Bug#861989: nautilus: Nautilus freezes/crashes when right clicking and selecting properties for 1 or more files

[Jason Crain]

On Sun, May 28, 2017 at 10:44:41AM +1000, Luke Christopher Clarke wrote:
> I ran both commands as non-sudo and sudo respectively.

You should run these as your regular user, not root.

> -- START tracker-control -S --
> luke@luk-main:~$ tracker-control -S
> Store:
> ^C

It is odd that tracker-control freezes too.  I was expecting a message
about how the tracker store is initializing or reindexing.

> -- tracker-control -k store; TRACKER_VERBOSITY=3 G_DEBUG=all --
> G_MESSAGES_DEBUG=all /usr/lib/tracker/tracker-store
> luke@luk-main:~$ tracker-control store; TRACKER_VERBOSITY=3 G_DEBUG=all
   ^

This is supposed to be "tracker-control -k store".  The point is to kill
(-k) the existing tracker-store process before running tracker-store
with debug messages enabled, because otherwise it will just complain
that there is already a process running.

> G_MESSAGES_DEBUG=all /usr/lib/tracker/tracker-store
> Unrecognized options: 'store'
> Tracker-Message: Starting tracker-store 1.2.4
...
> Tracker-Message: Registering D-Bus service...
>   Name:'org.freedesktop.Tracker1'
> 
> (tracker-store:12568): Tracker-CRITICAL **: D-Bus service
> name:'org.freedesktop.Tracker1' is already taken, perhaps the daemon is
> already running?
> Trace/breakpoint trap

As it does here.

Bug#861989: nautilus: Nautilus freezes/crashes when right clicking and selecting properties for 1 or more files

[Jason Crain]

On Fri, May 26, 2017 at 08:03:26PM +1000, Luke Christopher Clarke wrote:
> I have not gone out of my way to install any extra nautilus extensions to
> my knowledge. Is there a way I can find that out?
> 
> Also the following is the output from your command. I also went ahead and
> replicated the bug.
> 
...
> (nautilus:1733): Tracker-DEBUG: New TrackerTagsView with 1 files
> (nautilus:1733): Tracker-DEBUG: tracker-backend.vala:37: Waiting for
> service to become available...
> 
> -- END TERMINAL OUTPUT --

It looks like it's frozen because tracker is stuck.  What's the output
of "tracker-control -S"?  And what is the output of running
"tracker-control -k store; TRACKER_VERBOSITY=3 G_DEBUG=all
G_MESSAGES_DEBUG=all /usr/lib/tracker/tracker-store" ?

Bug#863419: evince: selecting text and trying to drag it causes crash

[Jason Crain]

Control: tags -1 + moreinfo

On Fri, May 26, 2017 at 06:25:52PM +0300, Sotiris Kouvopoulos wrote:
> When I select some text, for example a simple word and try to drag the
> selection (hold left mouse button and move mouse) the application crashes
> immediately.
> 
> This is the output when evince is started from terminal:
> (evince:7977): Gdk-ERROR **: The program 'evince' received an X Window
> System error.
> This probably reflects a bug in the program.
> The error was 'BadWindow (invalid Window parameter)'.
>   (Details: serial 3201 error_code 3 request_code 141 (Composite)
> minor_code 8)
>   (Note to programmers: normally, X errors are reported asynchronously;
>that is, you will receive the error a while after causing it.
>To debug your program, run it with the GDK_SYNCHRONIZE environment
>variable to change this behavior. You can then get a meaningful
>backtrace from your debugger if you break on the gdk_x_error() function.)
> Trace/breakpoint trap

I can't reproduce this.  What desktop environment are you using?  If you
are able, a backtrace with debug symbols and GDK_SYNCHRONIZE=1 as in
that error message would be helpful.

Bug#861989: nautilus: Nautilus freezes/crashes when right clicking and selecting properties for 1 or more files

[Jason Crain]

On Sun, May 14, 2017 at 11:04:23AM +1000, Luke Christopher Clarke wrote:
> The issue will occur on any particular type of file. I have created a video
> recording me replicating the bug using just a JPG.
> https://youtu.be/eUI2tVpZ6nY

It looks like the problem is bigger than just the properties dialog since in
your video the nautilus window doesn't even appear the first time you run it.
Can you run nautilus with this command and post the output?
TRACKER_VERBOSITY=3 G_DEBUG=all NAUTILUS_DEBUG=all G_MESSAGES_DEBUG=all 
nautilus

Also, if you have installed any extra nautilus extensions, try removing them.

Bug#863016: poppler: CVE-2017-9083

[Jason Crain]

On Sat, May 20, 2017 at 10:30:17AM +0200, Salvatore Bonaccorso wrote:
> the following vulnerability was published for poppler.
> 
> CVE-2017-9083[0]:
> | poppler 0.54.0, as used in Evince and other products, has a NULL
> | pointer dereference in the JPXStream::readUByte function in
> | JPXStream.cc. For example, the perf_test utility will crash
> | (segmentation fault) when parsing an invalid PDF file.

Does this apply to Debian's poppler?  I think uses openjpeg instead of
the internal JPX decoder.

There's a discussion on the poppler mailing list about making it more
explicit that the internal decoders are unmaintained:
https://lists.freedesktop.org/archives/poppler/2017-May/012230.html

Bug#862860: gnome-shell: Optionally allow the user to disable grouping in the Alt-Tab window switch

[Jason Crain]

On Wed, May 17, 2017 at 08:29:38PM +0200, Carlo Marchiori wrote:
> optionally allow the user to disable grouping in the Alt-Tab window switch.
> In fact, I find this grouping feature annoying and not handy at all.

There is already an "AlternateTab" extension which you can enable either
through gnome-tweak-tool or through https://extensions.gnome.org/.

Bug#861902: [gnome-core] Installing gnome-core precises apache2-bin and libapache2-mod-dnssd

[Jason Crain]

On Fri, May 05, 2017 at 04:03:35PM +0200, Rastersoft wrote:
> Installing the gnome-core package in debian SID forces to install
> apache2-bin and libapache2-mod-dnssd. Both packages shouldn't be needed for
> a full gnome shell session.

They're pulled in because of gnome-user-share.  Removing it would
disable the file sharing feature of GNOME.  

Bug#861989: nautilus: Nautilus freezes/crashes when right clicking and selecting properties for 1 or more files

[Jason Crain]

Control: tags -1 + moreinfo

On Sun, May 07, 2017 at 12:41:46PM +1000, Luke wrote:
> Method
> ~~
> 1.) Open up Nautilus
> 2.) Navigate to a folder where you want to find the file size of contents
> 3.) Select the file(s)/folder(s), right hand click
> 4.) Click on the option "Properties" in the menu
> 
> Expected Results
> 
> Properties dialog will appear
> 
> 
> Actual Results
> ~~
> Nautilus will lock up, will not be able to use it even if I have multiple
> windows opened. Eventually Gnome will inform me that Nautilus has potentially
> crashed.
> Other times, it has taken over a minute to load the properties dialog.
> 
> Notes
> ~
> Does not happen all the time, about 50% for me. When it does occur Nautilus
> will not open unless I either;
> - Open using root permissions
> - Use command 'pkill nautilus'
> 
> I can replicate this issue on two separate computers running Debian 8.7

I don't see this happening on either jessie or sid.  Does this only
happen under some particular circumstance, like a particular file, a
large directory, a network filesytem, etc?

Bug#862249: Mounting an SFTP share with path may lead to data being deleted

[Jason Crain]

On Wed, May 10, 2017 at 10:42:51AM +0200, Jonas Meurer wrote:
> I just discovered a severe bug in the sftp protocol support of nautilus:
> 
> I tried to mount a remote folder via SFTP/SSH by using a syntax similar
> to the following:
> 
> 'sftp:///path/to/directory'. Instead of displaying
> '/path/to/directory' on the remote host, nautilus kept giving warnings
> that it doesn't know what to do with file 'directory' and moved to the
> home directory on the remote host.
> 
> I tried it with different syntax (colon between host and path, 'user@'
> in front of host, using 'ssh://' instead of 'sftp://') and I tried both
> using the 'Andere Orte' (something like 'different locations' in
> english) and the address bar (+). One time nautilus even
> crashed (the Files window got closed).
> 
> After some time, I went back to the remote console SSH session and was
> shocked to realize that the whole directory '/path/to/directory' was
> removed on the remote host. Luckily I had backups.
> 
> I don't have time to do further debugging right now as I'm quite busy,
> but I will do further debugging and try to find a clear reproducer in
> the following days.

Please do try to find a way to reproduce this.  I suspect that nautilus'
warning messages may be caused by the path being missing or a filesystem
problem, instead of nautilus deleting the path.  I've tried opening
paths in nautilus several ways and I can't find any problem with it.

Bug#861575: tilix: Does not start on Stretch (symbol lookup error: undefined symbol)

[Jason Crain]

On Sun, Apr 30, 2017 at 06:43:18PM -0400, Mathieu Basille wrote:
> Following the name change from terminix to tilix, I updated 
> my Stretch system (which uses apt pining to install Sid 
> packages whenever possible) to use the new developments of 
> tilix.
> 
> Install of v1.5.4-1 works without any problem, but I'm unable
> to run tilix:
> 
> tilix: symbol lookup error: tilix: undefined symbol:
>  
> _D7gobject7ObjectG7ObjectG11addOnNotifyMFDFC7gobject9ParamSpec9ParamSpecC7gobject7ObjectG7ObjectGZvAyaE4gtkc12gobjecttypes13GConnectFlagsZm
...
> 
> Versions of packages tilix depends on:
> ii  dconf-gsettings-backend [gsettings-backend]  0.26.0-2+b1
> ii  libc62.24-10
> ii  libgtkd-3-0  3.3.1-2
> ii  libphobos2-ldc71 1:1.1.1-1
> ii  libvted-3-0  3.3.1-2
> ii  libx11-6 2:1.6.4-3
> ii  tilix-common 1.5.4-1

tilix seems to require libgtkd-3-0 >= 3.5.1-1 and should have its
Depends increased.

Bug#860796: appears to be linked to the lack of a running "nautilus --gapplication-service"

[Jason Crain]

On Fri, Apr 28, 2017 at 09:15:34AM +0100, Ben Green wrote:
> So, I'm back to thinking that I should have 'nautilus
> --gapplication-service' running, but somehow managing not to terminate, and
> that that is all the problem is.

That's not the problem because 'nautilus --gapplication-service' is
supposed to exit after a while.  nautilus-desktop contacts nautilus over
D-Bus, D-Bus runs 'nautilus --gapplication-service' if it's not already
running, and then nautilius exits after a short time if it's not being
used.

Bug#860796: appears to be linked to the lack of a running "nautilus --gapplication-service"

[Jason Crain]

On Tue, Apr 25, 2017 at 02:20:08PM +0100, Ben Green wrote:
> I do have KDE, mate and xfce4 installed. Having removed and purged KDE from
> my laptop, which shares the same issue, there's no difference.
> 
> There used to be a problem that clicking on a desktop file would launch
> dolphin. I've fixed this in the past by removing:
> 
> /usr/share/dbus-1/services/org.kde.dolphin.FileManager1.service
> 
> Though it's presence, tested through reboots, makes no difference to the
> current situation.
> 
> Here's a listing of /usr/share/dbus-1/services/ FWIW
> 
[snip]
> 
> I'll next try removing mate and xfce4 to see if that helps.

I think that if you remove org.mate.freedesktop.FileManager1.service,
that might fix it.  These all provide the org.freedesktop.FileManager1
dbus service and they might be interfering with each other.  Ideally KDE
would use dolphin, mate would use caja, and gnome would use nautilus,
but I don't know if there's a way to tell dbus to start different
services for different DEs.

On Tue, Apr 25, 2017 at 02:28:19PM +0100, Ben Green wrote:
> having had a looking journalctl and /var/log/syslog, I note that there is no
> output produced when clicking desktop icons.

I'm surprised there's not at least something about dbus not being able
to start the org.freedesktop.FileManager1 service, or nautilus not being
able to show items.

Bug#860445: libvte9: U+1F3DB CLASSICAL BUILDING incorrectly rendered as a singlewidth character

[Jason Crain]

Control: reassign -1 libvte-2.91-0 0.46.1-1
Control: forwarded -1 https://bugzilla.gnome.org/781676

On Sun, Apr 16, 2017 at 07:34:25PM -0400, Alexis Hunt wrote:
> I noticed recently that U+1F3DB CLASSICAL BUILDING is incorrectly being 
> rendered
> as a singlewidth character when it, like most (all?) emoji, is actually
> doublewidth. The result is that it overlaps with the subsequent character. It
> should be rendered as doublewidth.
> 
> I have reproduced this in Terminator and in rxvt-unicode.

I assume that you meant to file this against vte2.91  Terminator uses
vte2.91.  rxvt-unicode doesn't use any version of vte though so it won't
be affected by any possible vte bugs.

The Unicode spec is ambiguous about some of those emoji characters.  It
lists the range 1F3D4..1F3DF as neutral (N) rather than wide (W), but
elsewhere it says that emoji are considered to be wide.  I've forwarded
this to the vte developers so they can comment on whether they consider
it a bug.

Bug#858924: gnome-calendar: Segfault in gcal_manager_create_event when adding event

[Jason Crain]

Control: forwarded -1 https://bugzilla.gnome.org/show_bug.cgi?id=779733
Control: tags -1 + patch

On Tue, Mar 28, 2017 at 01:36:45PM -0400, Daniel Gnoutcheff wrote:
> Package: gnome-calendar
> Version: 3.22.3-1
> Severity: important
> 
> To reproduce:
> 
> 1. Click on an empty day to add a new event
> 2. Click "Edit Details..."
> 3. Click "Done"
> 
> Result: gnome-calendar segfaults.  

This crash is fixed upstream with the attached patch.
>From 6f87ada70dbeae71e3428ee3a63f79b8c918f121 Mon Sep 17 00:00:00 2001
From: Yash Singh 
Date: Fri, 10 Mar 2017 17:58:52 +0530
Subject: [PATCH] window: set edit dialog's event to NULL after saving

Earlier the app used to crash when a new event was added using the
'Edit Details' buttion/dialog. This was happening because edit dialog's
event was set to NULL before the event was being stored in the calendar
through the 'edit_dialog' and hence the app was crashing.

