RPM Community Forums

Mailing List Message of <rpm-devel>

Re: Support of non-Linux/*BSD platforms?

From: Jeff Johnson <n3npq@mac.com>
Date: Wed 10 Sep 2008 - 20:41:15 CEST
Message-id: <417D044F-4B75-49DB-A756-27524ABAB729@mac.com>

On Sep 10, 2008, at 7:39 AM, Frank Fegert wrote:

>
> Thanks to everyones work on rpm5 the built went much better compared  
> to
> my last attempt with 4.4.7. Only rpmmtree seems a bit too Linux/*BSD- 
> ish
> with its strtouq, asprintf and st_mtimespec usage. Luckily CVS already
> had some fixes. This brought up the question if the rpm5 development  
> is
> now only targeted towards Linux/*BSD-ish platforms, or if its just  
> that
> no non-Linux/*BSD build problems have been reported? IOW, are there  
> plans
> on dropping other platforms? (No offense intended!)
>

Re rpmmtree portability:

IIRC the strotuq was a recent QNX patch, as ported the POSIX function  
was used.

A portable (from XAR) replacement for asprintf is included. Personally I
believe that all usages os malloc'ing printf's are pointless. Other  
opinions
are possible, and in particular, rpm.org has decided to use rasprintf()
throughout their code base.

st_mtimespec is in glibc (and in POSIX iirc). I persoanlly see little  
reason for
Yet Another Timespec but I did not rip st_mtreespec out of the  
rpmmtree port because
st_mtimespec is also in glibc (which likely means in POSIX, but I've  
not checked).

Other *BSDisms (very few linux usage cases for mtree(8)) have been  
incorporated
within rpmmtree.c because it was expedient. There should be few  
portability issues
that are not solvaqble by
    Don't build rpmmtree if it doesn't compile on your platform.

The code will be rewritten as necessary if rpmmtree appears to be  
useful (I think mtree(8) is useful, but its
a silly opinion to discuss at length).

hth

73 de Jeff
Received on Wed Sep 10 20:41:34 2008
Driven by Jeff Johnson and the RPM project team.
Hosted by OpenPKG and Ralf S. Engelschall.
Powered by FreeBSD and OpenPKG.