The root cause of this problem with openscad is this upstream bug in Qt:
https://bugreports.qt.io/browse/QTBUG-95108
The symptoms are that in some keyboard layouts, the Qt::GroupSwitchModifier
is handled incorrectly, causing some keys to not respond correctly in
certain applications.
This bug
"Dr. Nikolaus Klepp" writes:
> Ok, setting the keyboard layout to US makes working again. Here 3x
> :
Thanks for the additional test. That data point does suggest that the QT bug
is the root cause.
>> 3. It would be interesting to try to downgrade Qt to some 5.15.4 version and
>> see if that
Thanks a lot for the additional information,
"Dr. Nikolaus Klepp" writes:
> This is what I get when I run "openscad --debug=scintillaeditor":
>
> Pressing 3 times:
> scintillaeditor: KeyRelease - modifiers: shift ctrl alt meta keypad GROUP
> scintillaeditor: KeyRelease - modifiers: shift ctrl
I did not manage to find a way to reproduce myself. But I found that
openscad has some debugging option that may provide additional information.
Can you try to run openscad with the --debug option:
openscad --debug=scintillaeditor
and then reproduce the problem with the ENTER key and note
Kristian Nielsen writes:
> I had some problems running openscad in my unstable/sid environment to try to
> reproduce this. I will try to solve this, but meanwhile...
I managed to run the latest openscad package in unstable. I did not see the
problem, the key worked fine to insert a line
"Dr. Nikolaus Klepp" writes:
> Package: openscad
> Version: 2021.01-5+b2
> After yesterdays upgrade openscad ignores the -key
>
> Expected:
> - press --> editor performs linebreak.
>
> Observed:
> - press --> nothing happen, the key is ignored.
> - press + --> editor performs linebreak.
Control: reassign
Thanks for the report, I will upload shortly a fixed openscad package.
- Kristian.
Paul Gevers writes:
> With a recent upload of ghostscript the autopkgtest of openscad fails
> in testing when that autopkgtest is run with the binary packages of
> ghostscript from unstable.
Public upstream bug reports:
https://github.com/openscad/openscad/issues/4037
https://github.com/openscad/openscad/issues/4043
Source: openscad
Severity: important
Dear Maintainer,
*** Reporter, please consider answering these questions, where appropriate ***
* What led up to the situation?
Upstream has reported two out-of-bounds memory access bugs, which have been
assigned CVEs:
The openscad bug tracking this issue is #996020:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=996020
- Kristian.
Package: release.debian.org
Severity: normal
Tags: buster
User: release.debian@packages.debian.org
Usertags: pu
[ Reason ]
This is a fix for two minor security issues in buster:
https://security-tracker.debian.org/tracker/CVE-2020-28599
Package: openscad
Version: 2019.01~RC2-2
Severity: important
There is a bug in the import() function in OpenSCAD when importing STL
files. Certain invalid files can cause out-of-bounds accesses, potentially
causing arbitrary code execution.
The bug is associated with these CVEs:
Not 100% clear what the issue(s) reported here are, but I looked through the
patches, and saw these:
1. Some libz/libzip changes. I'm guessing these are to avoid having to add
-lz/-lzip to the using project? There was a related change in the recent
NMU, 1.8.1+ds-3.1, which instead adds links to
(CVE-2021-21772), backporting fix from v2.1.1
+(Closes: #985092)
+
+ -- Kristian Nielsen Thu, 01 Apr 2021 21:25:54
+0200
+
lib3mf (1.8.1+ds-3.1) unstable; urgency=medium
* Non-maintainer upload.
diff -Nru lib3mf-1.8.1+ds/debian/control lib3mf-1.8.1+ds/debian/control
--- lib3mf-1.8.1+ds
Package: sponsorship-requests
Severity: important
Dear mentors,
I am looking for a sponsor for a new upload for package "openscad", which I
have been maintaining for a couple years. This upload fixes a FTBFS and is
needed to get openscad back into the archive.
* Package name: openscad
?
- Kristian.
commit 7225800b534a36e5ad84c56c274889b8d0edc0ce (HEAD -> newstuffs, salsa/newstuffs)
Author: Kristian Nielsen
Date: Tue Dec 24 11:05:45 2019 +0100
Add a patch to work-around test failures against latest mesa
diff --git a/debian/changelog b/debian/changelog
index 9ca95
I managed to reproduce without openscad, using a slightly modified glxgears
from mesa-demos:
---
--- glxgears.c.orig 2019-12-23 20:57:04.565366089 +0100
+++ glxgears.c 2019-12-23 20:57:53.933591974 +0100
@@ -503,6 +503,8 @@
Some additional data: The error occurs when openscad tries to run
glXSwapBuffers().