This patch fixes the above-mentioned issue by saving the event before
setting the edit dialog's event to NULL.

https://bugzilla.gnome.org/show_bug.cgi?id=779733
---
 src/gcal-window.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/src/gcal-window.c b/src/gcal-window.c
index 263be51..b44af8f 100644
--- a/src/gcal-window.c
+++ b/src/gcal-window.c
@@ -1135,7 +1135,6 @@ edit_dialog_closed (GtkDialog *dialog,
   event = gcal_edit_dialog_get_event (edit_dialog);
   view = GCAL_VIEW (window->views[window->active_view]);
 
-  gcal_edit_dialog_set_event (edit_dialog, NULL);
   gtk_widget_hide (GTK_WIDGET (dialog));
 
   switch (response)
@@ -1182,6 +1181,8 @@ edit_dialog_closed (GtkDialog *dialog,
   break;
 
 }
+
+  gcal_edit_dialog_set_event (edit_dialog, NULL);
 }
 
 static void
-- 
2.11.0

Bug#838795: gnome-shell: Startup notification fails on Wayland

[Jason Crain]

Version: 3.22.2-3

On Sat, Sep 24, 2016 at 11:23:07PM +0100, Tony Houghton wrote:
> When launching an application the spinner continues to show next to its
> name in the taskbar, and the app menu can not be opened, for several
> seconds after the app has opened, presumably until it times out. Sorry
> if this is the wrong package, I don't know what's responsible if it
> isn't gnome-shell.

There's a workaround for this included in mutter 3.22.2-3 and later.

Bug#851061: Collections don't work at all, do they?

[Jason Crain]

Version: 3.22.1-1

On Wed, Jan 11, 2017 at 10:21:02PM +0100, Richard B. Kreckel wrote:
> - Nothing happens when I click on the collection item.

Fixed in version 3.22.1.

Bug#855449: nautilus-sendto: Thunderbird debranding breaks the program

[Jason Crain]

On Sat, Feb 18, 2017 at 12:54:28PM +0100, Andrea Stacchiotti wrote:
> Package: nautilus-sendto
> Version: 3.8.4-2
> Severity: important
> 
> Dear maintainer,
> 
> following the debranding of Icedove into Thunderbird, this package does not
> work anymore with Thunderbird.

This is fixed in a recent version of icedove.

Bug#856652: RFS: xpdf/3.0.4.real-4

[Jason Crain]

On Sat, Mar 04, 2017 at 08:19:43PM +0100, Svante Signell wrote:
> And BTW: poppler upstream seems to be freedesktop.org, i.e. gnome. Who
> can trust gnome nowadays, especially people preferring systemd-free
> software?

Sniping at GNOME aside, freedesktop.org is independent from GNOME, and
poppler's lead maintainer is a KDE developer.

Bug#856652: RFS: xpdf/3.0.4.real-4

[Jason Crain]

On Sat, Mar 04, 2017 at 07:28:37PM +0100, Svante Signell wrote:
> On Sat, 2017-03-04 at 09:46 -0600, Jason Crain wrote:
> > The upstream xpdf source contains a file misc/hello.pdf for testing
> > purposes, according to the INSTALL file.  It would likely need to be
> > repacked to remove that file.
> 
> What's the problem with that file? It is not part of any xpdf*.deb.

Listed under "Source package missing source" on the FTP master's reject FAQ¹:

"Source packages are part of the distribution. As such source must be
present for all files in the source package itself, this includes
convenience copies of other projects."

I think this applies whether or not the file is used in the binary
package.  The source for that PDF is not included, making it not DFSG
free.

¹ https://ftp-master.debian.org/REJECT-FAQ.html

Bug#856652: RFS: xpdf/3.0.4.real-4

[Jason Crain]

On Sat, Mar 04, 2017 at 08:02:29AM -0700, Sean Whitton wrote:
> It sounds like the source should not in fact be repacked.  What do you
> think, Svante?

The upstream xpdf source contains a file misc/hello.pdf for testing
purposes, according to the INSTALL file.  It would likely need to be
repacked to remove that file.

Bug#698935: I found how to fix.

[Jason Crain]

On Sat, Mar 04, 2017 at 04:46:58PM +0330, Farhad Mohammadi Majd wrote:
> unfortunately "BiCon" is not available on any of Debian repositories,
> please someone do packaging it.

BiCon was actually requested some time ago,
https://bugs.debian.org/446813, but no one has packaged it yet.

If you want it packaged, you can package it yourself and become the
maintainer.  Start by reading the New Maintainer's Guide
https://www.debian.org/doc/devel-manuals#maint-guide.  Ask questions on
the debian-mentors mailing list or on #debian-mentors on irc and read
the Mentors FAQ https://wiki.debian.org/DebianMentorsFaq.  Then request
sponsorship when you are ready to have someone review it.

Bug#856652: RFS: xpdf/3.0.4.real-4

[Jason Crain]

On Fri, Mar 03, 2017 at 11:29:44AM +0100, Svante Signell wrote:
> I am looking for a sponsor for my ITA package "xpdf"

If you're going to adopt the Xpdf package, I thought you might want to
know a little about Xpdf first and why the Debian package is the way
that it is.  The Debian package modifies Xpdf to make it use poppler
(https://poppler.freedesktop.org) for rendering instead of using it's
own internal code.  I think this was done for security and bug fixes,
and because the poppler upstream is more responsive than Xpdf.

Poppler is a fork of Xpdf.  Before poppler, it was common for
applications which wanted to read PDFs to embed a copy of the Xpdf code.
Poppler was created to turn sections of the Xpdf code into a library so
that it could be more easily reused, to have a centralized place for
development, and to put APIs over the code.

Debian's Xpdf is significantly different from upstream Xpdf.  The
package has build rules and patches which modify the code to be
compatibile with poppler.  Also, since the poppler devs do not guarantee
stability for the old Xpdf functions, because those functions were never
intended to be a stable interface, the patches will occasionally need to
be modified to match any changes in poppler.

Bug#855977: evince: hyperlinks not clickable - feature request: automatic links

[Jason Crain]

On Fri, Feb 24, 2017 at 12:30:50AM +0100, Petr Vanek wrote:
> I have been using Evince for years and am happy with it, but for
> whatever reason, it doesn't make links clickable in my own created
> documents, see example here: http://www.penguin.cz/~vanous/out.pdf - the
> numbers in circles have hyperlinks underneath.

This particular bug has already been reported upstream to the evince
developers at https://bugzilla.gnome.org/669107.  As such, I am closing
this bug because I do not feel that keeping it open is helpful.  If you
have further comments for this bug, please direct them to the GNOME
bugzilla.

Bug#856383: libglib2.0-0: Epiphany SIGSEVs in slab_allocator_free_chunk at ././glib/gslice.c:1347

[Jason Crain]

Control: tags -1 + moreinfo

On Tue, Feb 28, 2017 at 03:31:41PM +0200, Andres Gomez wrote:
> Package: libglib2.0-0
>
> my locally built epiphany SIGSEVed in slab_allocator_free_chunk, after
> visiting several pages.

On Tue, Feb 28, 2017 at 05:41:02PM +0200, Andres Gomez wrote:
> Package: libgtk-3-0
>
> My locally built epiphany SIGSEVed in gtk_css_static_style_new_compute, after
> visiting several pages.

Both of these are more likely to be a bug in webkit or epiphany than in
glib or gtk.  For a next step, try setting the environment variable
G_SLICE=debug-blocks before running and see if you can get an abort()
closer to your bug.  If that doesn't work and you're feeling
adventurous, you can try some other things like compiling with
AddressSanitizer (-fsanitize) or running under valgrind.

The problem with memory corruption bugs is that the stack trace you've
provided only shows where the memory corruption was detected.  It
doesn't show where the memory corruption actually happened.

Bug#856327: unblock: gexiv2/0.10.4-2

[Jason Crain]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock

Please unblock package gexiv2

The unstable version adds a patch that fixes an assertion.  The current testing
version can abort when processing certian images from Minolta cameras, and
cause shotwell to crash.

Debian bug: https://bugs.debian.org/856101
Gnome bug: https://bugzilla.gnome.org/776233
RedHat bug: https://bugzilla.redhat.com/1401706

debdiff is attached

unblock gexiv2/0.10.4-2

-- System Information:
Debian Release: 9.0
  APT prefers testing
  APT policy: (9000, 'testing')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.9.0-1-amd64 (SMP w/2 CPU cores)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
diff -Nru gexiv2-0.10.4/debian/changelog gexiv2-0.10.4/debian/changelog
--- gexiv2-0.10.4/debian/changelog  2016-10-06 09:39:18.0 -0500
+++ gexiv2-0.10.4/debian/changelog  2017-02-26 05:16:34.0 -0600
@@ -1,3 +1,11 @@
+gexiv2 (0.10.4-2) unstable; urgency=medium
+
+  * Add patch 0004-get_orientation-Fix-abort-on-Minolta-meta-data.patch.
+Fixes an assertion when processing metadata from a not-rotated image
+from Minolta cameras. (Closes: #856101)
+
+ -- Jason Crain <ja...@inspiresomeone.us>  Sun, 26 Feb 2017 05:16:34 -0600
+
 gexiv2 (0.10.4-1) unstable; urgency=medium
 
   * New upstream version 0.10.4
diff -Nru 
gexiv2-0.10.4/debian/patches/0004-get_orientation-Fix-abort-on-Minolta-meta-data.patch
 
gexiv2-0.10.4/debian/patches/0004-get_orientation-Fix-abort-on-Minolta-meta-data.patch
--- 
gexiv2-0.10.4/debian/patches/0004-get_orientation-Fix-abort-on-Minolta-meta-data.patch
  1969-12-31 18:00:00.0 -0600
+++ 
gexiv2-0.10.4/debian/patches/0004-get_orientation-Fix-abort-on-Minolta-meta-data.patch
  2017-02-26 05:16:34.0 -0600
@@ -0,0 +1,57 @@
+From: Jens Georg <m...@jensge.org>
+Date: Sun, 18 Dec 2016 17:02:42 +0100
+Subject: get_orientation: Fix abort on Minolta meta-data
+
+Properly handle rotation value 72 (not rotated)
+
+Origin: upstream, 
https://git.gnome.org/browse/gexiv2/commit/?id=6810c92d7b2978f97ac9750ddf8a6a62b44be4cc
+Bug: https://bugzilla.gnome.org/776233
+Bug-Redhat: https://bugzilla.redhat.com/1401706
+Bug-Debian: https://bugs.debian.org/856101
+Last-Update: 2017-02-23
+---
+ gexiv2/gexiv2-metadata.cpp | 17 +++--
+ 1 file changed, 11 insertions(+), 6 deletions(-)
+
+diff --git a/gexiv2/gexiv2-metadata.cpp b/gexiv2/gexiv2-metadata.cpp
+index 8aa4d45..7d1be08 100644
+--- a/gexiv2/gexiv2-metadata.cpp
 b/gexiv2/gexiv2-metadata.cpp
+@@ -395,26 +395,31 @@ GExiv2Orientation gexiv2_metadata_get_orientation 
(GExiv2Metadata *self) {
+ // Because some camera set a wrong standard exif orientation tag,
+ // We need to check makernote tags first!
+ if (gexiv2_metadata_has_exif_tag(self, "Exif.MinoltaCs7D.Rotation")) {
+-switch (gexiv2_metadata_get_exif_tag_long(self, 
"Exif.MinoltaCs7D.Rotation")) {
++long orientation = gexiv2_metadata_get_exif_tag_long(self, 
"Exif.MinoltaCs7D.Rotation");
++switch (orientation) {
+ case 76:
+ return GEXIV2_ORIENTATION_ROT_90;
+-
+ case 82:
+ return GEXIV2_ORIENTATION_ROT_270;
++case 72:
++return GEXIV2_ORIENTATION_UNSPECIFIED;
+ default:
+-g_assert_not_reached();
++g_debug ("Unknown Minolta rotation value %ld, ignoring", 
orientation);
+ }
+ }
+-
++
+ if (gexiv2_metadata_has_exif_tag(self, "Exif.MinoltaCs5D.Rotation")) {
+-switch (gexiv2_metadata_get_exif_tag_long(self, 
"Exif.MinoltaCs5D.Rotation")) {
++long orientation = gexiv2_metadata_get_exif_tag_long(self, 
"Exif.MinoltaCs5D.Rotation");
++switch (orientation) {
+ case 76:
+ return GEXIV2_ORIENTATION_ROT_90;
+ 
+ case 82:
+ return GEXIV2_ORIENTATION_ROT_270;
++case 72:
++return GEXIV2_ORIENTATION_UNSPECIFIED;
+ default:
+-g_assert_not_reached();
++g_debug ("Unknown Minolta rotation value %ld, ignoring", 
orientation);
+ }
+ }
+ 
diff -Nru gexiv2-0.10.4/debian/patches/series 
gexiv2-0.10.4/debian/patches/series
--- gexiv2-0.10.4/debian/patches/series 2016-10-06 09:39:18.0 -0500
+++ gexiv2-0.10.4/debian/patches/series 2017-02-26 05:16:34.0 -0600
@@ -1,3 +1,4 @@
 exiv2_pc.patch
 fix-typelib-install-directory.patch
 always-install-vapi-file.patch
+0004-get_orientation-Fix-abort-on-Minolta-meta-data.patch

Bug#856151: RFS: gexiv2/0.10.4-2

[Jason Crain]

Control: tags -1 - moreinfo

On Sun, Feb 26, 2017 at 05:12:44PM -0700, Sean Whitton wrote:
> I built the package and then ran Lintian.  It produces an error
> 
> E: libgexiv2-2:
> symbols-file-contains-current-version-with-debian-revision on symbol 
> _ZN5Exiv28XmpdatumaSIlEERS0_RKT_@Base and 1 others

I noticed that those types of errors showed up in a couple of the buildd
logs after the last upload.  It's because this is a C++ library that
exports a C interface.  Most of the C++ symbols are marked as optional
in the symbols file, but it's hard to catch all of them because they
vary between architectures.  My long term plan is to copy something I've
seen in libpoppler-glib and use a regex to mark all C++ symbols as
optional, but I didn't think in a freeze was a good time to do that.

Bug#856151: RFS: gexiv2/0.10.4-2

[Jason Crain]

Control: tags -1 - moreinfo

On Sat, Feb 25, 2017 at 07:03:33PM -0700, Sean Whitton wrote:
> You are missing a # before the number of the bug you are closing in your
> changelog.

