glibc bug in LFS-6.1 and its nALFS profile

Steve Prior sprior at geekster.com
Thu Oct 6 08:49:43 PDT 2005


Jeremy Huntwork wrote:
> Alexander E. Patrakov wrote:
> 
>> Tough question. The problem while Jeremy Huntwork was the project 
>> leader was that exactly the same versions of packages had to be used 
>> in the book and on the LiveCD (with the exception of ncurses because 
>> of "xterm -lc" compatibility needed for i18n purposes). I'd rather ask 
>> Justin to build the 6.1-4 CD with the following fixes:
> 
> 
> I still feel very strongly that released versions of the LiveCD should 
> follow as closely as possible the LFS book that they are based upon. 
> Their first and main purpose is to successfully build *that* version of 
> the LFS book. Any other use is secondary. The change to ncurses on those 
> CDs was negligible as far as how it affected the rest of the system, so 
> I OK'd that change.
...
> 
> -- 
> JH

I don't disagree with that philosophy JH.  What I think should be the issue
now is to determine how serious this GLIBC issue really is.  If it is a
serious security issue and therefore makes "LFS 6.1 stable" a defective release (don't
mean this as bad work against the LFS group, just that a bad apple got into
the pile), then I think it might be time for LFS 6.1.1 which the LiveCD could then track
faithfully.

Calling LFS 6.1 stable means that someone coming along and building it from
the book (or LiveCD) believes they have a reasonable platform to build from.
If they're going to run into trouble as soon as they try to put ssh on top of
it (which is about the first thing that anyone would/should want to do with
it or would you prefer telnet) they run into trouble, then I think that's worth
a point release.

It also sounds to me like there is still a problem in SSH in BLFS - from the
sound of things this is a matter of whether the sshd gets an error code or a
segfault for something it is doing incorrectly.  The error code lets it get by
and the segfault doesn't, but something is still not working as intended.

This doesn't mean that glibc should be segfaulting and apparently should be upgraded
or patched because this affects multiple programs, but either there's a bug in
ssh or at least the BLFS instructions for ssh might need to change.


Steve



More information about the alfs-discuss mailing list