Here are the attributes requested from glXChooseFBConfig() and the list of
XVisualInfo returned:
int attributes[] = {
GLX_DRAWABLE_TYPE, GLX_WINDOW_BIT | GLX_PIXMAP_BIT | GLX_PBUFFER_BIT,
Source: mesa
Version: mesa-19.3.1-2
Severity: important
Dear Maintainer,
Since upload to unstable of mesa 19.3.1-1/19.3.1-2, the openscad testsuite
breaks. This also shows up as a regression on
https://tracker.debian.org/pkg/mesa .
The root cause is the following failure to run openscad in a
How about this? It uses --max-parallel to limit build to one process per 3
GB of memory:
https://salsa.debian.org/knielsen-guest/openscad/commit/3c64461f18762bbc2a1eebefe3c5f8880726791a
This should prevent builds failing due to out-of-memory on small machines,
and still allow most machines to
Sebastien Bacher writes:
> because it uses more memory than available. The attached patch makes it
> build with --no-parallel which workarounds the issue, what would you
Hm, this seems to be the wrong place to put this?
If a build machine has too little memory for a --parallel build of a
Helmut Grohne writes:
> openscad fails to cross build from source, because its python3
> dependency cannot be installed for the host architecture. It actually
> wants to run python3 during build, so it needs the build architecture
> python3. Thus it should be annotated :any. Please consider
On Sat, 27 Jul 2019 17:06:50 +0200 Helmut Grohne wrote:
> openscad fails to cross build from source, because its build dependency
> on openscad-mcad is unsatisfiable. In general, Architecture: all
> files (textual). Please mark it Multi-Arch: foreign.
Thanks, Helmut, sounds good. I will
I believe this is caused by a general limitation of OpenSCAD. When multiple
shapes are combined, the shapes must not share an edge or face exactly.
There is some explanation of this in the OpenSCAD manual:
https://en.wikibooks.org/wiki/OpenSCAD_User_Manual/STL_Import_and_Export#STL_Export
One
Otto Kekäläinen writes:
> Tags: unreproducible
>
> 2017-01-25 0:21 GMT+02:00 Julian Gilbey :
>> After upgrading to 10.1 from 10.0, I purged the old
>> mariadb-server-10.0 package, but this had two quite unpleasant
>> effects: it shut down the server and it
Otto Kekäläinen writes:
> On a quick look the solution suggested by Kristian looks OK. I will
> eventually do something to implement and test it, but if somebody has
> time to do it right now and send me a git merge request (or Github
> pull request) then things would progress
"Norvald H. Ryeng" writes:
> On Thu, 12 Jan 2017 13:16:32 +0100
> Dominik George wrote:
>> This package does not require Oracle's MySQL, it requires a root user
>> that can authenticate to the database and use it in its entirety, a
>> feature which
Otto Kekäläinen writes:
> - example configs are outdated anyway, probably no point including them at all
Sounds reasonable.
> - mytop maybe?
There is a separate `mytop' package in Debian, so probably not. I remember
previous discussions about this. If something is missing in
Otto Kekäläinen <o...@debian.org> writes:
> 2016-10-30 1:45 GMT+03:00 Kristian Nielsen <kniel...@knielsen-hq.org>:
>> Maybe something like this is needed on the cmake command?
>>
>> -DINSTALL_PLUGINDIR=/usr/lib/x86_64-linux-gnu/mysql/lib18/plugin
>>
Helmut Grohne writes:
> | dpkg: error processing archive
> /tmp/apt-dpkg-install-qGtx4F/3-libmariadbclient18_10.0.27-2_i386.deb
> (--unpack):
> | trying to overwrite shared '/usr/lib/mysql/plugin/dialog.so', which is
> different from other instances of package
FYI,
James Cowgill writes:
> mips-machine.patch
> Fix the value of DEFAULT_MACHINE on 32-bit mips platforms.
I've upstreamed this patch, it should appear in MariaDB 10.0.28.
> mips-groonga-atomic.patch
> Link groonga against libatomic as it uses 64-bit atomic operations
James Cowgill writes:
>>> Currently mariadb-10.0 FTBFS on mips64el due to various testsuite
>>> failures. The two attached patches should resolve this. I am also
>>> mips-errno-enotempty.patch
>>> On mips* architectures, ENOTEMPTY == 93 which wasn't handled by two test
>>>
John David Anglin writes:
> The build fails because the following two tests fail:
> rpl.rpl_drop_db binlog.binlog_database
> These two tests fail because the error number for ENOTEMPTY is 247 on hppa:
> dave@mx3210:/usr/include$ find . -name errno.h|xargs grep 247
>
Kristian Nielsen <kniel...@knielsen-hq.org> writes:
> And I'll see if I can fix mysql-test-run.pl upstream to not need '.' in
> @INC. Otherwise, this problem will probably show up in other places sooner
> or later...
