook? In other words, a simple file move in the data.tar?

If dpkg can handle that even with the symlink, then i think, yes.

> 
> > I have no clue how the correct way to make this switch in the debian
> > packaging is, but i hope this messages helps to confirm what is happening
> > in this situation and someone with more solid clue about these kind of
> > problems in packaging can resolve this.
> 
> It is as simple as adding a mv invocation to an
> appropriate override_dh_something to debian/rules. That's something that
> easily passes an unblock request with the release team.
> 
> Probably, appending to override_dh_auto_install is sufficient:
> 
> mv 
> debian/cadabra2/usr/lib/python3/dist-packages/notebook/static/components 
> debian/cadabra2/usr/lib/share/javascript
> rmdir debian/cadabra2/usr/lib/python3/dist-packages/notebook/static
> rmdir debian/cadabra2/usr/lib/python3/dist-packages/notebook
> 

It seems what is needed is this:

   mv 
debian/cadabra2/usr/lib/python3.*/site-packages/notebook/static/components 
debian/cadabra2/usr/share/javascript
   rmdir debian/cadabra2/usr/lib/python3.*/site-packages/notebook/static
   rmdir debian/cadabra2/usr/lib/python3.*/site-packages/notebook

(The wildcards are used to avoid hard coding the current debian default
python version)

That results in a debdiff like this:
 
Files in second .deb but not in first
-
-rw-r--r--  root/root   /usr/share/javascript/codemirror/mode/cadabra/cadabra.js

Files in first .deb but not in second
-
-rw-r--r--  root/root   
/usr/lib/python3/dist-packages/notebook/static/components/codemirror/mode/cadabra/cadabra.js


Which looks ok.

both before the patch and after the following file exists when
python3-notebook was installed before installing cadabra:

/usr/share/javascript/codemirror/mode/cadabra/cadabra.js


I just noticed that the override_dh_auto_install also contains a line not
yet updated to python 3.11

> sed -i s,python3.10,python3,g debian/cadabra2/usr/bin/cadabra2

Updateing that might be a good idea too. Switching that to 3.11 results in
the further debdiff:

Control files: lines which differ (wdiff format)

Depends: fonts-cmu, python3 (<< 3.12), python3 (>= 3.11~), [-python3.11:any,-] 
python3:any (>= 3.6~)

Which seems ok too.

@G�rkan Myczko

Could you see if that seems sane and if it is do an upload with the moved
file? (I'm not a DD, so i can't help out with that)

 - Martin



Bug#1036021: cadabra2,python3-notebook: undeclared file conflict on /usr/lib/python3/dist-packages/notebook/static/components/codemirror

2023-05-19 Thread Martin Hostettler
On Fri, 19 May 2023 08:48:46 +0200 Helmut Grohne  wrote:
> Hi,
> 
> On Thu, May 18, 2023 at 10:30:48PM +0400, Andrey Rakhmatullin wrote:
> > Looks like the file in cadabra2, let's say, provides integration between
> > it and python3-notebook. 
> 
> I'm not sure what you want to achieve here. If this is about two
> competing implementations where one is strictly superior to the other,
> then maybe dpkg-diverting the lesser one would be reasonable? The
> diversion mechanism works entirely without adding a relationship between
> the packages.
> 
> Another option is use of update-alternatives, which also allows
> specifying a preference on the packaging level while also providing a
> user choice.
> 
> I lack the context of the package at hand, so I'll leave that up to you,
> but I think there is sufficient options on the table for dealing with it
> in a good way. Please choose wisely.
> 

>From what i see the intended result is not a file replacement but adding a
file into codemirror.

If everything works well the sequence is that python3-notebook pulls in
libjs-codemirror which contains various files in
/usr/share/javascript/codemirror/

python3-notebook then ships
/usr/lib/python3/dist-packages/notebook/static/components/codemirror
as
symlink (../../../../../../share/javascript/codemirror aka
/usr/lib/share/javascript/codemirror )
pointing to the codemirror library.

Now cadabra2 wants to place an additional file (mode/cadabra/cadabra.js)
for integration into the codemirror library used by python3-notebook.

Currently it does that by using the path
/usr/lib/python3/dist-packages/notebook/static/components/codemirror/mode/cadabra/cadabra.js
in data.tar which "works" when python3-notebook was installed previously,
but is the subject of this bug because when installation order if different
this does the wrong thing.

So all the ideas of diverts and alternatives seems to be unneeded. What
seems to be needed is finding the correct way for cadabra2 to always place
the integration file directly into /usr/lib/share/javascript/codemirror
without depending on the symlink from python3-notebook that might or might
not be there.

I have no clue how the correct way to make this switch in the debian
packaging is, but i hope this messages helps to confirm what is happening
in this situation and someone with more solid clue about these kind of
problems in packaging can resolve this.

- Martin



Bug#1032553: magic-wormhole: FTBFS in testing: dh_auto_test: error: pybuild --test -i python{version} -p 3.11 returned exit code 13

2023-04-07 Thread Martin Hostettler
tags 1032553 + patch

thanks

On Wed, Apr 05, 2023 at 09:22:45AM +0200, Sascha Steinbiss wrote:
> Hi all,
> 
> [...]
> > This is mentioned in
> > https://github.com/magic-wormhole/magic-wormhole/issues/458 as likely
> > a "timing issue". Not sure if it's fixed upstream. >
> 
> Could it make sense to also patch the tests to include the delay that is
> mentioned in the GitHub issue comments?

I've tried adding a 2 second delay in the failing test and that yields a
package that builds reliably for me. I just rebuild the package with the
patch 250 times successfully in a row.

This patch works without any other commits backported.

--- magic-wormhole-0.12.0/src/wormhole/test/test_wormhole.py2021-02-05 
02:16:21.0 +
+++ magic-wormhole-0.12.0.fix/src/wormhole/test/test_wormhole.py
2023-04-07 14:44:26.792819150 +
@@ -6,6 +6,7 @@
 from twisted.internet import reactor
 from twisted.internet.defer import gatherResults, inlineCallbacks, returnValue
 from twisted.internet.error import ConnectionRefusedError
+from twisted.internet.task import deferLater
 from twisted.trial import unittest
 
 import mock
@@ -505,6 +506,7 @@
 yield self.assertFailure(w1.get_welcome(), WrongPasswordError)  # late
 
 yield self.assertFailure(w1.close(), WrongPasswordError)
+yield deferLater(reactor, 2, lambda: None)
 yield self.assertFailure(w2.close(), WrongPasswordError)
 
 @inlineCallbacks


I'm not a DD, so i can't upload any fixes, but i would really appreciate
if we can get this fixed before the auto removal strikes.

Regards,

 - Martin
--- magic-wormhole-0.12.0/src/wormhole/test/test_wormhole.py	2021-02-05 02:16:21.0 +
+++ magic-wormhole-0.12.0.fix/src/wormhole/test/test_wormhole.py	2023-04-07 14:44:26.792819150 +
@@ -6,6 +6,7 @@
 from twisted.internet import reactor
 from twisted.internet.defer import gatherResults, inlineCallbacks, returnValue
 from twisted.internet.error import ConnectionRefusedError
+from twisted.internet.task import deferLater
 from twisted.trial import unittest
 
 import mock
@@ -505,6 +506,7 @@
 yield self.assertFailure(w1.get_welcome(), WrongPasswordError)  # late
 
 yield self.assertFailure(w1.close(), WrongPasswordError)
+yield deferLater(reactor, 2, lambda: None)
 yield self.assertFailure(w2.close(), WrongPasswordError)
 
 @inlineCallbacks