[Scrollkeeper-devel] scrollkeeper-update segfaults or behaves weird.

[Dag W. <da...@wi...>]

Hi,

I'm the author of soapbox, which is a program that tricks processes ;) An=
d=20
it seems that scrollkeeper-update doesn't like this and eithers segfaults=
=20
or gets involved in a recursive bug and then segfaults.

First of all, soapbox fakes succesful completion of libc functions, so I=20
can understand that this isn't in itself a bug in scrollkeeper-update
that would occur in real world situations, but it is bugging me a lot and=
=20
should obviously be fixed.

More information about soapbox at:

	http://dag.wieers.com/home-made/soapbox/


BTW These problems occur on scrollkeeper-0.3.10-7 on RH8 and on =20
scrollkeeper-0.3.11-3 on RH9.


Problem 1: segfault
=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8
If you install soapbox, and you do:

	soapbox scrollkeeper-update -p /var/scrollkeeper

you trigger the bug where it segfaults.

	[dag@lisse ~]# soapbox scrollkeeper-update -p=20
	/var/scrollkeeper -o /usr/share/omf/eog/
	soapbox: Attempt to open64("/dev/tty", O_NONBLOCK, 0000).
	soapbox: Attempt to open64("/dev/tty", O_NONBLOCK, 0000).
	soapbox: Attempt to mkdir("/var/scrollkeeper", 0755).
	Segmentation fault

Probably because mkdir("/var/scrollkeeper") seemed to have worked and the=
=20
following function-call isn't safe. (Beware that a directory can be gone=20
after it is created)


Problem 2: recursive error
=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=A8=
=A8
If I build programs that use scrollkeeper-update in the make install=20
phase, it does something like:

	scrollkeeper-update -p /var/tmp/root-netspeed_applet-0.8/var/scrollkeepe=
r -o /var/tmp/root-netspeed_applet-0.8/usr/share/omf/netspeed_applet
	Could not create directory /var/tmp/root-netspeed_applet-0.8/var/scrollk=
eeper/dev:/dev:/dev:/dev:/dev:/dev:/dev:/dev:/dev:dev:/dev:/dev:/dev:/dev=
:/dev:/dev:/dev:/dev:/dev:/dev:/dev:/dev:/:/dev:/dev: : File name too lon=
g
	make[4]: *** [install-data-hook-omf] Segmentation fault

And segfaults eventually. The cause of this might be due the same=20
or a very similar problem.


I'm not subscribed to his list, so it would be nice to keep me CC'ed.

Thanks in advance !
--   dag wieers,  da...@wi...,  http://dag.wieers.com/   --
=ABAny errors in spelling, tact or fact are transmission errors=BB