MainInclusionReportSvgalib
Main Inclusion Report for svgalib
Requirements
Availability: http://archive.ubuntu.com/ubuntu/pool/universe/s/svgalib, available for i386 and amd64
Rationale:
- Build dependency of usplash in order to provide high-resolution vesa support
- Should only include libsvga1 and libsvga1-dev - binaries are unnecessary
Security:
3 CVE issues in the past
- No binaries running as root or suid/sgid, except in svgalib-bin package (not requested for inclusion); however, we generally support all binaries from a source package, so these need to be dealt with.
- Does not open any port.
Quality assurance:
- Package works out of the box without configuration.
- Package does not ask any debconf questions higher than priority 'normal'.
No showstopper Debian bugs, but lots of hardware specific crashes.
Good maintenance in Debian.
- Upstream is approximately dead, large chunks of the code are rotten
- Claims (but tends to fail) to deal with ancient hardware which we cannot support.
Standards compliance:
Meets the FHS, Debian Policy
Meets Debian library packaging guide standards.
- Standard debhelper packaging, custom, but reasonable patch system.
Dependencies:
- All in main.
Comments
<mdz> mjg59: the primary reason we keep it out of main is to prevent nasty suid binaries from creeping in through other packages <mjg59> mdz: Right <mdz> mjg59: though the main inclusion process should prevent that just as well nowadays except where Debian introduces something without our knowledge <mjg59> mdz: To be honest, I'd be quite tempted to say that we can't support all of its code <mjg59> But the chances of anyone actually having hardware old enough that svgalib claims to support it (other than through vesa) is, well, small <mdz> mjg59: would it be feasible to copy the bits you need into usplash, as with bogl? <mjg59> mdz: Not in any remotely trivial manner <mjg59> The build system is a nightmare * jsgotangco (n=jsg123@ubuntu/member/jsgotangco) has joined #ubuntu-devel <mdz> I'd like to find a way to let it into main for usplash without exposing us to any potential evils <mjg59> Sure <mdz> maybe I'm paranoid; it's probably unlikely that new suid programs are popping up due to svgalib in this day and age one would hope
Reviewers
MartinPitt: traditionally we support source packages, so the suid-ness should be either removed or thorougly audited. TBH I'm not happy about this package at all, and I feel that a hi-res boot splash does not really justify having to support this package. If this is covered by an approved spec, or mdz explicitly approves this updated report, so be it; otherwise I veto.
MainInclusionReportSvgalib (last edited 2008-08-06 16:15:08 by localhost)