[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-commits
Subject:    Re: kdesupport/mimelib
From:       Waldo Bastian <bastian () suse ! de>
Date:       2000-04-18 2:38:19
[Download RAW message or body]

On Mon, 17 Apr 2000, Don Sanders wrote:
> On Mon, 17 Apr 2000, CVS by waba wrote:
> > kdesupport/mimelib datetime.cpp,1.6,1.7
> > Author: waba
> > CVSROOT: /home/kde
> > Mon Apr 17 08:47:51 MET DST 2000
> > Update of /home/kde/kdesupport/mimelib
> > In directory zeus:/tmp/cvs-serv13521
> >
> > Modified Files:
> > 	datetime.cpp
> > Log Message:
> > WABA: Fixed 64-bit bug with time_t.
>
> Waldo are you sure about this fix?

Ehm.. well, I was :-)

> For a long time KMail wasn't working on the Alpha architecture, it was
> crashing when a composer window was created. This was due to a KMail bug
> rather than a mimelib bug, I fixed this in the HEAD branch and later in
> KMail 1.0.29.1.
>
> A week or so after I fixed it in the HEAD branch someone else independently
> fiixed it, but they fixed it in a different way. I have a vague
> recollection of this other fix being same as the one you have committed.
>
> This fix may work but I don't believe it is necessary (the real bug was in
> KMail) and it doesn't look right to me.

I dunno if this fixes a crash or not, but passing this DwUint32 to functions 
which expect a time_t is IMO asking for trouble. When/if time_t happens to be 
64-bit this function might write 64 bit into a 32 bit space and I have no 
idea where the other 32 bits go.

Looking better at the code, it is interesting to see that it has all these 
#if define(HAVE_GMTIME_R) and #if defined(USE_PORTABLE_GMTIME)  checks but 
that nobody ever defines these things. (At least not in the HEAD branch)

Cheers,
Waldo

[prev in list] [next in list] [prev in thread] [next in thread]