Discussion:
Crash disabling plugins and (related) bug report about handling of image files
Daniel Leidert
2014-02-11 11:02:47 UTC
Permalink
Hi,

During examining [1] I found a reproducible crash:

I've created a test SVG file (empty one is good enough) test.svg, start
bluefish. Now double-clicking this file in the bluefish filebrowser
inserts an <img> tag - I _guess_, this is done by the zencoding plugin -
and this is what [1] is about. Can you confirm, this is caused by the
zencoding plugin?

Now with the next step opening the preferences dialog -> plugins: As
soon as I disable the zencoding plugin and hit "Apply", bluefish
crashes. The backtrace is attached. This one needs a fix.

Regards, Daniel
Daniel Leidert
2014-02-11 11:07:38 UTC
Permalink
Am Dienstag, den 11.02.2014, 12:02 +0100 schrieb Daniel Leidert:

[..]
Post by Daniel Leidert
Now with the next step opening the preferences dialog -> plugins: As
soon as I disable the zencoding plugin and hit "Apply", bluefish
crashes. The backtrace is attached. This one needs a fix.
Hm. Reversing these actions (first disable plugin, then double-click the
file) and closing bluefish without saving the file makes bluefish crash
directly at startup. I wonder if this is the cause of [2]

[1] https://bugs.launchpad.net/ubuntu/+source/bluefish/+bug/1248595
[2] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=731821

Regards, Daniel
Olivier Sessink
2014-02-11 21:07:01 UTC
Permalink
Post by Daniel Leidert
[1] https://bugs.launchpad.net/ubuntu/+source/bluefish/+bug/1248595
fixed in svn!

Olivier
--
Bluefish website http://bluefish.openoffice.nl/
Blog http://oli4444.wordpress.com/
Olivier Sessink
2014-02-11 15:56:31 UTC
Permalink
Post by Daniel Leidert
Hi,
I've created a test SVG file (empty one is good enough) test.svg, start
bluefish. Now double-clicking this file in the bluefish filebrowser
inserts an <img> tag - I _guess_, this is done by the zencoding plugin -
and this is what [1] is about. Can you confirm, this is caused by the
zencoding plugin?
I don't think it is related to the plugin. this is hardcoded in
filebrowser2.c. However, the crash report showed a real problem if an
invalid URL is passed to the filebrowser.

disabling a plugin only sets a 0 in a list instead of a 1, only after
restart this has any effect. So I guess this was caused by earlier
memory corruption.

try to debug with valgrind `grep valgrind src/*.c` to get the best
options to run valgrind. That will give better information where the
memory coruption is started.

Olivier
--
Bluefish website http://bluefish.openoffice.nl/
Blog http://oli4444.wordpress.com/
Daniel Leidert
2014-02-13 20:34:57 UTC
Permalink
Post by Olivier Sessink
Hi,
[..]
Post by Olivier Sessink
disabling a plugin only sets a 0 in a list instead of a 1, only after
restart this has any effect. So I guess this was caused by earlier
memory corruption.
try to debug with valgrind `grep valgrind src/*.c` to get the best
options to run valgrind. That will give better information where the
memory coruption is started.
Ok. Another thing I missed earlier: I created test.svn inside /tmp. From
a running bf instance I changed from $(HOME) to / to /tmp in the bf
filebrowser. This somehow leads to the following entry in the bf
Post by Olivier Sessink
recent_dirs: file://
and this seems to have caused the issue. After your last change to
Post by Olivier Sessink
** (bluefish:4808): WARNING **: cannot handle uri file://
cleanup_scanner, memory scancache 0(0Kb+0Kb) found 0(0Kb) fcontext 0(0Kb) = 0Kb
** (bluefish:4808): WARNING **: failed to list directory in filebrowser: Vorgang wird nicht unterstützt (15)
IMHO entries like the following shouldn't be found in session-2.0,
because these are not valid URIs:

recent_dirs: file:/
recent_dirs: file://

These are not valid URIs. How do they appear?

Regards, Daniel
--
To unsubscribe from this list: send the line "unsubscribe bluefish-dev" in
the body of a message to listar-QLpEr2logwzILq5++***@public.gmane.org or visit the list control panel at http://www.ems.ru/cgi-bin/listargate.cgi
Bluefish web site: http://bluefish.openoffice.nl/
Andrius
2014-02-14 14:34:12 UTC
Permalink
When I select root in fb, in sesion-2.0 file the entry looks like

recent_dirs:file:///
However, when I reopen bf with this entry as default dir, the fb dir_menu shows "file://" which is not valid directory. When root selected it should show just "/"

Andrius
________________________________
Sent: Thursday, February 13, 2014 10:34 PM
Subject: Re: Crash disabling plugins and (related) bug report about handling of image files
Post by Olivier Sessink
Hi,
[..]
Post by Olivier Sessink
disabling a plugin only sets a 0 in a list instead of a 1, only after
restart this has any effect. So I guess this was caused by earlier
memory corruption.
try to debug with valgrind `grep valgrind src/*.c` to get the best
options to run valgrind. That will give better information where the
memory coruption is started.
Ok. Another thing I missed earlier: I created test.svn inside /tmp. From
a running bf instance I changed from $(HOME) to / to /tmp in the bf
filebrowser. This somehow leads to the following entry in the bf
Post by Olivier Sessink
recent_dirs: file://
and this seems to have caused the issue. After your last change to
Post by Olivier Sessink
** (bluefish:4808): WARNING **: cannot handle uri file://
cleanup_scanner, memory scancache 0(0Kb+0Kb) found 0(0Kb) fcontext 0(0Kb) = 0Kb
** (bluefish:4808): WARNING **: failed to list directory in filebrowser: Vorgang wird nicht unterstützt (15)
IMHO entries like the following shouldn't be found in session-2.0,
recent_dirs: file:/
recent_dirs: file://
These are not valid URIs. How do they appear?
Regards, Daniel
--
To unsubscribe from this list: send the line "unsubscribe bluefish-dev" in
Bluefish web site: http://bluefish.openoffice.nl/
Olivier Sessink
2014-02-15 21:03:27 UTC
Permalink
Post by Daniel Leidert
Post by Olivier Sessink
Hi,
[..]
Post by Olivier Sessink
disabling a plugin only sets a 0 in a list instead of a 1, only after
restart this has any effect. So I guess this was caused by earlier
memory corruption.
try to debug with valgrind `grep valgrind src/*.c` to get the best
options to run valgrind. That will give better information where the
memory coruption is started.
Ok. Another thing I missed earlier: I created test.svn inside /tmp. From
a running bf instance I changed from $(HOME) to / to /tmp in the bf
filebrowser. This somehow leads to the following entry in the bf
Post by Olivier Sessink
recent_dirs: file://
and this seems to have caused the issue. After your last change to
fixed in svn!

Olivier
--
Bluefish website http://bluefish.openoffice.nl/
Blog http://oli4444.wordpress.com/
Loading...