I've fixed the changelog entry.  I think I got it wrong because I
misread the gbp documentation on meta tags.  Fixed and reploaded to
mentors:

https://mentors.debian.net/debian/pool/main/g/gexiv2/gexiv2_0.10.4-2.dsc

Bug#856184: nautilus: hidden directories becomes the output standard of the Nautilus every time when i press CTRL + H inside on "move file" to move (transfer) some file to a hidden directory.

[Jason Crain]

On Sun, Feb 26, 2017 at 10:44:19AM +0100, Michael Biebl wrote:
> Am 26.02.2017 um 08:35 schrieb Leandro:
> >* What led up to the situation? Press CTRL + H inside of the "move to" 
> > of the nautilus.
> >* What exactly did you do (or not do) that was effective (or
> >  ineffective)? Reconfigure the nautilus on the Dconfig. This solved the 
> > problem, but he returns every time when i do this.
> 
> ctrl+h is the shortcut to show hidden files. This is intended behaviour.

And going through dconf is not necessary.  You can just Ctrl+H again to
hide files or through the header bar menu.

Bug#856151: RFS: gexiv2/0.10.4-2

[Jason Crain]

Package: sponsorship-requests
Severity: normal

Dear mentors,

I am looking for a sponsor for my package "gexiv2"

 * Package name: gexiv2
   Version : 0.10.4-2
   Upstream Author : Jim Nelson <j...@yorba.org>
 * URL : https://wiki.gnome.org/Projects/gexiv2
 * License : GPL-2+
   Section : libs

It builds those binary packages:

 gir1.2-gexiv2-0.10 - GObject-based wrapper around the Exiv2 library - 
introspection data
 libgexiv2-2 - GObject-based wrapper around the Exiv2 library
 libgexiv2-dev - GObject-based wrapper around the Exiv2 library - development 
file

To access further information about this package, please visit the following 
URL:

https://mentors.debian.net/package/gexiv2

Alternatively, one can download the package with dget using this command:

  dget -x 
https://mentors.debian.net/debian/pool/main/g/gexiv2/gexiv2_0.10.4-2.dsc

More information about gexiv2 can be obtained from 
https://wiki.gnome.org/Projects/gexiv2.

Changes since the last upload:

  * Add patch 0004-get_orientation-Fix-abort-on-Minolta-meta-data.patch.
Fixes an assertion when processing metadata from a not-rotated image
from Minolta cameras. (Closes: 856101)

I am intending to have this in unstable so I can apply for an unblock.

Regards,
 Jason Crain

Bug#856101: libgexiv2-2: assertion when reading unrotated image from Minolta camera

[Jason Crain]

Package: libgexiv2-2
Version: 0.10.4-1
Severity: important

Tags: patch
Forwarded: https://bugzilla.gnome.org/776233

If the image at https://bugzilla.redhat.com/attachment.cgi?id=1228168 is opened
in shotwell, it aborts with the following message:

**
ERROR:gexiv2/gexiv2-metadata.cpp:405:GExiv2Orientation
gexiv2_metadata_get_orientation(GExiv2Metadata*): code should not be reached
Aborted

A patch to fix the issue is available at https://bugzilla.gnome.org/776233.

backtrace:

#0  0x7fffef0bdfdf in __GI_raise (sig=sig@entry=6) at
../sysdeps/unix/sysv/linux/raise.c:58
#1  0x7fffef0bf40a in __GI_abort () at abort.c:89
#2  0x7fffef8e4515 in g_assertion_message (domain=domain@entry=0x0,
file=file@entry=0x77905ba1 "gexiv2/gexiv2-metadata.cpp",
line=line@entry=405, func=func@entry=0x77906860
 "GExiv2Orientation
gexiv2_metadata_get_orientation(GExiv2Metadata*)",
message=message@entry=0x7fffb4001f00 "code should not be reached") at
././glib/gtestutils.c:2432
#3  0x7fffef8e45aa in g_assertion_message_expr (domain=domain@entry=0x0,
file=file@entry=0x77905ba1 "gexiv2/gexiv2-metadata.cpp",
line=line@entry=405, func=func@entry=0x77906860
 "GExiv2Orientation
gexiv2_metadata_get_orientation(GExiv2Metadata*)", expr=expr@entry=0x0) at
././glib/gtestutils.c:2455
#4  0x778f7939 in gexiv2_metadata_get_orientation(GExiv2Metadata*)
(self=0x569a20a0 [GExiv2Metadata]) at gexiv2/gexiv2-metadata.cpp:405
#5  0x5562b9c4 in photo_metadata_get_orientation (self=0x568401e0
[PhotoMetadata]) at ./src/photos/PhotoMetadata.vala:1121
#6  0x55734e8d in photo_prepare_for_import
(params=params@entry=0x7fffb000d5e0) at ./src/Photo.vala:1218
#7  0x55761224 in prepared_file_import_job_real_execute
(base=) at ./src/BatchImport.vala:1983
#8  0x555e7068 in workers_thread_start (ignored=,
self=0x569890d0) at ./src/threads/Workers.vala:96
#9  0x555e7068 in _workers_thread_start_gfunc (data=,
self=0x569890d0) at ./src/threads/Workers.vala:31
#10 0x7fffef8e5d3e in g_thread_pool_thread_proxy (data=) at
././glib/gthreadpool.c:307
#11 0x7fffef8e5345 in g_thread_proxy (data=0x56a7d720) at
././glib/gthread.c:784
#12 0x7fffef430424 in start_thread (arg=0x7fffc5cfe700) at
pthread_create.c:333
#13 0x7fffef1739bf in clone () at
../sysdeps/unix/sysv/linux/x86_64/clone.S:105



-- System Information:
Debian Release: 9.0
  APT prefers testing
  APT policy: (9000, 'testing')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.9.0-1-amd64 (SMP w/2 CPU cores)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages libgexiv2-2 depends on:
ii  libc6 2.24-9
ii  libexiv2-14   0.25-3
ii  libgcc1   1:6.3.0-6
ii  libglib2.0-0  2.50.2-2
ii  libstdc++66.3.0-6

libgexiv2-2 recommends no packages.

libgexiv2-2 suggests no packages.

-- no debconf information

Bug#855449: nautilus-sendto: Thunderbird debranding breaks the program

[Jason Crain]

Control: block -1 by 855334

On Sat, Feb 18, 2017 at 12:54:28PM +0100, Andrea Stacchiotti wrote:
> following the debranding of Icedove into Thunderbird, this package does not
> work anymore with Thunderbird.

I asked about this on https://bugs.debian.org/855406.

The new package has a wrapper script for Thunderbird, which takes care
of migrating from icedove to thunderbird.  The wrapper parses command
line arguments so it is now necessary to add "--" to the command line to
separate the wrapper's arguments from thunderbird's arguments.  In other
words, nautilus-sendto currently calls thunderbird like this:

/usr/bin/thunderbird -compose attachment='file:///tmp/attachment'

It will be necessary to call thunderbird like this:

/usr/bin/thunderbird -- -compose attachment='file:///tmp/attachment'

Though that won't work either until bug https://bugs.debian.org/855334
is fixed, because the wrapper doesn't forward arguments correctly.

Bug#855406: [thunderbird] commandline argument "-compose attachment" not recognized

[Jason Crain]

On Sun, Feb 19, 2017 at 12:24:51PM +0100, Carsten Schoenert wrote:
> without feedback it's obviously impossible to solve this report.
> I'll close the report then now. The command line option '-compose' is
> working so far with the fix from Daniel.

Having this same problem with nautilus-sendto: https://bugs.debian.org/855449

nautilus-sendto calls icedove/thunderbird as "/usr/bin/icedove -compose
attachment='file:///tmp/attachment'" when trying to email a file.  This
no longer works with the new thunderbird.  However, after applying the
patch in https://bugs.debian.org/855334, it will work if the command is
changed to "/usr/bin/icedove -- -compose attachment='file:///tmp/attachment'".

Is this the recommended way to do it now?  Debian's icedove/thunderbird
is going to now require "--" before any options?

Bug#855516: gnome-shell: Crashes when Shotwell's Adjust dialog is moved

[Jason Crain]

Version: 3.22.3-1

Fixed by mutter in unstable

Bug#855681: Bug#855516: gnome-shell: Crashes when Shotwell's Adjust dialog is moved

[Jason Crain]

Control: severity -1 important
Control: tags -1 + patch

Shotwell 0.25.4+really0.24.5-0.1, currently in unstable, crashes under
wayland when the adjustment dialog is moved.  Same for the crop toolbar
and probably others elements.  This has been fixed upstream in the
development version.  Please consider applying the patch.

patch:
https://git.gnome.org/browse/shotwell/commit/?id=1eec412e97c255c9ca

backtrace:
#0  0x7fa2b9395129 in wl_proxy_marshal (proxy=0x0, opcode=opcode@entry=5) 
at ../src/wayland-client.c:692
#1  0x7fa2c346cfd3 in zxdg_toplevel_v6_move (serial=209, seat=, zxdg_toplevel_v6=) at 
../../gdk/wayland/xdg-shell-unstable-v6-client-protocol.h:1332
#2  0x7fa2c346cfd3 in gdk_wayland_window_begin_move_drag (window=, device=0x55db1c71c160 [GdkWaylandDevice], button=, 
root_x=, root_y=, timestamp=)
at ././gdk/wayland/gdkwindow-wayland.c:3481
#3  0x55db1b0acec2 in 
editing_tools_editing_tool_window_real_button_press_event (base=, event=) at ./src/editing_tools/EditingTools.vala:86
#4  0x7fa2c38f9b71 in _gtk_marshal_BOOLEAN__BOXEDv (closure=0x55db1c75ea70, 
return_value=0x7ffd15dfed80, instance=0x55db1d61c2b0, args=, 
marshal_data=, n_params=, 
param_types=0x55db1c775aa0) at ././gtk/gtkmarshalers.c:131
#5  0x7fa2c1d4b1a4 in _g_closure_invoke_va 
(closure=closure@entry=0x55db1c75ea70, 
return_value=return_value@entry=0x7ffd15dfed80, 
instance=instance@entry=0x55db1d61c2b0, args=args@entry=0x7ffd15dfee50, 
n_params=, param_types=0x55db1c775aa0) at 
././gobject/gclosure.c:867
#6  0x7fa2c1d653a1 in g_signal_emit_valist (instance=0x55db1d61c2b0, 
signal_id=, detail=0, var_args=var_args@entry=0x7ffd15dfee50) at 
././gobject/gsignal.c:3300
#7  0x7fa2c1d65fbf in g_signal_emit 
(instance=instance@entry=0x55db1d61c2b0, signal_id=, 
detail=detail@entry=0) at ././gobject/gsignal.c:3447
#8  0x7fa2c3a464ac in gtk_widget_event_internal (widget=0x55db1d61c2b0 
[EditingToolsAdjustToolAdjustToolWindow], event=0x55db1d06c020) at 
././gtk/gtkwidget.c:7723
#9  0x7fa2c3a66ea3 in _gtk_window_check_handle_wm_event 
(event=event@entry=0x55db1d06c020) at ././gtk/gtkwindow.c:8135
#10 0x7fa2c38f8891 in gtk_main_do_event (event=0x55db1d06c020) at 
././gtk/gtkmain.c:1748
#11 0x7fa2c3409e15 in _gdk_event_emit (event=event@entry=0x55db1d06c020) at 
././gdk/gdkevents.c:73
#12 0x7fa2c34659a2 in gdk_event_source_dispatch (base=, 
callback=, data=) at 
././gdk/wayland/gdkeventsource.c:121
#13 0x7fa2c0b8d7f7 in g_main_dispatch (context=0x55db1c716e70) at 
././glib/gmain.c:3203
#14 0x7fa2c0b8d7f7 in g_main_context_dispatch 
(context=context@entry=0x55db1c716e70) at ././glib/gmain.c:3856
#15 0x7fa2c0b8da60 in g_main_context_iterate 
(context=context@entry=0x55db1c716e70, block=block@entry=1, 
dispatch=dispatch@entry=1, self=) at ././glib/gmain.c:3929
#16 0x7fa2c0b8db0c in g_main_context_iteration 
(context=context@entry=0x55db1c716e70, may_block=may_block@entry=1) at 
././glib/gmain.c:3990
#17 0x7fa2c202d70d in g_application_run (application=0x55db1c7862b0 
[GtkApplication], argc=argc@entry=0, argv=argv@entry=0x0) at 
././gio/gapplication.c:2381
#18 0x55db1b298c59 in application_start (self=0x55db1c7900e0, argv=0x0, 
argv_length1=0) at ./src/Application.vala:170
#19 0x55db1b18a822 in library_exec (mounts=, 
mounts_length1=) at ./src/main.vala:197
#20 0x55db1b18b50c in _vala_main (args=, 
args_length1=) at ./src/main.vala:423
#21 0x55db1b083e70 in main (argc=, argv=) at 
./src/main.vala:317

Bug#843819: gnome-terminal: Parenthesis in the directory name leads to crash

[Jason Crain]

Control: tags -1 - moreinfo unreproducible
Control: reassign -1 bash 4.4-4

On Sat, Feb 04, 2017 at 07:49:22PM +0200, Prometheus wrote:
> Ok I managed to reproduce this:
> 
> Create a new folder in nautilus 
> 
> /mnt/new/Temp/data/syncc/unklll//ΠΠ-07-Ααα βββ
> γδ //-άάά 2-θό ώ
> ύυα/Plot-LLC (3350)
> 
> now go one folder behind the last and open terminal
> 
> cd Plot-LLC\ \(3350\)/
> 
> The terminal closes immediately.

