Re: [DebianGIS] GRASS 6.0.1 for sarge and on-going work...

[Hamish <hamish_nospam@yahoo.com>]

> > .. one that gives me headaches. AFAIK G_asprintf() writes to a tmp
> > file and then reads from it which IMO is too inefficent for
> > widespread use. Still waiting for something better or perhaps a
> > better understanding of why the existing version really isn't so bad
> > after all.
> 
> A patch to a portable snprintf() which realloc() on demand, for
> instance? 


I think this really needs to be sorted out on the grass5 devel mailing
list, but as portability and correctness is a big part of this problem
maybe the Debian crowd is the perfect source for some help on this one?


There have been numerous discussions by folks much smarter & more
experienced than myself on the subject, going back years. Most recently
in these threads--  (mostly related)

http://thread.gmane.org/gmane.comp.gis.grass.devel/8768
http://thread.gmane.org/gmane.comp.gis.grass.devel/7928
http://thread.gmane.org/gmane.comp.gis.grass.devel/7566
http://thread.gmane.org/gmane.comp.gis.grass.devel/6739
 and
http://thread.gmane.org/gmane.comp.gis.grass.devel/1078
  (G_asprintf() is introduced starting halfway down)


etc.

I'm pretty sure there is more; luckily it's a pretty good search term-

search recent GRASS devel mailing list archives:
  http://news.gmane.org/gmane.comp.gis.grass.devel

search full archive:
  http://grass.ibiblio.org/searchgrass.php




Hamish