Bug 161887 - net-libs/libsoup <=2.2.3, <=2.2.98 missing input sanitizing Denial of Service (CVE-2006-5876)
|
Bug#:
161887
|
Product: Gentoo Security
|
Version: unspecified
|
Platform: All
|
|
OS/Version: Linux
|
Status: RESOLVED
|
Severity: normal
|
Priority: P2
|
|
Resolution: FIXED
|
Assigned To: security@gentoo.org
|
Reported By: aetius@gentoo.org
|
|
Component: Vulnerabilities
|
|
|
URL:
http://seclists.org/fulldisclosure/2007/Jan/0254.html
|
|
Summary: net-libs/libsoup <=2.2.3, <=2.2.98 missing input sanitizing Denial of Service (CVE-2006-5876)
|
|
Keywords:
|
|
Status Whiteboard: B3 [noglsa] aetius
|
|
Opened: 2007-01-13 12:26 0000
|
setting status and cc'ing herd.
libsoup-2.2.99 is in the tree now as ~arch, which includes the fix for upstream
bug 391970 as linked above.
If this bug is considered a security fix that should get quick stabilization,
please CC arches yourself or let me know to do that.
@comment #2 -
Do we want to stabilize a patch on any of the lower versions? I recall
something about 2.2.9x being a development branch?
2.2.9x versions have been the minimum for GNOME since GNOME-2.14 -
ftp://ftp.gnome.org/pub/GNOME/teams/releng/2.14.0/versions
We have 2.16 stable now.
So apparently upstream considers it stable. Plus many of the (stabilized)
libsoup users in the tree demand at least 2.2.90.
As for SLOT=0 (1.99.28), I hope to get rid of that completely very soon, though
users will have to notice to uninstall it themselves, as nothing would force an
unmerge through a block.
Understood. Arches, please test and mark stable:
net-libs/libsoup-2.2.99
KEYWORDS="alpha amd64 arm hppa ia64 mips ppc ppc64 ppc sparc x86"
amd64 stable first and the best!
thanks
Created an attachment (id=107068) [details]
Test errors on Alpha
I get arather impressive amount of test errors (537212952 to be precise :) on
alpha using 2.2.99. 2.2.94 passes tests with no errors. I've attached test part
of the emerge log.
Any ideas what could cause this?
SPARC is seeing the same failures when it comes to testing as Alpha is in
comment #8
Stable for HPPA with precisely 1076425976 test errors.
A negative amount failed on x86. header-parsing is a new test introduced with
.99, as the ones also available in .98 pass successfully.
-156140 errors
FAIL: header-parsing
x86 stable, as the software works with libsoup...damn tests.
So? Should we ignore the testsuite?
How about we start using RESTRICT="test" for known failures?
sparc stable and disabled tests in the ebuild since they're known broken.
Stable on Alpha and IA64.
/vote no, it's a client DoS.
I don't know how I voted twice, with conflicting votes, but I really did mean
to vote no.
noglsa feel free to reopen if you disagree