From:	"Andrew Crabtree" <andrewc AT rosemail DOT rose DOT hp DOT com>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: Function Sizes
Date:	19 Aug 1997 20:52:40 GMT
Organization:	Hewlett Packard
Lines:	21
Message-ID:	<01bcace2$00dea970$45111d0f@ros51675cra>
References:	<Pine DOT SUN DOT 3 DOT 91 DOT 970819161648 DOT 21250R-100000 AT is> <33f9d88f DOT 336368 AT news DOT eclipse DOT co DOT uk>
NNTP-Posting-Host:	ros51675cra.rose.hp.com
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

> will you please summarise the *current* situation
> (RE: Using an "empty" function to determine function end)
> 
> Do any current optimisations break this method?
> 
> What about PGCC releases (Andrew?)

None that I know of.  The most likely thing to cause problems here would be
feeding profile data to the 
linker and asking that it use the profile data to reorder function
locations to maximize instruction cache hits.

AFAIK the gnu linker doesn't support this, but Robert H would know far
better than I.  

I'd agree with Eli though that it should really be avoided.  Relying on a
(mis)feature will always get you into trouble in the 
long run.

Andrew

- Raw text -