Re: egcs, PATCH to fix warnings in gperf generated files ...

["Kaveh R. Ghazi" <ghazi at caip dot rutgers dot edu>]

 > From: Mark Mitchell <mark@markmitchell.com>
 > 
 > >>>>> "Kaveh" == Kaveh R Ghazi <ghazi@caip.rutgers.edu> writes:
 > 
 >     Kaveh> 	I sent mail to
 >     Kaveh> bruno@linuix.mathematik.uni-karlsruhe.de which bounced, and
 >     Kaveh> also to haible@ilog.fr which went into a black hole.  Since
 >     Kaveh> I am unable to reach the gperf maintainer, how about I
 >     Kaveh> install my patch?  We can always remove it if gperf is ever
 >     Kaveh> fixed to produce output which passes -W -Wall...
 > 
 > I think these kind of changes are not worth the effort.  Proper
 > prototypes, avoiding non-standard conversions, etc. are all worth
 > while, but I don't think that removing every warning, even that in
 > machine-generated code via editing that code, is worth it.  

	I disagree.  Clutter warnings distract people from real ones,
and also encourage people to partially or completely turn off warning
flags.  (This is not a theoretical concern, but I won't name names. :-) )

	To the extent that I can remove clutter warnings without a
serious technical downside, it provides value.  My patch removes a few
hundred warnings with only a couple of two liner patches.  So I think
its well worth it. 



 > I agree
 > that we should fix gperf, though; it should generate clean code.


	Agreed.  Once gperf is fixed, and we are using it I'll certainly
remove the workaround code. 


 > I'm sure Bruno can be found; if not, obviously, the FSF needs to find
 > a new gperf maintainer.
 > -- 
 > Mark Mitchell 			mark@markmitchell.com


	I do intend to follow up on gperf, even if my patch is
installed.  Do you have an address for Bruno which is known to work?
Is he simply on vacation? :-)

		--Kaveh
--
Kaveh R. Ghazi			Engagement Manager / Project Services
ghazi@caip.rutgers.edu		Icon CMT Corp.