Message-ID:	<B0000030603@stargate.astr.lu.lv>
From:	"Andris Pavenis" <pavenis AT lanet DOT lv>
To:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
Date:	Thu, 21 May 1998 17:49:27 +0300
MIME-Version:	1.0
Subject:	Re: gcc 2.8.1 outputs are much larger than gcc 2.7.2 ... why ?? (fwd)
CC:	djgpp-workers AT delorie DOT com
References:	<B0000030590 AT stargate DOT astr DOT lu DOT lv>
In-reply-to:	<Pine.SUN.3.91.980521165059.1049C-100000@is>

Date sent:      	Thu, 21 May 1998 16:54:40 +0300 (IDT)
From:           	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To:             	Andris Pavenis <pavenis AT lanet DOT lv>
Copies to:      	djgpp-workers AT delorie DOT com
Subject:        	Re: gcc 2.8.1 outputs are much larger than gcc 2.7.2 ... why ?? (fwd)

> The message I forwarded claimed that exceptions support adds a lot of 
> local labels, and that's why it bloats the binary.
> 
> > There is one more problem. I did not suceed to get rid of local labels in
> > libstdcxx.a  and  libgpp.a  (I compiled without debugging information).
> 
> Let me see if I understand.  Are you saying that you *can* get rid of 
> local labels in object files, but not in libstdcxx?  How's that possible?
> 
I was not very correct. There was no big changes in object file
sizes for C programs (or also for C++ programs compiled with -fno-exceptions)
but object files from both compiler versions contained references to
local labels. So they of course remained also in libraries.

There is not very much local labels in
a single C++ source file so the increase in size is not so well visible. 
But there is MANY C++ sources in libstdcxx.a so the increase in executable
size is visible much better.

- Raw text -