Hi, On Nov 19, 2007 8:12 PM, The Rasterman Carsten Haitzler <[EMAIL PROTECTED]> wrote: > On Sat, 17 Nov 2007 11:11:55 -0600 "Nathan Ingersoll" <[EMAIL PROTECTED]> > babbled: > > > Is this a proper fix or hiding a bug in the calling application (or > > another part of the lib)? I would think the application should be > > responsible of knowing when they should unref the connection. > > personally i like defensive programming. it's better than every process having > to trap its segv's/aborts and give u a debug trace and suddenly lose all your > work/state. i think a compile or runtime definable "please btich to stdout/err > about these problems" is the best way to go (all of evas and ecore are very > defensively programmed - set environment variables to turn on "bitch mode" and > you can also enable "abort mode" for aborting on caught failures).
I'm all for debug/trace messages with runtime tunnables and so on. However, what I was calling "defensive programming" is if conn shouldn't be NULL when calling some function (specially for internal calls), we better spit out some error message or just let the program segfault so we know there's something wrong and we can fix it. Otherwise, we might be hiding some other error and it just ends up more difficult to find it. -- Ulisses ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
