[Top][All Lists]

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

Re: strange NSException code

From: David Ayers
Subject: Re: strange NSException code
Date: Fri, 02 Dec 2005 14:20:48 +0100
User-agent: Mozilla Thunderbird 1.0.2 (X11/20051002)

Hash: SHA1

Richard Frith-Macdonald schrieb:
> I'd like to -
> 1. abolish the distinction between debug and non-debug libraries (the 
> _d suffix to the library name)

I /violently/ agree!  Note this should porbably include the profile
suffix also.

> 2. build with the -g flag at all times (debug symbols always on)
> 3. build with 'debug=yes' by default ... to define the DEBUG 
> preprocessor constant and have debug logging facilities compiled  in ...
> so you have to explcitly build with 'debug=no' for a  production release.

I have no strong prefernce here...

> Point 1 gets rid of the problems we have where we can get debug and 
> non-debug versions of the library linked into the same executable.
> Point 2 means we *always* have some change to debug our code even in 
> production environments
> Point 3 means we don't have to keep adding 'debug=yes' during 
> development (after all, we develop a lot, but release production code 
> only occasionally) ... but it also means we have all the debug 
> facilities on a production system unless we *really* want to turn  them
> off for some good reason.

Maybe we want "CVS/SVN" versions default to debug=yes and that release
tags/tarballs default to debug=no similar to gcc's --enable-checking.
OTOH since the code generation is different, we might start missing
issues.  I believe a ran into a lot of issues in gdl2 or gsweb because
it previously was only used with either debug=yes or debug=no and DEBUG
tags where misplaced wrt variable declarations and control flow issues.

But I'm fine with a strict default set to either yes or no.

Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org


reply via email to

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