[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Getting a fakeroot-ng fix into Lenny


I have just uploaded to Sid fakeroot-ng version 0.12-2. It has three changes, all but one minor in scope. 1. Upstream (which is me, but anyways) had a serious bug (#493061). It was marked "normal" when originally filed against debhelper. After some research it turned out that the problem was due to the submitter setting "fakeroot" to point to "fakeroot-ng" through the alternatives mechanism. While it may be "normal" for debhelper, for fakeroot-ng it is at least "important", possibly even "grave".

The fix had to do with initializing order of state variables in the handling of "unlinkat". Like I said - an upstream bug. Since, as upstream, I have already readied version 0.13, which does some changes to the build system, I decided to backport just this fix.

Why should it go in: From Mark's email: An "important" bug in package in an "extra" package, uploaded through unstable. One may claim that not being able to erase files when going through fakeroot is actually a "grave" level bug.

2. While building the update I noticed a FTBS twice problem with the package (bug #495174). According to info from the debian devel list, FTBS twice is a release goal for Lenny. The fix is an extremely trivial one - add a TAB or erase empty lines in the debian/rules "clean" rule. Since the demand for a "clean" target in debian/clean target is part of the Debian Policy (4.9) as a "must" instruction, I tagged this bug as Serious.

Why should it go in: fix for a release critical bug. Requires changes in white spaces only, extremely trivial.

3. A minor issue, but an equally minor fix. Lintian gives a false positive regarding fakeroot-ng 0.12-1. When attempting to detect command invocations in the maintainer scripts, it does not treat continuation lines properly. A bug was filed against it (495176). In order to remove the Lintian warning from fakeroot-ng, I change a line in postinst that had a command spread across four lines for readability to spread across three lines instead. The entire change involves removal of a backslash and a newline from a script.

Why should it go in: EXTREMELY small change, with no possibility of side effects, which cleans up the report for the package.

Due to all of the above, I kindly ask that you allow fakeroot-ng version 0.12-2 from Sid into Lenny.


Reply to: