[uClibc] Re: uclibc perl - use miniperl

[Guido Schimmels]

Am 05.11.2003 07:24:19 schrieb(en) Manuel Novoa III:
> On Wed, Nov 05, 2003 at 05:58:07AM +0100, Guido Schimmels wrote:
> > Am 30.10.2003 03:39:49 schrieb(en) Miles Bader:
> >
> > >Of course, 800K is still pretty big; if possible it's better to
> just
> > >rewrite in awk (mawk is about 80K using shared libraries, and 115K
> > >_statically_ linked;
> >
> > Or 95k (instead of 115k) if you use the dietlibc.
> 
> Over a month ago, I spent about an hour browsing their stdio  
> subsystem
> and spotted about 10 obvious bugs... 3 or 4 of which could result in
> data loss.  I even posted a list of files to check to the dietlibc
> mailing list.  See  
> http://article.gmane.org/gmane.linux.lib.dietlibc/332
> for the post.

Probably because FvL approaches stdio with garlic and a cross :)
He uses his -lowfat library for that.

> If you want to use dietlibc, go ahead.  I've got nothing personal
> against it, and in fact have spent perhaps 10 hours in the last
> 2 months writing emails to both Felix and the dietlibc list to try
> to wake them up to the need for testing and code review.  For some
> of the other public postings, see

I use it mainly for daemons. The memory footprint is simply too sweet  
to resist. As I mostly use the fefe/djb/skaware/dreesen/pape  
replacements for the ancient crufty securitynightmarish unix stuff I  
guess I'm safe. O, I've linked the recent module-init-tools against  
dietlibc too. No obvious problems so far.
Due to the limited capabilities of the dietlibc you can't use it for  
most things anyway.