This looks like a problem in bash. gnome-terminal closes because the
bash process ends. I can reproduce it with the following commands:

mkdir -p "/tmp/data/unkl/ll/ΠΠ-07-Ααα βββ 
γδ //-άάά 2-θό ώ 
ύυα/Plot-LLC (3350)/"
cd "/tmp/data/unkl/ll/ΠΠ-07-Ααα βββ γδ 
//-άάά 2-θό ώ ύυα/Plot-LLC 
(3350)/"

To crash, it requires the terminal width to be narrow, under 80 columns
or so, and a fancy prompt. I'm using the colored prompt from the default
.bashrc:

PS1="\[\e]0;\u@\h: 
\w\a\]${debian_chroot:+($debian_chroot)}\[\033[01;32m\]\u@\h\[\033[00m\]:\[\033[01;34m\]\w\[\033[00m\]\$
 "

Backtrace is attached.
Program received signal SIGSEGV, Segmentation fault.
0x004b8e73 in rl_redisplay () at ../../.././lib/readline/display.c:823
823 ../../.././lib/readline/display.c: No such file or directory.
(gdb) bt
#0  0x004b8e73 in rl_redisplay ()
at ../../.././lib/readline/display.c:823
#1  0x004a5f6c in readline_internal_setup ()
at ../../.././lib/readline/readline.c:443
#2  0x004a7152 in readline_internal ()
at ../../.././lib/readline/readline.c:669
#3  readline (prompt=) at ../../.././lib/readline/readline.c:376
#4  0x00423555 in yy_readline_get ()
at /usr/homes/chet/src/bash/src/parse.y:1456
#5  0x00425b0e in yy_getc ()
at /usr/homes/chet/src/bash/src/parse.y:1390
#6  shell_getc (remove_quoted_newline=1)
at /usr/homes/chet/src/bash/src/parse.y:2299
#7  0x00428dca in read_token (command=0)
at /usr/homes/chet/src/bash/src/parse.y:3115
#8  0x0042c781 in yylex () at /usr/homes/chet/src/bash/src/parse.y:2675
#9  yyparse () at y.tab.c:1834
#10 0x00422dcf in parse_command () at .././eval.c:261
#11 0x00422ed8 in read_command () at .././eval.c:305
#12 0x004230bb in reader_loop () at .././eval.c:149
#13 0x00421d4e in main (argc=1, argv=0x7fffdfe8, 
env=0x7fffdff8) at .././shell.c:792

Bug#847935: test report

[Jason Crain]

On Mon, Jan 16, 2017 at 09:44:38AM +0100, Erik Thiele wrote:
> I reinstalled fonts-texgyre
> 
> then I installed your package:
> 
> ii libpoppler46:amd64 0.26.5-2+deb8u1.1 amd64
> 
> then I found out that neither atril not evince is linking against
> libpoppler by running ldd on those binaries. Some of my own software
> links against libpoppler and there i find your library. maybe evince is
> somehow loading libpoppler at runtime. or it does not use libpoppler!

Poppler is linked to from the backend
/usr/lib/evince/4/backends/libpdfdocument.so.

> anyway. running atril or evince does not show the micro sign.

I could be wrong about how poppler's build system works, but I think the
way the cairo backend is linked, you might need an updated
libpoppler-glib instead of libpoppler.

Bug#847935: [Pkg-fonts-devel] Bug#847935: Bug#847935: Bug#847935: Bug#847935: MICRO sign does not show

[Jason Crain]

On Thu, Jan 12, 2017 at 04:32:53PM +0100, Fabian Greffrath wrote:
> Am Donnerstag, den 12.01.2017, 12:25 +0100 schrieb Fabian Greffrath:
> > Although poppler isn't really the root of the problem, it is best to
> > fix
> > it there. I re-assign this bug to poppler (I assume you are using the
> > version in stable, right?) and try to find a patch over the weekend
> > for
> > upstream application.
> 
> Well, strange! I was just going to reproduce this issue on my own
> testing/unstable system, but -- although I am using the fonts from
> fonts-texgyre as substitutes for the base fonts -- I am not able to
> reproduce this bug.
> 
> Well, there are 22 revisions between the poppler versions in stable and
> testing, so I assume the bug has been fixed in between (it has not been
> fixed in the fonts-texgyre package between stable and testing, I have
> checked that).
> 
> So, not sure what to do with this bug report then...

I believe this has been partially fixed in poppler with
https://bugs.freedesktop.org/96994, though it's probably still possible
to have missing characters, depending on the PDF and on which fonts are
chosen.

At one time I had attempted to improve on the way poppler chooses fonts,
but it's a little tricky to do that in a way that doesn't harm the
rendering of some documents.  Ideally more documents would embed the
fonts so these issues could be avoided.

Bug#850408: terminix: (apparently) does not notify it has started

[Jason Crain]

On Fri, Jan 06, 2017 at 09:45:04AM +0100, Xavier Bestel wrote:
> Package: terminix
> Version: 1.4.0-1
> Severity: minor
> 
> Dear Maintainer,
> 
> When launching Terminix in gnome-shell, the menu has a spinning wheel
> which should stop as soon as the application has started - see:
> https://www.freedesktop.org/wiki/Software/startup-notification/
> But with Terminix the wheel is spinning for too long, apparently it just
> stops after a while because of a timeout. The problem is that the menu
> is then unusable until the timeout happens.

I assume you are using wayland?  For me, it's not just Terminix, it's
all apps when running under wayland.

Bug#848998: evince search in jessie entirely broken, already fixed in 3.22 that is not available in jessie

[Jason Crain]

On Wed, Dec 21, 2016 at 07:36:39PM +0300, Sergey Organov wrote:
> The "search" function in PDF documents in evince 3.14 (currently in stable
> 'jessie') won't jump to corresponding page of the document for found matches.
> 
> This is already fixed in 3.22 available in testing/unstable. Please consider 
> to
> backport this version to jessie.

Version 3.22 is not going to be backported.  That would require
backporting half of GNOME and at that point, you may as well just run
testing.  It might be possible to backport the patches that fix that
specific issue as a stable update.  The patches in
https://bugzilla.gnome.org/730252 are pretty simple.  But it depends on
whether Debian's GNOME maintainers and the stable release managers
consider the issue important enough for a stable update and I have
doubts about whether it meets the 'truly critical functionality problem'
guideline.

There is a simple workaround; just view the document in continuous mode
instead of non-continuous mode.

Bug#847128: evince: different rendering of two copies of same PDF

[Jason Crain]

On Mon, Dec 05, 2016 at 09:49:30PM +0100, Lionel Elie Mamane wrote:
> When I load
> https://www.belgocontrol.be/html/belgocontrol_static/eaip/eAIP_Main/pdf/EB_Amdt_2016_012_en.pdf
> in Iceweasel, which is configured to open PDF files with Evince, the
> PDF opens in evince and displays bad.

Your images looks similar to https://bugzilla.gnome.org/775123, where
the text is stretched oddly.  When I looked at it previously, it was a
heisenbug, where the rendering is sometimes right, and trying to run it
in a debugger made the problem go away.

Bug#845530: evince: Search doesn't navigate (by index or 'next') to found results

[Jason Crain]

Control: tags -1 + moreinfo

On Thu, Nov 24, 2016 at 10:29:33AM +, Conrad J.C. Hughes (for Debian 
package stuff) wrote:
> When I search for text in Evince, it finds results but doesn't seem to
> support navigation to those results.  For example, if I search for
> "thanks to" in this document:
> 
>   http://www.blogs.hss.ed.ac.uk/crag/files/2016/10/algocracy.pdf
> 
> I get a result, but neither by clicking or double-clicking on the result
> in the results index pane, nor by clicking on the "next" or "previous"
> arrows, can I navigate to that result.  I have to manually scroll through
> the document until I reach the (correctly highlighted) text.

Does it work if you view the document in continuous mode?  There was a
bug in evince https://bugzilla.gnome.org/730252.

Bug#809388: evince: evince aborts after a few seconds without an error message

[Jason Crain]

On Sun, Nov 20, 2016 at 06:36:31AM +0100, Janusz S. Bień wrote:
> On Fri, Nov 18 2016 at 21:50 CET, ja...@inspiresomeone.us writes:
> > Control: tags -1 - moreinfo unreproducible
> >
> > On Wed, Nov 16, 2016 at 05:20:52PM +0100, Janusz S. Bień wrote:
> >> Another crash,  still a not typical one:
> >
> > I might have found what's causing it.  Try setting environment variable
> > G_SLICE=always-malloc before evince runs and see if that stops the
> > crashes.
> 
> Unfortunately the crashes still occur.

I'm not familiar with AUCTeX.  I suppose you open the .tex file in
emacs, press some key combination, then it regenerates the PDF and runs
evince?  If it uses the dbus interface as I suspect, you'll have to go
through some extra hoops to set the environment variables.  I think
running something like "G_SLICE=always-malloc dbus-run-session emacs
file.tex" will work.  Try it by setting GTK_THEME=Adwaita:dark first to
see if that environment variable gets passed through to the evince
process.

The problem I found showed itself during document reloading.  As
mentioned in the bug report at https://bugs.freedesktop.org/98786, if
you open that document in evince and hold down Ctrl-R to continuously
reload the document, it will crash within a few minutes.

Bug#809388: evince: evince aborts after a few seconds without an error message

[Jason Crain]

Control: tags -1 - moreinfo unreproducible

On Wed, Nov 16, 2016 at 05:20:52PM +0100, Janusz S. Bień wrote:
> Another crash,  still a not typical one:

I might have found what's causing it.  Try setting environment variable
G_SLICE=always-malloc before evince runs and see if that stops the
crashes.

Bug#843810: libgtk-3-0: gtk3 >= 3.21.4-1 does not show a menu in Firefox, virt-manager and probably many others

[Jason Crain]

On Wed, Nov 09, 2016 at 09:26:34PM +0100, Marc-Christian Petersen wrote:
> there seems to be a problem with gtk3 >= 3.21.4-1. Please see #839029 also.
> When using gtk3 3.21.4-1 or above, virt-manager, Firefox and maybe also others
> are not able to show a menu (when pressing Alt-F for example). Downgrading 
> gtk3
> to 3.20.9-1 fixes it, so Firefox and virt-manager are usable again.

Alt-F in firefox and virt-manager works fine for me.  What version of
firefox are you using and where did you get it from?  What desktop
environment are you using?  Are you using any themes or the default
Adwaita theme?  Does it work differently if you try in wayland vs xorg?

Bug#841082: epiphany-browser: crash when trying to show desktop notifications

[Jason Crain]

On Mon, Oct 17, 2016 at 07:12:48PM +0200, Sergio Villar Senin wrote:
> Yes, it's properly installed 3.22.0. I've just moved to 3.22.1 and it
> seems it did the trick.

I see that epiphany-browser's depends on epiphany-browser-data needs to
be bumped.  The new package includes the schema and the previous version
doesn't.

Bug#841082: epiphany-browser: crash when trying to show desktop notifications

[Jason Crain]

On Mon, Oct 17, 2016 at 04:23:28PM +0200, Sergio Villar Senin wrote:
>* What led up to the situation?
> 
>   Visit https://davidwalsh.name/demo/notifications-api.php and click on 
> "Show a notification" button
> 
>* What was the outcome of this action?
> 
>  The browser crashed. Launching it from the terminal allowed me to see 
> this error.
> 
>  (epiphany:760): GLib-GIO-ERROR **: Settings schema 
> 'org.gnome.Epiphany.host' is not installed

1) Do you have this file present?
   /usr/share/glib-2.0/schemas/org.gnome.epiphany.host.gschema.xml

2) Do you have epiphany-browser-data installed and not in a half
   installed or half configured state?

3) Does reinstalling epiphany-browser-data fix it?

4) Does running the following command fix it?
   glib-compile-schemas /usr/share/glib-2.0/schemas/

Bug#832406: gnome: Region & Language: Language only English & Formats only US?

[Jason Crain]

Control: reassign -1 gnome-control-center 1:3.14.2-3
Control: severity -1 wishlist

On Mon, Jul 25, 2016 at 10:52:11AM +0300, Evan wrote:
>* What led up to the situation?
> trying to set the correct Format for my area
>* What exactly did you do (or not do) that was effective (or
>  ineffective)?
> clicked on Formats or Language then "3-dots", then typing in the search box, 
> trying to choose a different Format
>* What was the outcome of this action?
> "No regions found"
>* What outcome did you expect instead?
> A list of available Regions

The region panel only shows you locales which you have installed.  It
does "locale -a" essentially to make the list.  You can run
"dpkg-reconfigure locales" as root to actually generate those locales,
and afterward they will show in the region panel.

It would be nice though if the region panel could do that itself.

Bug#837771: gvfs-open: fails to attach file to email

[Jason Crain]

Control: unmerge -1
Control: reassign -1 icedove 1:45.2.0-1
Control: notforwarded -1
Control: retitle -1 icedove does not handle attach parameter in mailto url

On Wed, Sep 14, 2016 at 08:26:24AM -0400, Antoine Beaupré wrote:
> When I run Gnome (or at least configure my environment with
> XDG_CURRENT_DESKTOP=GNOME), xdg-email fails to send an email with an
> attachment. Usually, I can send an email with a file attached with:
> 
> xdg-email --utf8 --to anar...@example.com --subject test --body "foo"  
> --attach /etc/motd 
> 
> This starts mutt and everything works more or less correctly. But then
> with the above GNOME environment, xdg-email tries to start Icedove
> through gvfs-open. (If it was Thunderbird, mind you, it could work
> because there are special provisions for xdg-email to use Thunderbird,
> but this is Debian with its fork so that fails.)
> 
> It does this with:
> 
> gvfs-open 
> 'mailto:?to=anar...@example.com=test=foo=/etc/motd'
> 
> That starts Icedove with all fields properly populated, but no
> attachment.
> 
> It seems to me this is a bug in gvfs-open.

Sorry, I misunderstood what you were asking for when I reassigned this
to xdg.  But still it's not a bug in gvfs.  gvfs expects the mailto app
to handle the url.  It doesn't do any per application specialization
like xdg-email does, it just passes the url directly to icedove, so
icedove needs to be fixed to handle the attach parameter.

Bug#807319: gnome-documents: enable additional book types

[Jason Crain]

Control: forwarded -1 https://bugzilla.gnome.org/740970

On Mon, Dec 07, 2015 at 04:47:37PM +0530, Ritesh Raj Sarraf wrote:
> Attached is a patch which widens the supported format for GNOME Books.
> This patch adds/enables document type PDF and  COLLECTIONS.
> 
> This is my first time dealing with Javascript so I'm not much versed
> with it. I don't know what COLLECTIONS here is being referred to. But my
> patch will generalize it for all types.
> 
> I think it should be okay to have the book types (PDF and Collection)
> show up under Books too, than not showing/rendering anything at all.

I doubt upstream would accept a patch like that, since I think the idea
for PDFs is that they want a way to manually tag PDFs as being books so
receipts and other things don't show up in gnome books.  There's some
discussion about that in https://bugzilla.gnome.org/704316.

Bug#807320: TypeError: currentType is null

[Jason Crain]

On Sun, Oct 16, 2016 at 12:05:41AM +0530, Ritesh Raj Sarraf wrote:
> On Sat, 2016-10-15 at 10:23 -0500, Jason Crain wrote:
> > Are you still having this issue?  It's working for me and the upstream bug
> > report https://bugzilla.gnome.org/745741 has been closed as unreproducible.
> 
> I am not seeing the issue any more. But then, in its current form in 3.22.0, I
> can't really verify it.
> 
> In previous version (3.18) where I reported it, it was able to render epub
> format. Now (3.22), it doesn't really render anything at all. Just blank 
> pages.

