This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFC]: designing customizable format attributes


On Sat, 9 Jul 2005, Kaveh R. Ghazi wrote:

> It's only "elaborate" if you force it to be so.  It's possible to
> design an extensible syntax that could be implemented incrementally
> like was done for sentinels.  But airing the issues for a design is
> necessary now so that it can be implemented on time for stage1.

Attempting to design something extensible with datastructures in such a 
premature state is just something I think would add to the long-term 
maintenance burden here.  The only incremental things I think can be done 
as yet are the gradual improvement of the existing datastructures.

> If you don't want to help, fine.  But don't tell me how to allocate my
> own time.

The fact is that any list discussion consumes the time of readers trying 
to avoid problematic new designs being made or enshrined in code and takes 
away time the readers could otherwise spend fixing bugs.  A principle that 
if design or code for new features is done during stage 3 then all 
objections, comments, proposals for different designs, etc. made when the 
code is submitted in stage 1 are to be considered just as timely as if 
made during stage 3 might help, but I don't think it could fully solve the 
problem of new development not merely being a personal choice but pushing 
on the time of hundreds of others.  There is a feedback problem here; 
people expect (by history) a long stage 3, which encourages new 
development to avoid a long break from such, which both directly reduces 
the bug-fixing time available and so prolongs stage 3 but also consumes 
time from everyone else keeping up with the new development thereby 
further reducing the bugfixing and prolonging stage 3, thereby encouraging 
more new development and design rather than biding one's time in the 
interests of a short stage 3.

-- 
Joseph S. Myers               http://www.srcf.ucam.org/~jsm28/gcc/
    jsm@polyomino.org.uk (personal mail)
    joseph@codesourcery.com (CodeSourcery mail)
    jsm28@gcc.gnu.org (Bugzilla assignments and CCs)


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]