'tlf -n' segfaults immediately after start with SPOTLIST activated (as is the default in /usr/share/tlf/logcfg.dat). That happens both on x86 and on amd64 machines.
Digging into the problem it comes down to a classical 'index out of bounds' problem in one of the files. If called as stated above the program uses a negative array index at some point. The provided patch fixes the problem. Upstream is informed.
Created attachment 205995 [details, diff] Patch to fix the out of bounds access.
Created attachment 205996 [details, diff] Patch for tlf ebuild to apply the above patch
'tlf -n' has a nasty segfault, but the patch seems not to fix the issue
Hmmm. I had tested with stable system before. You are right, there are at least two more buffer overflows if compiled with gcc-4.4.1 and glibc-2.10. See corrected patch for fix for all three problems.
Created attachment 206254 [details, diff] New patch which includes also fixes for gcc-4.4.1 and glibc-2.10
Big thanks for the patches! +*tlf-0.9.31.2-r1 (08 Oct 2009) + + 08 Oct 2009; Víctor Ostorga <vostorga@gentoo.org> + +tlf-0.9.31.2-r1.ebuild, +files/tlf-0.9.31.2-segfault-fix.patch: + Fixing segfault if called with 'tlf -n', patch thanks to Thomas Beierlein + <tb@forth-ev.de> bug 287640 +