It's strange that epub doesn't work for you.  I tried an epub file from
http://www.gutenberg.org/ebooks/375 and it worked, though zooming is not
implemented which makes it much less useful.

Bug#807320: TypeError: currentType is null

[Jason Crain]

Control: tags -1 + moreinfo

On Mon, Dec 07, 2015 at 04:49:57PM +0530, Ritesh Raj Sarraf wrote:
> Package: gnome-documents
> Version: 3.18.2-1
> Severity: important
> 
> I get the following error when running GNOME Books. I am not sure if it
> is a warning or an error. Gjs tags it as a warning but JS treats it as
> an error with an exception.
> 
> I am filing it here, for now. If I get a fix, I'll post on this bug.
> 
> Dec 07 16:42:40 learner org.gnome.Books[9989]: (org.gnome.Books:16272): 
> Gjs-WARNING **: JS ERROR: TypeError: currentType is null
> Dec 07 16:42:40 learner org.gnome.Books[9989]: 
> QueryBuilder<._buildWhere/<@resource:///org/gnome/Books/js/query.js:106
> Dec 07 16:42:40 learner org.gnome.Books[9989]: 
> QueryBuilder<._buildWhere@resource:///org/gnome/Books/js/query.js:105
> Dec 07 16:42:40 learner org.gnome.Books[9989]: 
> wrapper@resource:///org/gnome/gjs/modules/lang.js:178
...

Are you still having this issue?  It's working for me and the upstream bug
report https://bugzilla.gnome.org/745741 has been closed as unreproducible.

Bug#805192: gnome-documents: crash when trying to search

[Jason Crain]

Control: tags - 1 + moreinfo

On Sun, Nov 15, 2015 at 08:41:51PM +0100, Sergio Villar Senin wrote:
>* What led up to the situation?
> 
>1- Open gnome-documents
>2- Click on search button
>3- Type 1 letter
> 
>* What exactly did you do (or not do) that was effective (or
>  ineffective)?
> 
>  When launching gnome-documents in console I get the following:
>  *** Error in `/usr/bin/gjs': malloc(): memory corruption (fast): 
> 0x7f9ab4115f30 ***
>  Abort
> 
>* What was the outcome of this action?
> 
>Crash.

Are you still having this problem?  If so, please install debug symbols
and post a stack trace.

Bug#840736: gnome-maps: always offline cant browse any map

[Jason Crain]

Control: tags -1 + moreinfo

On Fri, Oct 14, 2016 at 08:58:34PM +1100, Fulano Diego Perez wrote:
> cannot browse any maps
> 
> always reports offline with active internet comms
> 
> ignore missing pkgs below - report bug has confusion with this

Also, are you able to access http://tile.openstreetmap.org and do you
see any message if you run from a terminal, especially 'connection
failed' messages?

Bug#827335: evince: assertion failed: (wayland_display->cursor_theme_name)

[Jason Crain]

On Wed, Oct 12, 2016 at 03:13:33PM +0200, Michael Biebl wrote:
> I've tried running evince under GNOME/Wayland and I don't have a
> /run/user/*/weston-shared-* directory
> 
> Since I know nothing about apparmor, I'd prefer having input from the
> apparmor and wayland maintainers first, if the proposed patch is
> sensible. I don't like to simply guess.

I can't speak for what makes a sensible apparmor policy, but the
/run/user/*/weston-shared-XX file comes from libwayland-cursor.  It
creates, immediately unlinks, and mmaps the file.

Bug#840430: glib2.0: segfaults in nemo when opening files from the desktop

[Jason Crain]

On Wed, Oct 12, 2016 at 10:18:53AM +0900, Norbert Preining wrote:
> Hi Jason,
> 
> > It's a little bit suspicious that in frame 1, the _object pointer is
> > invalid (0x2).  Anyway, the debug packages for gtk are now automatic
> 
> Hmm, I retried, and now a get a different backtrace. It segfaults
> at 100% of instances, but it seems always in different regions:
> (gdb) bt
> #0  0x7f4e6dd5fb4e in g_slice_free1 (tmem=, 
> mem=0x736e6f697463413c, ix=0)
> at ././glib/gslice.c:855
> #1  0x7f4e6dd5fb4e in g_slice_free1 (mem_size=, 
> mem_block=0x736e6f697463413c)
> at ././glib/gslice.c:1122
> #2  0x7f4e6dd3daad in g_list_foreach (list=, 
> list@entry=0x2ebd4a8, func=0x7f4e6f98c210, user_data=user_data@entry=0x0) 
> at ././glib/glist.c:1005
> #3  0x7f4e6dd3dacb in g_list_free_full (list=0x2ebd4a8, 
> free_func=)
> at ././glib/glist.c:220
> 
> > debug packages, as in https://wiki.debian.org/AutomaticDebugPackages.
> 
> Thanks, activated and installed.
> 
> Do you have any idea how to debug such a beast?

You could try setting the environment variable G_SLICE=debug-blocks so
glib will abort if it detects invalid memory releases.  That will
hopefully tell you where the invalid release happens, instead of
possibly crashing much later after memory has been corrupted.

Alternatively, valgrind can help find invalid memory reads and writes.
You'll want to set environment variable G_SLICE=always-malloc while
using valgrind with glib, and you can ignore the leak checking stuff
since you aren't looking for memory leaks.

I don't know how much help I can be though, because nemo isn't crashing
on my machine.

Bug#827335: evince: assertion failed: (wayland_display->cursor_theme_name)

[Jason Crain]

On Tue, Oct 11, 2016 at 09:44:48PM +0200, Michael Biebl wrote:
> Am 11.10.2016 um 19:52 schrieb Diane Trout:
> > --- debian/evince/etc/apparmor.d/usr.bin.evince 2016-10-10
> > 23:32:09.826405495 -0700
> > +++ /etc/apparmor.d/usr.bin.evince  2016-10-11 10:45:06.401131452
> > -0700
> > @@ -65,8 +65,9 @@
> >owner @{HOME}/** rw,
> >owner /media/**  rw,
> >owner @{HOME}/.local/share/gvfs-metadata/** l,
> >owner /{,var/}run/user/*/gvfs-metadata/** l,
> > +  owner /{,var/}run/user/*/weston-shared-* rw,
> 
> Does anyone know why this rule includes /var/run/user?
> We are talking about XDG_RUNTIME_DIR here, which is /run/user/ and I
> never saw it referenced as /var/run/user

I don't know.  If I delete every occurance of {,var/} from the file it
still works.  It could be left over cruft from this release goal, to
transition from /var/run to /run:
https://wiki.debian.org/ReleaseGoals/RunDirectory

Bug#827335: evince: assertion failed: (wayland_display->cursor_theme_name)

[Jason Crain]

Control: tags -1 - moreinfo + patch

On Tue, Oct 11, 2016 at 10:52:20AM -0700, Diane Trout wrote:
> I made the change below to /etc/apparmor.d/usr.bin.evince and tested it
> first with rwl and then rw (both of which worked).

Great!  I've installed apparmor and I can confirm that it breaks under
wayland and that your patch fixes it.

Bug#840430: glib2.0: segfaults in nemo when opening files from the desktop

[Jason Crain]

On Tue, Oct 11, 2016 at 11:14:14PM +0900, Norbert Preining wrote:
> nemo (file manager of cinnamon) is happily crashing in gobject code.
> I have installed all kind of debug symbol packages, and even rebuilt
> libgtk3 and installed the (otherwise not available) debug symbols, 
> and then I get the following backtrace:
> Core was generated by `nemo'.
> Program terminated with signal SIGSEGV, Segmentation fault.
> #0  0x7f905fa15835 in g_type_check_instance_is_fundamentally_a (
> type_instance=type_instance@entry=0x2, 
> fundamental_type=fundamental_type@entry=80)
> at ././gobject/gtype.c:4022
> 4022  ././gobject/gtype.c: No such file or directory.
> [Current thread is 1 (Thread 0x7f9062885f40 (LWP 31652))]
> (gdb) bt
> #0  0x7f905fa15835 in g_type_check_instance_is_fundamentally_a 
> (type_instance=type_instance@entry=0x2, 
> fundamental_type=fundamental_type@entry=80) at ././gobject/gtype.c:4022
> #1  0x7f905f9f5af5 in g_object_unref (_object=0x2) at 
> ././gobject/gobject.c:3082
> #2  0x7f9061360e46 in  () at /usr/lib/x86_64-linux-gnu/libgtk-3.so.0
> #3  0x7f90613612de in  () at /usr/lib/x86_64-linux-gnu/libgtk-3.so.0
> #4  0x7f9061361272 in  () at /usr/lib/x86_64-linux-gnu/libgtk-3.so.0
> #5  0x7f9061361272 in  () at /usr/lib/x86_64-linux-gnu/libgtk-3.so.0
> #6  0x7f90613621f5 in  () at /usr/lib/x86_64-linux-gnu/libgtk-3.so.0
> #7  0x7f9060fbb3e8 in  () at /usr/lib/x86_64-linux-gnu/libgdk-3.so.0
> #8  0x7f905f71868a in g_main_context_dispatch (context=0x1b134a0) at 
> ././glib/gmain.c:3201
> #9  0x7f905f71868a in g_main_context_dispatch 
> (context=context@entry=0x1b134a0)
> at ././glib/gmain.c:3854
> #10 0x7f905f718a40 in g_main_context_iterate 
> (context=context@entry=0x1b134a0, block=block@entry=1, 
> dispatch=dispatch@entry=1, self=) at ././glib/gmain.c:3927
> #11 0x7f905f718aec in g_main_context_iteration 
> (context=context@entry=0x1b134a0, may_block=may_block@entry=1) at 
> ././glib/gmain.c:3988
> #12 0x7f905fcd370d in g_application_run (application=0x1b19130 
> [NemoApplication], argc=argc@entry=1, argv=argv@entry=0x7fff35af2cb8) at 
> ././gio/gapplication.c:2381
> #13 0x0042fe3f in main (argc=1, argv=0x7fff35af2cb8) at 
> nemo-main.c:104
> (gdb) 
> 
> Looking at the code in gobject/gtype.c:4022 I am a bit shocked that it
> is crashing there ... nothing suspicous.
> 
> Anyway, I am happy to rebuild all kind of stuff, install more dbg packages,
> and help find the problem.
> 
> BTW, this is 100% reproducible, nothing to worry.

It's a little bit suspicious that in frame 1, the _object pointer is
invalid (0x2).  Anyway, the debug packages for gtk are now automatic
debug packages, as in https://wiki.debian.org/AutomaticDebugPackages.

Bug#827335: evince: assertion failed: (wayland_display->cursor_theme_name)

[Jason Crain]

On Tue, Oct 11, 2016 at 12:04:49AM -0700, Diane Trout wrote:
> Interestingly if I rebuild the debian package and run it from the build
> directory it works.
> 
> ie in the result of running apt-get source evince
> 
> this command works
> 
> ./debian/tmp/usr/bin/evince ~/dl/ENCODE\ Consortium\ Meeting_\
> Encyclopedia\ 2016.pdf 
> 
> and this one doesn't
> 
> evince ~/dl/ENCODE\ Consortium\ Meeting_\ Encyclopedia\ 2016.pdf 
> 
> (evince:6887): Gdk-WARNING **: Failed to load cursor theme DMZ-White
> 
> (evince:6887): Gdk-WARNING **: Failed to load cursor theme DMZ-White
> **
> Gdk:ERROR:/build/gtk+3.0-dN7rjx/gtk+3.0-
> 3.22.1/./gdk/wayland/gdkdisplay-
> wayland.c:1012:_gdk_wayland_display_get_scaled_cursor_theme: assertion
> failed: (display_wayland->cursor_theme_name)

It's aborting because loading the cursor theme is failing for some
reason.  And this may be a wild guess, but can you check permissions on
$XDG_RUNTIME_DIR ?  You'll need write access to that directory.

Bug#827335: evince: assertion failed: (wayland_display->cursor_theme_name)

[Jason Crain]

Resending this to the bts because it seems it didn't make it through the
first time.

On Tue, Oct 11, 2016 at 12:04:49AM -0700, Diane Trout wrote:
> Interestingly if I rebuild the debian package and run it from the build
> directory it works.
> 
> ie in the result of running apt-get source evince
> 
> this command works
> 
> ./debian/tmp/usr/bin/evince ~/dl/ENCODE\ Consortium\ Meeting_\
> Encyclopedia\ 2016.pdf 
> 
> and this one doesn't
> 
> evince ~/dl/ENCODE\ Consortium\ Meeting_\ Encyclopedia\ 2016.pdf 
> 
> (evince:6887): Gdk-WARNING **: Failed to load cursor theme DMZ-White
> 
> (evince:6887): Gdk-WARNING **: Failed to load cursor theme DMZ-White
> **
> Gdk:ERROR:/build/gtk+3.0-dN7rjx/gtk+3.0-
> 3.22.1/./gdk/wayland/gdkdisplay-
> wayland.c:1012:_gdk_wayland_display_get_scaled_cursor_theme: assertion
> failed: (display_wayland->cursor_theme_name)

It's aborting because loading the cursor theme is failing for some
reason.  And this may be a wild guess, but can you check permissions on
$XDG_RUNTIME_DIR ?  You'll need write access to that directory.

Bug#723071: evince: Send To dunction does nothing

[Jason Crain]

Control: tags -1 + moreinfo

On Tue, Sep 13, 2016 at 06:18:31PM +0200, Darshaka Pathirana wrote:
> just wanted to tell you that the problem still exists in Debian/Jessie. But
> this time it does not fail silently but displays a big red banner with "Could
> not send current document" and "Failed to execute child process "(null)" (No
> such file or directory)". See attached screenshot.

Do you have the "nautilus-sendto" program installed?

Bug#827335: evince: assertion failed: (wayland_display->cursor_theme_name)

[Jason Crain]

On Tue, Jun 14, 2016 at 06:58:06PM -0700, Diane Trout wrote:
> I launched evince running under wayland from gnome terminal and it immediately
> crashed with an
> assertion failed.
> 
> (evince:4492): Gtk-WARNING **: Theme parsing error: :56:20: The 
> 'gtk-key-
> bindings' property has been renamed to '-gtk-key-bindings'
> [New Thread 0x7f50825f6700 (LWP 4500)]
> **
> Gdk:ERROR:/build/gtk+3.0-UYMeaD/gtk+3.0-3.20.6/./gdk/wayland/gdkdisplay-
> wayland.c:827:_gdk_wayland_display_get_scaled_cursor_theme: assertion failed:
> (wayland_display->cursor_theme_name)
> 
> Program received signal SIGABRT, Aborted.

I haven't been able to reproduce this.  Do you still see this problem
under more recent versions of gtk and evince?  Do you have some cursor
theme enabled and does the problem stop if you disable the cursor theme?
Does it only happen under wayland and not xorg?

Bug#840328: evince: Evince does not received correct keystroke under Wayland

[Jason Crain]

Control: tags -1 + moreinfo

On Mon, Oct 10, 2016 at 06:25:12PM +0200, Bruno BEAUFILS wrote:
> I use evince under gnome which runs on top of XWayland on stretch.
> 
> My laptop keyboard has a physicial UK layout.
> 
> I configured the keyboard through Gnome Control Center to use another layout
> (french one).
> 
> In any applications (or at least all the one I used so far) when I type the
> physical Q key, the application receive the A keystroke as expected.
> 
> It is not the case in evince (and as far as I know for now) only in evince:
> when I press the physical Q key, it receives the Q keystroke without respect 
> of
> any keyboard settings layout.
> 
> I am unsure where to report this bug (Xwayland, Gnome or Evince) but since I 
> am
> able to reproduce that bug only with evince so far, I try here.

I haven't been able to reproduce the behavior you describe.  Evince is
following the input sources the same as any other gnome app.  Here's
what I've tried:

1. Go into the gnome settings menu for region & language.
2. Add a new input source.  I picked French, so now I have two sources,
   English (US) and French.
3. An input source selecter widget appears in the gnome top bar showing
   "en" as my current input source.
4. I test this in gedit.
a. Type a few q's in gedit and see q's are entered.
b. Use the widget to switch to "fr" input source.
c. Type a few q's in gedit and see a's are entered.
d. Use widget to switch back to "en".
5. Test this in evince.  Open some random PDF.  Push the search button
   so I can see what I'm typing.
a. Same as above, type q's and see q's are entered in the search
   bar.
b. Use the widget to switch to "fr".
c. Type some q's and see a's are entered in the search bar.

Bug#822658: gnome-todo: Segmentation fault

[Jason Crain]

Control: forwarded -1 https://bugzilla.gnome.org/show_bug.cgi?id=772672
Control: tags -1 + patch

On Tue, Apr 26, 2016 at 03:40:48PM +0700, Trần Ngọc Quân wrote:
> On 26/04/2016 15:23, Emilio Pozuelo Monfort wrote:
> > Please install libglib2.0-0-dbg and libgtk-3-0-dbg. Also install
> > gnome-todo-dbgsym from
> > http://debug.mirrors.debian.org/debian-debug/pool/main/g/gnome-todo/gnome-todo-dbgsym_3.20.0-1_amd64.deb
> I'm using 32-bit system. I install gnome-todo-dbgsym_3.20.0-1_i386.deb
> > Then get a new backtrace.
> See the attachment.

I managed to reproduce this in a i386 chroot, and the backtrace looks
pretty much like yours.  It probably looks like that because it's in the
middle of a bunch of varargs functions.

The problem is that in a g_object_new call, the "xalign" property is
being passed an int when it should be a float/double.  x86-64 gets away
with it, I think due to alignment, but i386 gets the types and
parameters mismatched and crashes from trying to treat a TRUE as a
string.

Attached patch fixes it.  Forwarded upstream.
>From 095ddfb6cae4fdc36f70e237c3d4a6917688cf30 Mon Sep 17 00:00:00 2001
From: Jason Crain <ja...@inspiresomeone.us>
Date: Sun, 9 Oct 2016 19:42:02 -0500
Subject: [PATCH] panel-scheduled: pass xalign as float

Pass the "xalign" property to g_object_new as a float instead of an int
so it matches the property definition.  Passing it as an int crashes
i386 machines.

https://bugzilla.gnome.org/show_bug.cgi?id=772672
---
 plugins/eds/gtd-panel-scheduled.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/plugins/eds/gtd-panel-scheduled.c b/plugins/eds/gtd-panel-scheduled.c
index ffbc7cb..ddde78f 100644
--- a/plugins/eds/gtd-panel-scheduled.c
+++ b/plugins/eds/gtd-panel-scheduled.c
@@ -127,7 +127,7 @@ create_label (const gchar *text,
 "margin-left", 12,
 "margin-bottom", 6,
 "margin-top", first_header ? 6 : 18,
-"xalign", 0,
+"xalign", 0.0,
 "hexpand", TRUE,
 NULL);
 
-- 
2.9.3

Bug#840164: libglib2.0-bin: add gio executable

[Jason Crain]

On Sat, Oct 08, 2016 at 09:15:43PM -0500, Jason Crain wrote:
> glib, as of version 2.49.3, includes a new 'gio' commandline tool as a
> replacement for the gvfs-* commands.  Please include it in the package.
> It's already being built so just adding it to libglib2.0-bin.install
> should work.  Patch is attached.

FYI, I just checked the git log for gvfs, and the next (unreleased) gvfs
will replace the gvfs-* executables with scripts which print a
deprecation warning and run the new gio executable, so the next gvfs-bin
will probably need a dependency on libglib2.0-bin.

Bug#840164: libglib2.0-bin: add gio executable

[Jason Crain]

Package: libglib2.0-bin
Version: 2.50.0-2
Tags: patch
Severity: normal

Dear Maintainer,

glib, as of version 2.49.3, includes a new 'gio' commandline tool as a
replacement for the gvfs-* commands.  Please include it in the package.
It's already being built so just adding it to libglib2.0-bin.install
should work.  Patch is attached.


-- System Information:
Debian Release: stretch/sid
  APT prefers testing
  APT policy: (9000, 'testing')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.7.0-1-amd64 (SMP w/2 CPU cores)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages libglib2.0-bin depends on:
ii  libc62.24-3
ii  libelf1  0.166-2
ii  libglib2.0-0 2.50.0-2
ii  libglib2.0-data  2.50.0-2

libglib2.0-bin recommends no packages.

libglib2.0-bin suggests no packages.

-- no debconf information
Index: debian/libglib2.0-bin.install
===
--- debian/libglib2.0-bin.install	(revision 51254)
+++ debian/libglib2.0-bin.install	(working copy)
@@ -1,9 +1,11 @@
 usr/bin/gapplication
 usr/bin/gdbus
+usr/bin/gio
 usr/bin/gresource
 usr/bin/gsettings
 usr/share/bash-completion
 usr/share/man/man1/gio-querymodules.1*
+usr/share/man/man1/gio.1*
 usr/share/man/man1/glib-compile-resources.1*
 usr/share/man/man1/glib-compile-schemas.1*
 usr/share/man/man1/gdbus.1*

Bug#755415: ITA: gexiv2 -- GObject-based wrapper around the Exiv2 library

[Jason Crain]

On Tue, Oct 04, 2016 at 12:04:39PM +0200, Luca Falavigna wrote:
> Hi Jason,
> 
> 2016-10-03 13:23 GMT+02:00 Jason Crain <ja...@aquaticape.us>:
> > I would like to adopt gexiv2.  I've done a bit of work already for
> > packaging the new version and I plan on uploading it to collab-maint
> > soon.  I'll need a sponsor though.  Would you be willing to sponser it?
> 
> Great!
> Feel free to ping me when ready, I'll be glad to sponsor it.

I've uploaded the package to mentors.  I'm a little bit nervous about
upstream's changes to gexiv2.vapi.  The new version removes some things
about streams and a log hander, but I don't know enough about vala to
know if that's a problem, or whether anything in Debian uses the vala
bindings.

* Package name: gexiv2
  Version : 0.10.4-1
  Upstream Author : Jens Georg <m...@jensge.org>
* URL : https://wiki.gnome.org/Projects/gexiv2
* License : GPL-2+
  Section : libs

It builds those binary packages:

gir1.2-gexiv2-0.10 - GObject-based wrapper around the Exiv2 library - 
introspection da
libgexiv2-2 - GObject-based wrapper around the Exiv2 library
libgexiv2-dev - GObject-based wrapper around the Exiv2 library - development 
file

To access further information about this package, please visit the following 
URL:

https://mentors.debian.net/package/gexiv2

Alternatively, one can download the package with dget using this command:

  dget -x 
https://mentors.debian.net/debian/pool/main/g/gexiv2/gexiv2_0.10.4-1.dsc

More information about gexiv2 can be obtained from 
https://wiki.gnome.org/Projects/gexiv2.

Changes since the last upload:

  * New upstream version 0.10.4
  * Make compile warnings nonfatal.  Add --disable-Werror to debian/rules
override_dh_auto_configure.
  * Add fix-typelib-install-directory.patch from upstream.  The .typelib files
were being installed in $(datadir)/gir-1.0.  This patch moves them to the
correct location $(libdir)/girepository-1.0.
  * Add always-install-vapi-file.patch from upstream.  The .vapi file was only
being installed when valac was present.  This patch causes the pre-built
vapi file to be used when valac is not available.
  * Update to standards version 3.9.8
  * Switch to automatic dbgsym packages
  * Move from debhelper 9 to 10
- Remove --parallel and --with autoreconf, they're default now
  * Enable hardening flags.  Add DEB_BUILD_MAINT_OPTIONS = hardening=+all.
  * Add linker flags -Wl,-z,defs -Wl,-O1 -Wl,--as-needed
  * Remove shlibs:Depends for gir1.2-gexiv2-0.10.  There's no use for it when
the package only contains python and typelib files and it causes a warning
message.
  * Update debian/copyright.
- Upstream has assigned copyright to SFC.
- Add copyright of several new m4 macros.
- Add comment clarifying origin of gexiv2-managed-stream.h.
  * Add debian/upstream/metadata
  * Remove Build-Depends on libtool-bin.  It's no longer necessary since
upstream fixed their call to g-ir-scanner in
https://bugzilla.gnome.org/765827.
  * Add debian/control VCS fields
  * Update debian/libgexiv2-2.symbols.
  * New maintainer (Closes: #755415)

 Regards,
  Jason Crain

Bug#769080: Please reopen and fix this bug

[Jason Crain]

On Thu, Oct 06, 2016 at 03:46:36PM +0200, Cristian Rigamonti wrote:
> On Mon, Oct 03, 2016 at 10:36:39PM -0500, Jason Crain wrote:
> > It's not an evince bug.  It's a bad interaction between gtk and some window
> > managers.  I do see that the window decorations are missing in icewm on 
> > jessie,
> > but it's working on sid.  I'm suspecting that it was fixed in
> > https://bugzilla.gnome.org/746222, but backporting that might cause more
> > problems than it solves.
> 
> Thank you very much for the explanation.
> 
> I upgraded my libgtk-3-{0,bin,common} packages to the sid version 3.22.1,
> but I still get the bug. I guess I should also upgrade evince to the sid 
> version, right?
> 
> Anyway I'm not going to do it because this requires upgrading way too many 
> packages on my system,
> so I will switch to zathura or mupdf.

If you're mixing stable packages with testing or unstable packages...
please don't do that.  I'm surprised that upgrading gtk only resulted in
the bug not being fixed.  I would expect that it would make all gnome
applications stop working.

Bug#769080: Please reopen and fix this bug

[Jason Crain]

On Tue, Oct 04, 2016 at 12:21:10AM +0200, Cristian Rigamonti wrote:
> I don't understand why this bug was closed just because a workaround for a 
> particular window manager
> was found.
> 
> I'm still experiencing this bug with another window manager (icewm), so I 
> think this is clearly an
> evince bug that should be fixed.

It's not an evince bug.  It's a bad interaction between gtk and some window
managers.  I do see that the window decorations are missing in icewm on jessie,
but it's working on sid.  I'm suspecting that it was fixed in
https://bugzilla.gnome.org/746222, but backporting that might cause more
problems than it solves.

Bug#755415: ITA: gexiv2 -- GObject-based wrapper around the Exiv2 library

[Jason Crain]

Control: retitle -1 ITA: gexiv2 -- GObject-based wrapper around the Exiv2 
library
Control: owner -1 ja...@inspiresomeone.us

On Sun, Jul 20, 2014 at 05:07:43PM +0200, Luca Falavigna wrote:
> I request an adopter for gexiv2 source package.
> Upstream is very responsive, and the software is well maintained.

I would like to adopt gexiv2.  I've done a bit of work already for
packaging the new version and I plan on uploading it to collab-maint
soon.  I'll need a sponsor though.  Would you be willing to sponser it?

Bug#839178: codecrypt impossible dependencies

[Jason Crain]

On Thu, Sep 29, 2016 at 10:22:52PM +0200, ban...@openmailbox.org wrote:
>   Dear maintainer, I am trying to install codecrypt from a pinned sid repo
> on Jessie.

You are mixing the stable and unstable releases.  Do not do this.  It
will break things.  See https://wiki.debian.org/DontBreakDebian.

>   After including other dependency packages as well, apt has run into a
> unresolvable dependency requirement. libstdc++6 insists on libkolabxml1
> 1.1.0-3 while the newest version available in sid is 1.0.2-2

The problem is that codecrypt and other packages on sid use a newer
libstdc++6 version.  Packages on jessie use an older libstdc++6 version.
It's warning you that if you upgrade libstdc++6, you will break every
c++ program you have installed.

>   AFAICT this dependency conflict is a blocker even if the user is running
> unstable.

The dependencies worked fine for me when I installed it on a fully sid
system.

If you want it to work on jessie, it needs to be recompiled for jessie.
The build-depends don't look bad so you shouldn't run into problems
making a backport.

Bug#838131: [gnome-photos] Does not work on Intel Core 2 Duo CPU T6600 (64 bit)

[Jason Crain]

On 2016-09-17, Nils Asmussen  wrote:
> after upgrading to the Version 3.21.92-1, I always get an error stating
> "illegal instruction":
> 
> bash$ gnome-photos
> Illegal instruction
> 
> The relevant assembly instruction is "rdtscp".
> I checked this by substituting it by "nop"s:
> With a hexeditor I replaced "0F 01 F9" by "90 90 90". As a result I got
> the gnome-photos main screen flashing on my screen for a short time. (I
> didn't see anything of gnome-photos before.)
> 
> I conclude that changing the compiler flags to support older cpus would
> fix the problem.

