This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [Patch debug] Do not emit debug sections until they are required.
- From: IainS <developer at sandoe-acoustics dot co dot uk>
- To: Tom Tromey <tromey at redhat dot com>
- Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>, Richard Henderson <rth at redhat dot com>, Mike Stump <mrs at gcc dot gnu dot org>
- Date: Wed, 3 Nov 2010 13:55:09 +0000
- Subject: Re: [Patch debug] Do not emit debug sections until they are required.
- References: <A3086084-F186-42E0-92EB-1359151FF2D7@sandoe-acoustics.co.uk> <m3oca75t4h.fsf@fleche.redhat.com>
On 2 Nov 2010, at 21:10, Tom Tromey wrote:
"IainS" == IainS <developer@sandoe-acoustics.co.uk> writes:
IainS> In order to do this, macinfo data are saved in a vec, and
emitted from
IainS> dwarf2out_finish () rather than 'on the fly'.
Does this impact gcc's memory use?
the size of the macro strings + a small overhead per macro.
(i.e. it should be dominated by the strings in most cases).
one would generally expect this to number in kb rather than Mb -
although I'm sure we could come up with a scenario which would exceed
this.
(And does it matter to anybody?)
... given the discussion on irc last night about >4Gb objects, it
seems unlikely.
It's judgement on whether the improvement in determinism for the
section order and readability of the asm are worth the increase in
footprint.
cheers
Iain