I've now pushed to upstream 10.0 a patch so that mysql-test
Paul Gevers writes:
> I rather expect in the tool-chain. Has anybody seen the recent change in
> Perl? Your remarks about including ./ in the Perl path sounds exactly
> like the change that was recently made:
>
>
Otto Kekäläinen <o...@debian.org> writes:
> 2016-09-09 23:04 GMT+03:00 Kristian Nielsen <kniel...@knielsen-hq.org>:
>> ci.debian.net, maybe the docs have something, I can try looking. How is this
>> configured, is everything in the debian packaging, read by autopkgt
Otto Kekäläinen writes:
> Thank you very much Kristian for looking into this and producing a
> potential fix. Unfortunately I don't have any environment where I
> could test it, and I am a bit reluctant to apply the patch and test it
> in production unless somebody confirms that
Otto Kekäläinen writes:
> I uploaded now 10.0.27 and even with the recent fixes, it is still
> failing. I don't know how to fix it. Can you Paul help debug it?
> https://ci.debian.net/data/packages/unstable/amd64/m/mariadb-10.0/20160909_103759.autopkgtest.log.gz
> starting
Otto Kekäläinen o...@seravo.fi writes:
Upstream MariaDB only ships with libmysqlclient.so, while I have
changed the Debian version to ship the library using the name
libmariadbclient.so, because I got feedback that it is not good policy
for the MariaDB package to ship a so name that also
Jakub Wilk jw...@debian.org writes:
If a package is marked as Multi-Arch: same, it should be
co-installable with itself. This is currently not the case because of
the two files mentioned in my original mail:
Selecting previously unselected package libmariadbclient18:amd64.
Preparing to
intrig...@debian.org writes:
debian/rules only installs the usr.sbin.mysqld AppArmor profile on
Ubuntu (no idea why this is done in override_dh_installlogrotate-arch,
by the way; looks like a buggy merge to me, but anyway, that's
off-topic for now).
In my experience, there are a lot of
intrigeri intrig...@debian.org writes:
Hi,
Kristian Nielsen wrote (21 Jan 2014 09:18:05 GMT) :
In my experience, there are a lot of problems with installing an apparmor
profile by default for the MySQL server. This is from 4 years of experience
maintaining MariaDB .deb packages.
Thank you
Otto Kekäläinen o...@seravo.fi writes:
Ok, we can postpone the upload but could we now try to figure out the
technical solution of what will be eventually uploaded?
Maybe we could handle this by initially changing the Depends: on
libdbd-mysql-perl to a Recommends: ? That would allow to upload
Jonathan Aquilina eagles051...@gmail.com writes:
it was mentioned to give a user choice as to what to use. i have seen for
instance when you install gnome aside kde it asks you what desktop manager
you want to use. Wouldnt something like that be needed in this situation?
Gnome and KDE are
Jonathan Aquilina eagles051...@gmail.com writes:
then in that case why not setup mysql-server as a dependency for MariaDB?
I don't understand. Mariadb provides mysql-server. It doesn't depend on it.
- Kristian.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a
Lionel Elie Mamane lio...@mamane.lu writes:
The main issue that needs solving is to let MySQL and MariaDB packages
co-exist in the same repository.
So they _do_ need some more work :)
Yes. The existing packages are not suitable as-is for inclusion in Debian
because of this issue.
The
Lionel Elie Mamane lio...@mamane.lu writes:
On Tue, Feb 09, 2010 at 10:40:59PM +0100, Kristian Nielsen wrote:
Some additional information that might be useful:
- apt-able repositories of MariaDB are available from http://ourdelta.org/
- Package scripts for .deb are maintained on Launchpad
Some additional information that might be useful:
- apt-able repositories of MariaDB are available from http://ourdelta.org/
- Package scripts for .deb are maintained on Launchpad:
https://launchpad.net/ourdelta
- Build of packages are automated in Buildbot:
-a
Linux ymer 2.6.13.4 #3 Thu Oct 27 01:07:26 CEST 2005 i686 GNU/Linux
$ ls -l /lib/libc.so.6
lrwxrwxrwx 1 root root 13 2006-05-14 15:06 /lib/libc.so.6 - libc-2.3.6.so
--
Kristian Nielsen, Software Developer
MySQL AB, www.mysql.com
Are you MySQL certified? www.mysql.com/certification
49 matches
Mail list logo