>From gnome-photos' configure.ac:
> AC_MSG_CHECKING([for fast counters with rdtscp])
> AC_RUN_IFELSE([AC_LANG_SOURCE([[
> #include 
> int main (int argc, char *argv[]) { int cpu; __builtin_ia32_rdtscp (); 
> return 0; }]])],
>   [have_rdtscp=yes],
>   [have_rdtscp=no])
>   AC_MSG_RESULT([$have_rdtscp])
>   AS_IF([test "$have_rdtscp" = "yes"],
>   [CFLAGS="$CFLAGS -DEGG_HAVE_RDTSCP"])

It looks like they've recently added something called EggCounter which
which compile-time conditionally uses rdtscp.

Bug#837762: evince: warnings about "Allocating size to GtkEntry 0x... without calling gtk_widget_get_preferred_width/height"

[Jason Crain]

On Wed, Sep 14, 2016 at 12:45:19PM +0200, Julien Cristau wrote:
> I've started seeing these annoying warnings recently:
> 
> (evince:28346): Gtk-WARNING **: Allocating size to EvSidebar 0x55dcca0c2530 
> without calling gtk_widget_get_preferred_width/height(). How does the code 
> know the size to allocate?
> 
> (evince:28346): Gtk-WARNING **: GtkIconView 0x55dcca12e680 is drawn without a 
> current allocation. This should not happen.
> 
> These look like debugging messages to evince developers, not to me, I
> shouldn't have to see them.

Those messages probably point to real problems, and the EvSidebar
message has already been reported upstream at
https://bugzilla.gnome.org/767136.  Though if you really want to hide
the messages, the way to do that is to compile gtk with
--enable-debug=no.

Bug#827195: evince plugin won't open postscript

[Jason Crain]

On Wed, Sep 14, 2016 at 12:17:47AM +0200, Leon Meier wrote:
> A file which cannot be opened is, e.g.,
> http://www.math.chalmers.se/~palbin/rickard.ps
> 
> The browser hangs for a few minutes, then the blank tab is displayed.
> 
> In fact, no postscript file can be opened from the browser.

Firefox apparently does not implement whatever GTK3 support the plugin
needs.  I might need to poke a couple upstream bug reports so that the
plugin at least shows an error message instead of freezing.

It does work reasonably well in epiphany browser.

Bug#834860: libgtop: "glibtop: statvfs '/run/user/0/gvfs' failed" ten times every second in syslog

[Jason Crain]

On 2016-08-20, jpp  wrote:
> Hello, I get 2/3 messages every second in the syslog :
> org.mate.panel.applet.MultiLoadAppletFactory[21255]: glibtop: statvfs 
> '/run/user/0/gvfs' failed
> 
> More than 132000 such messages a day in the syslog, the system was freshly 
> re-installed on new disks.

I'm guessing that if you disable harddisk monitoring on your system
monitor applet it will stop?

Also, are you logging in as root?  It's strange that it's looking in
/run/user/0.  $XDG_RUNTIME_DIR is set to /run/user/$UID and $UID it
usually around 1000.  I suspect that part of the problem is that your
$XDG_RUNTIME_DIR is not set correctly or permissions on that directory
are not set correctly.

Bug#834526: gnome-shell: Application icon in "side panel" lacks transparency

[Jason Crain]

On Tue, Aug 16, 2016 at 07:27:11PM +0200, Felix Natter wrote:
> I am the maintainer of a java called freeplane [1], which has an icon
> with transparent edges. However, in current Debian unstable, the edges
> are no longer transparent in GNOME3's left icon panel. See the green
> icon in the attached screen shot.

FWIW, I tried this with Debian testing and unstable, I set freeplane as
a favorite to add it to the icon panel / dash, and the icon shows up
fine with a transparent background.  On both xorg and wayland, if that
makes a difference.

Bug#811866: More C++ help needed (Was: Bug#811866: fixed in hyphy 2.2.6+dfsg-4)

[Jason Crain]

On 2016-08-14, Andreas Tille  wrote:
> Since there was no answer to this question I just did this since it
> seems like a working solution.  Unfortunately there are further build
> issues I can't deal with:
> 
...
> /build/hyphy-2.2.6+dfsg/src/gui/HYChartWindow.cpp:3010:54: error: no matching 
> function for call to '_Formula::_Formula(_String&, NULL, bool)'
>  _Formula f (*thisString,nil,false);
>   ^
...
> /build/hyphy-2.2.6+dfsg/src/core/include/formula.h:87:5: note: candidate: 
> _Formula::_Formula(_String&, _VariableContainer*, _String*)
>  _Formula (_String&,_VariableContainer* theParent=nil,_String* 
> errorString = nil);
>  ^~~~
> /build/hyphy-2.2.6+dfsg/src/core/include/formula.h:87:5: note:   no known 
> conversion for argument 3 from 'bool' to '_String*'

Previous versions of the compiler allow conversion from false to a null
pointer, so this used to work, though maybe with a warning.  Replacing
false with nullptr or nil or whatever upstream uses for null will
probably fix it.

And also, it does look like there's a response for your other queston on
https://bugs.debian.org/811866

Bug#827195: browser-plugin-evince: browser-plugin-evince doesn't work.

[Jason Crain]

On Mon, Jun 13, 2016 at 03:19:07PM +, shirish शिरीष wrote:
> Package: browser-plugin-evince
> Version: 3.20.0-4
> Severity: important
> 
> Dear Maintainer,
> 
> I was trying to print an e-ticket and the plugin wasn't able to
> function. The following is what came on the CLI.

What web browser are you using?  The browser plugin does not work with
firefox.

Bug#702082: evince doesn't handle double-struck capital K (핂) in the bookmarks

[Jason Crain]

On Thu, Jul 28, 2016 at 03:55:56AM +0200, Leon Meier wrote:
> I believe that the file in the attachment is another manifestation of the
> same bug: we wish to have A핂B displayed in the table of contents, but we
> don't get it.

Yes, it's the same bug.  Reassigning to poppler.  I've attached a patch
at the upstream bugtracker https://bugs.freedesktop.org/97156 .

Bug#830565: evince: Segmentation fault in Evince when searching after opening a PDF

[Jason Crain]

Control: tags -1 + unreproducible moreinfo

On Sat, Jul 09, 2016 at 03:01:22PM +0200, Erik Brangs wrote:
> I was trying to use Evince to search in a PDF. After opening the PDF
> and pressing CTRL+F, I entered a single letter into the search bar.
> After this, evince crashed with a segmentation fault.

Stack trace is the same as the one in
https://bugzilla.gnome.org/724602 but unfortunately, I'm not able to
reproduce the crash with the PDF in that bug report, which makes
figuring out what happened much harder.

Can you trigger this crash reliably or did it happen just once?

Does it crash with other PDFs also?

If you are able to reliably reproduce this crash with your PDF, having
that PDF would be helpful for fixing the problem.

Bug#769080: new Evince without title

[Jason Crain]

On Wed, Jun 08, 2016 at 04:02:23PM -0500, Jason Crain wrote:
> On Tue, Jun 07, 2016 at 11:08:29AM +0200, Harald Dunkel wrote:
> > On 06/07/16 10:53, Harald Dunkel wrote:
> > > Which .fvwm2rc did you use?
> > > 
> > PPS: Would you mind to try
> > 
> > /usr/share/doc/fvwm/examples/system.fvwm2rc.gz
> 
> Using that .fvwm2rc evince acts weird.  It has a window border, so it
> doesn't look like your screenshots, but it's not resizable.  Looks like
> evince doesn't do well with fvwm2's MWMDecor option.

After some testing, I see this was "fixed" by gtk commit
03213b9509fc1df16c66194ea168aed6c15110e9.  It doesn't make the resize
handles visible, but at least the windows are resizable again.  I don't
expect this will make it in to jessie though, because it's not a very
simple patch and it also requires changes to the theme.

As a workaround, someone running into this problem on fvwm can try
editing ~/.fvwm2rc and removing the MWMDecor option to have fvwm ignore
the mwm flags.

Bug#769080: new Evince without title

[Jason Crain]

On Tue, Jun 07, 2016 at 11:08:29AM +0200, Harald Dunkel wrote:
> On 06/07/16 10:53, Harald Dunkel wrote:
> > Which .fvwm2rc did you use?
> > 
> PPS: Would you mind to try
> 
>   /usr/share/doc/fvwm/examples/system.fvwm2rc.gz

Using that .fvwm2rc evince acts weird.  It has a window border, so it
doesn't look like your screenshots, but it's not resizable.  Looks like
evince doesn't do well with fvwm2's MWMDecor option.

Bug#728358: evince: Disappearance of pictures in full screen (xephyr)

[Jason Crain]

On Fri, Jun 03, 2016 at 11:49:18AM -0500, Jason Crain wrote:
> On Thu, Oct 31, 2013 at 08:38:40AM +0100, Christophe TROESTLER wrote:
> > The attached document displays well in windowed mode and in full
> > screen, EXCEPT when the full screen happens inside a Xephyr window.
> > All other pictures of the document were displayed fine (including
> > within xephyr).
> 
> This is not related to zephyr, but the image in this document sometimes
> disappears based on zooom levels.  You can see it by running:
> 
> pdftocairo -r 100 -singlefile -png -f 1 -l 1 D.pdf a
> vs.
> pdftocairo -r 150 -singlefile -png -f 1 -l 1 D.pdf a

The PDF contains drawing commands with large numbers like:

0 0 m
0 2147483647 l
1098 2147483647 l
1098 0 l
W* n

And cairo doesn't deal well with it.  Whether anything is drawn varies
with the scale.  Attached C program reproduces this through just cairo.
#include 

int main (int argc, char *argv[]) {
cairo_surface_t *surface = cairo_image_surface_create (CAIRO_FORMAT_ARGB32,
   100, 100);
cairo_t *cr = cairo_create (surface);

// behavior varies with scale, uncommenting the below line makes it work.
// some scalings work and some don't, I don't see a pattern.
// cairo_scale (cr, 0.9, 0.9);

// cairo doesn't like this big number, clamping it also fixes it
cairo_rectangle (cr, 0, 0, 100, 2147483647);
cairo_clip (cr);
cairo_rectangle (cr, 10, 10, 80, 80);
cairo_fill (cr);

cairo_surface_write_to_png (surface, "test.png");
cairo_destroy (cr);
cairo_surface_destroy (surface);
return 0;
}

Bug#728358: evince: Disappearance of pictures in full screen (xephyr)

[Jason Crain]

On Thu, Oct 31, 2013 at 08:38:40AM +0100, Christophe TROESTLER wrote:
> The attached document displays well in windowed mode and in full
> screen, EXCEPT when the full screen happens inside a Xephyr window.
> All other pictures of the document were displayed fine (including
> within xephyr).

This is not related to zephyr, but the image in this document sometimes
disappears based on zooom levels.  You can see it by running:

pdftocairo -r 100 -singlefile -png -f 1 -l 1 D.pdf a
vs.
pdftocairo -r 150 -singlefile -png -f 1 -l 1 D.pdf a

Bug#608084: evince: does not display last page of PS file

[Jason Crain]

Control: reassign -1 libspectre 0.2.7-2
Control: found -1 0.2.7-4

On Mon, Dec 27, 2010 at 01:29:35AM +, brian m. carlson wrote:
> When viewing the document at
> , evince never
> displays the last page; it always claims to be "Loading...".  This is
> true whether I open it directly from Chromium or by hand from the
> terminal.  The terminal displays the following:
> 
>   undefined -21
[snip]

I get these same errors when using libspectre's test programs:

$ test/spectre-test SGIFreeSWLicB.1.1.ps out
Testing libspectre version: 0.2.7
process 2775: spectre_document_load: assertion `document != NULL' failed 
(spectre-document.c:57)
Rendering document
Rendering document at 2x
undefined -21
Error exporting page 4 as pdf: exporter error
Document format: PS-Adobe-3.0
Postscript language level: 2
Encapsulated PostScript: No
Number of pages: 5
Title: Microsoft Word - SGIFreeSWLicBWordV1.1020022.doc
Creator: Windows NT 4.0
For: (null)
Creation date: 12:39 4/24/2000
Document Orientation: Portrait
Page 0
Page label: 1
Page size: 586 x 767
Page orientation: Portrait
Page 1
Page label: 2
Page size: 586 x 767
Page orientation: Portrait
Page 2
Page label: 3
Page size: 586 x 767
Page orientation: Portrait
Page 3
Page label: 4
Page size: 586 x 767
Page orientation: Portrait
Page 4
Page label: 5
Page size: 586 x 767
Page orientation: Portrait
undefined -21
Error rendering page 4: render error
Rendering page 0 slice 195, 255, 195, 255
Rendering page 0 in a4 page

Reassigning to libspectre

Bug#825418: evince-previewer: Should not register as the default viewer for PDFs or other documents

[Jason Crain]

On Thu, May 26, 2016 at 12:12:43PM -0700, Josh Triplett wrote:
> Then evince-previewer should not register itself as the default viewer
> for various file types, including PDFs.  By default, opening a PDF from
> mutt or another program that respects the mailcap database (as updated
> by update-mime) will open that PDF in evince-previewer, rather than
> evince.  If evince-previewer only exists to handle print previews (which
> seems consistent with its window title of "Print Preview"), then it
> should not register itself as a handler for the MIME types that evince
> does.
> 
> /usr/share/applications/evince-previewer.desktop and
> /usr/share/applications/evince.desktop both contain an identical
> MimeType line.  I don't see anything in either file that sets a
> priority, so I don't see any way that an application (or update-mime)
> could know to prefer one over the other.

There is not a way to specify priorities for applications using .desktop
files.  The FDO spec sort of has a way to specify global defaults in a
/usr/share/applications/mimeapps.list file, but it doesn't have a good
priority mechanism and Debian doesn't include that file anyway, leaving
it up to either the desktop environment or someone using xdg-mime to set
per-user defaults.

I think it might be a good idea to remove the mime info from
evince-previewer.desktop though.  I don't see anything break when I
change it to "MimeType=" and as a bonus, this removes it from some menus
where it shouldn't be visible anyway.

Bug#502399: diarhesis characters make selection of words impossible

[Jason Crain]

On Thu, Oct 16, 2008 at 10:43:56AM +0200, martin f krafft wrote:
> Given the attached PDF file, it's impossible with evince to select
> single words for copying, which seems to be due to the diarhesis
> characters.

