[Top][All Lists]

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

Re: Request: srand48/ drand48

From: Bruno Haible
Subject: Re: Request: srand48/ drand48
Date: Fri, 1 May 2009 02:22:12 +0200
User-agent: KMail/1.9.9

Hello Stefan,

> I assume everybody still remembers the drand48 issue ;-)

Yes: it resurfaces repeatedly :-)

> I just had the time to dive into Richards code and the comments on it. 
> Would'nt it be easier to fetch the whole thing from glibc again? Instead 
> of pulling Richards code apart and placing everything back where it came 
> from?

If it appears to be more effective to you to go that way, why should we
object? You evaluated Richard's earlier proposed code. It's your decision
to start from it, or start from scratch.

> Or is it recommended to write stuff for gnulib frm scratch?

We try to reuse code from glibc where possible, because
  - it allows us to provide the desired modules with minimum effort,
  - this code is well tested,
  - it helps glibc if we find bugs in glibc code.

Every now and then, a bidirectional sync happens, to merge modifications
from glibc into gnulib, and to propose changes done in gnulib for inclusion
in glibc. In order to ease these merges, we try to do the porting in a way
that minimizes the effect on the source code. For example, instead of
replacing '__open' with 'open' everywhere, we put
  #if !_LIBC
  # define __open open
at the top of the file.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]