This is fixed in more recent versions of poppler.  There are probably
several commits that improved this, but the latest one is commit
040b316f0cb1ac933dce616fabe24c93f96fe1cd in poppler 0.33.0.

Bug#825226: jessie-pu: package evince/3.14.1-2+deb8u1

[Jason Crain]

Package: release.debian.org
Severity: normal
Tags: jessie
User: release.debian@packages.debian.org
Usertags: pu

I would like to patch a couple of bugs in the stable release of evince.

* reload-page-count.patch.  Fix crash when document has pages removed and
is reloaded.  Update the end page index when the document is reloaded
(Closes: #805276).  This effects people who use evince as a previewer while
working in latex or a similar typesetter.

* check-load-job-success.patch.  Fix crash in recent documents view when
a recent document fails to load.  Check whether a document's load job failed
before creating it's thumbnail (Closes: #762719).  It's possible to get into a
situation where evince is unusable because you've recently viewed a password
protected PDF.  Evince is unable to create a thumbnail for the PDF and crashes
on startup.

debdiff is attached
diff -Nru evince-3.14.1/debian/changelog evince-3.14.1/debian/changelog
--- evince-3.14.1/debian/changelog  2015-03-06 02:36:49.0 -0600
+++ evince-3.14.1/debian/changelog  2016-05-19 13:40:51.0 -0500
@@ -1,3 +1,14 @@
+evince (3.14.1-2+deb8u1) stable; urgency=medium
+
+  * Add reload-page-count.patch.  Fix crash when document has pages removed and
+is reloaded.  Update the end page index when the document is reloaded.
+(Closes: #805276)
+  * Add check-load-job-success.patch.  Fix crash in recent documents view when
+a recent document fails to load.  Check whether a document's load job
+failed before creating it's thumbnail.  (Closes: #762719)
+
+ -- Jason Crain <ja...@inspiresomeone.us>  Thu, 19 May 2016 13:03:32 -0500
+
 evince (3.14.1-2) unstable; urgency=medium
 
   * Team upload.
diff -Nru evince-3.14.1/debian/patches/check-load-job-success.patch 
evince-3.14.1/debian/patches/check-load-job-success.patch
--- evince-3.14.1/debian/patches/check-load-job-success.patch   1969-12-31 
18:00:00.0 -0600
+++ evince-3.14.1/debian/patches/check-load-job-success.patch   2016-05-19 
13:40:50.0 -0500
@@ -0,0 +1,23 @@
+Description: Check whether load job succeeded
+ Check whether document load job succeeded before creating it's thumbnail.
+ This fixes a crash in the recent documents view when a document fails to load.
+Origin: upstream, 
https://git.gnome.org/browse/evince/commit/?id=921211ea2dfcff79df172e39a380074883e2b1a2
+Author: Marek Kasik <mka...@redhat.com>
+Bug: https://bugzilla.gnome.org/744049
+Bug-Debian: https://bugs.debian.org/762719
+Last-Update: 2016-05-19
+
+Index: evince-3.14.1/shell/ev-recent-view.c
+===
+--- evince-3.14.1.orig/shell/ev-recent-view.c
 evince-3.14.1/shell/ev-recent-view.c
+@@ -366,7 +366,8 @@ document_load_job_completed_callback (Ev
+ EvRecentViewPrivate *priv = data->ev_recent_view->priv;
+ EvDocument  *document = EV_JOB (job_load)->document;
+ 
+-if (g_cancellable_is_cancelled (data->cancellable) || !document) {
++if (g_cancellable_is_cancelled (data->cancellable) ||
++ev_job_is_failed (EV_JOB (job_load))) {
+ get_document_info_async_data_free (data);
+ return;
+ }
diff -Nru evince-3.14.1/debian/patches/reload-page-count.patch 
evince-3.14.1/debian/patches/reload-page-count.patch
--- evince-3.14.1/debian/patches/reload-page-count.patch1969-12-31 
18:00:00.0 -0600
+++ evince-3.14.1/debian/patches/reload-page-count.patch2016-05-19 
13:40:44.0 -0500
@@ -0,0 +1,29 @@
+Description: Check legal boundaries of accessable pages
+ Keep the accessible view end page under the limits of the document.  Sometimes
+ when a document is reloaded, it may have fewer pages making the end page
+ higher than the actual number of pages.
+Origin: backport, 
https://git.gnome.org/browse/evince/commit/?id=e6e0d29d9fed63599e736003f06428a1aea87121
+Author: Germán Poo-Caamaño <g...@gnome.org>
+Bug: https://bugzilla.gnome.org/735744
+Bug-Debian: https://bugs.debian.org/805276
+Last-Update: 2016-05-17
+
+Index: evince-3.14.1/libview/ev-view-accessible.c
+===
+--- evince-3.14.1.orig/libview/ev-view-accessible.c
 evince-3.14.1/libview/ev-view-accessible.c
+@@ -389,6 +389,14 @@ initialize_children (EvViewAccessible *s
+   child = ev_page_accessible_new (self, i);
+   g_ptr_array_add (self->priv->children, child);
+   }
++
++/* When a document is reloaded, it may have less pages.
++ * We need to update the end page accordingly to avoid
++ * invalid access to self->priv->children
++ * See https://bugzilla.gnome.org/show_bug.cgi?id=735744
++ */
++  if (self->priv->end_page >= n_pages)
++  self->priv->end_page = n_pages - 1;
+ }
+ 
+ static void
diff -Nru evince-3.14.1/debian/patches/series 
evince-3.14.1/debian/pa

Bug#825126: RFS: scantailor/0.9.12.2-2

[Jason Crain]

Package: sponsorship-requests
Severity: normal

Dear mentors,

I am looking for a sponsor for my package "scantailor"

 Package name: scantailor
 Version : 0.9.12.2-2
 Upstream Author : Nate Craun <n...@natecraun.net>
 URL : http://scantailor.org/
 License : GPL-3+
 Section : graphics

It builds those binary packages:

  scantailor - interactive post-processing tool for scanned pages

To access further information about this package, please visit the following 
URL:

https://mentors.debian.net/package/scantailor


Alternatively, one can download the package with dget using this command:

  dget -x 
https://mentors.debian.net/debian/pool/main/s/scantailor/scantailor_0.9.12.2-2.dsc

More information about scantailor can be obtained from http://scantailor.org/.

Changes since the last upload:

  * Enable hardening flags.


Regards,
 Jason Crain

Bug#822216: Info received (Bug#822216: /usr/bin/evince: PDFs protected against printing can't be viewed)

[Jason Crain]

On Tue, Apr 26, 2016 at 09:53:03AM +0200, Andreas Wirooks wrote:
> The PDF from the Link above is now unprotected and can't be used as an
> example for "protected against printing" anymore.

Do you have another PDF that could show this problem?

Bug#824723: rhythmbox: after one minute skips the song

[Jason Crain]

Control: tags -1 + unreproducible moreinfo

On Thu, May 19, 2016 at 10:19:19AM +0200, Lawrence Nuyts wrote:
> rhythmbox: after one minute skips the song
>
> Package: rhythmbox
> Version: 3.1-1
> Severity: normal

Please provide output from running 'rhythmbox -D player' in a terminal
while this occurs.

Bug#805932: libsigc++-2.0-dev: Return-statement with a value, in function returning 'void'

[Jason Crain]

On Tue, Nov 24, 2015 at 07:22:46PM +1100, Riley Baird wrote:
> I currently maintain granule. In the process of C++11 migration, I have come
> across the following error message:
> 
> /usr/include/sigc++-2.0/sigc++/adaptors/adaptor_trait.h:256:21: error: return-
> statement with a value, in function returning 'void' [-fpermissive]
> { return functor_(); }
> 
> I had thought that it was a problem with my package, but then I noticed that 
> it
> was also present in another, unrelated package (#805681), so I think that it
> might be a problem with libsigc++-2.0-dev.

I downloaded granule, applied the patch in https://bugs.debian.org/805203,
and compiled it.  I get an error at this line:

src/EditControls.cpp:
...
194  m_actgroup->add (Gtk::Action::create ((*iter).get_name (),
195
(*iter).get_stock_id (),
196
(*iter).get_name (),
197
(*iter).get_tooltip ()),
198   Gtk::AccelKey ((*iter).get_accel_key ()),
199   bind(mem_fun (*this,
200   
::on_edit_button_clicked),
201(*iter).get_name 
()));
...

gcc seems to be getting confused about whether you want std::bind or
sigc::bind and it picks the wrong one.  Maybe there is a stray using
statement somewhere.  Changing it to sigc::bind fixes that error.

Bug#824891: gtk+3.0: latest version breaks external themes

[Jason Crain]

Control: block -1 by 821949

On Fri, May 20, 2016 at 11:40:36PM +0200, Christoph Anton Mitterer wrote:
> On Fri, 2016-05-20 at 23:36 +0200, Ximin Luo wrote:
> > I've filed bug reports at #824892 and #824893
> I think mate-themes are also broken.

This is the bug for mate-themes.

Bug#769080: new Evince without title

[Jason Crain]

On Fri, May 20, 2016 at 10:24:59AM +0200, Harald Dunkel wrote:
> On 05/18/16 01:45, Jason Crain wrote:
> > Is this still a problem?  I try evince in fvwm2 and it looks fine,
> > though maybe there's been a gtk fix in the meantime.
> > 
> Yes, the problem still exists (using Jessie). Attached is a
> sample snapshot, created using imagemagick's "import -frame".
> 
> Evince's properties window has the usual decorations (see
> the second attachment).

For comparison, here's what I see on Jessie, running evince on Fvwm.
evince 3.14.1-2, libgtk-3-0 3.14.5-1+deb8u1, fvwm 1:2.6.5.ds-3

It still doesn't work well on windowmaker, but from what I can tell, on
my system it looks like a normal-ish Fvwm window.

Bug#691843: Package Adoption

[Jason Crain]

On Thu, May 19, 2016 at 09:08:34PM -0400, Javier Prats wrote:
> I'm interested in picking up this package.  How may I proceed?

These adoption and orphan bugs tend to not be monitored by anyone, so if
you want to contact the maintainer, you need to cc them on the email.

If you're new to packaging, I suggest you start with the debian-mentors
FAQ (https://wiki.debian.org/DebianMentorsFaq) and various documents it
links to.  If after reading that you still have questions, ask on the
#debian-mentors IRC channel or the debian-mentors mailing list.

Bug#645276: evince crashes or produces a blank page when printing a colour tiff image

[Jason Crain]

No longer crashes as of Jessie, but the print preview is still blank in
Sid.

Bug#769080: new Evince without title

[Jason Crain]

Control: tags -1 + moreinfo unreproducible

On Tue, Nov 11, 2014 at 09:50:32AM +0100, Harald Dunkel wrote:
> Since the most recent upgrade the evince window doesn't have a 
> title and the usual window decorations anymore. AFAICT it sets 
> some kind of Notitle attribute in the window manager hints. This 
> makes Evince almost unusable. Looks like a transient window, but
> of course its not.
> 
> Using Wheezy there is no such problem.
> 
> Window manager is fvwm2.

Is this still a problem?  I try evince in fvwm2 and it looks fine,
though maybe there's been a gtk fix in the meantime.

Bug#824405: evince: scrollbar regression with libgtk-3-0 3.20

[Jason Crain]

Control: close -1

On Mon, May 16, 2016 at 04:19:01PM +0200, Jörg-Volker Peetz wrote:
> Thanks for your hints.
> The environment variable setting shows the scrollbars permanently what I 
> don't want.
> 
> The CSS syntax has changed indeed. Your suggestion didn't do the trick 
> without a
> little modification. Using
> 
> scrollbar {
>   background-color: Thistle2; }
>   * {
> -GtkScrollbar-has-backward-stepper: true;
> -GtkScrollbar-has-forward-stepper: true; }
> 
> the scrollbar buttons appear again. For the horizontal scrollbar they are a 
> bit
> clumsy especially in the right buttom corner where vertical and horizontal
> scrollbar overlap.

Yeah, that's why I recommended the environment variable, to fix that
weirdness.

> The config file has to be ~/config/gkt-3.0/gtk.css.
> Trying ~/config/gkt-3.20/gtk.css doesn't work.
> 
> For all I care, you could close this bug.
> Many thanks for your support.

Welcome.

Bug#824405: evince: scrollbar regression with libgtk-3-0 3.20

[Jason Crain]

On Sun, May 15, 2016 at 07:23:18PM +0200, Jörg-Volker Peetz wrote:
> Jason Crain wrote on 05/15/16 19:08:
> > On Sun, May 15, 2016 at 05:47:14PM +0200, Jörg-Volker Peetz wrote:
> >> keeping it sober, I use just openbox + lxpanel, no desktop theme set. In
> >> ~/.config/gtk-3.0/gtk.css I have, beside some color settings, these 
> >> defined:
> >>
> >> .scrollbar {
> >>   -GtkScrollbar-has-backward-stepper: true;
> >>   -GtkScrollbar-has-forward-stepper: true;
> >> }
> >> .scrollbar.trough {
> >>   background-color: Thistle2;
> >> }
> >>
> >> Concerning Adwaita or themes, all dependencies of libgtk-3-common are 
> >> installed.
> > 
> > The steppers have been missing from the default gnome theme for a while,
> > so this is more about how your css override has stopped working, right?
> > 
> 
> Yes.
> With gtk-3 3.18 this configuration works.
> I also see this problem with emacs24 and with the firefox version linked 
> against
> gtk-3.
> I wasn't able to find something helpful in the web.

They've changed the syntax of the CSS.  I think the new equivalent for
that is:

scrollbar {
  -GtkScrollbar-has-backward-stepper: true;
  -GtkScrollbar-has-forward-stepper: true;
  background-color: Thistle2;
}

And you might try setting the environment variable GTK_OVERLAY_SCROLLING=0

Bug#778715: Edge scrolling is quite annoying

[Jason Crain]

Control: tags -1 + unreproducible moreinfo

On Wed, Feb 18, 2015 at 10:32:22PM +0100, Vincent Bernat wrote:
> Since some versions, evince scrolls its content when the mouse is on
> the top edge or the bottom edge. This also happens when evince doesn't
> have the focus or when the mouse is on another screen. The behavior is
> also quite unexpected.
> 
> If there is some compile-time option to disable this, it would be
> welcome. I have an hard time to understand why such a feature has been
> implemented.

I can't reproduce this on either Jessie or Sid.  Can you confirm if you
still have this problem?  What desktop environment are you using?  Does
this effect any other gtk3 apps like gedit?