This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH, committed] Add ability to set the optimization options and on ix86 target options on a function specific basis
- From: "Manuel López-Ibáñez" <lopezibanez at gmail dot com>
- To: "Michael Meissner" <gnu at the-meissners dot org>, gcc-patches at gcc dot gnu dot org, karthikkumar at gmail dot com
- Date: Wed, 23 Jul 2008 13:32:43 +0200
- Subject: Re: [PATCH, committed] Add ability to set the optimization options and on ix86 target options on a function specific basis
- References: <20080723103750.GA9075@tiktok.the-meissners.org>
2008/7/23 Michael Meissner <gnu@the-meissners.org>:
> I committed the following patch to revision 138075 on the mainline to add
> function specific optimization/target option support. I did not move the
> optimization and target options data structures from being trees containing the
> structure to the structure itself that Steven Bosscher suggested, because it
> got complicated with regard to garbage collection and hashing. I might look at
> it later, but I wanted to get the main patches checked in.
I am sorry I comment on this so late but I just found out about the
current syntax. If my comments have already been answered, feel free
to point me to some thread where I can read the rationale for the
design decisions.
* Why don't make a 1-to-1 mapping between the command-line option and
the argument to option("") ?
It seems pretty cumbersome to have to strip the -m for each option.
Moreover, if we extend the current code to handle -Ox -f* -W* and
whatever we wish to handle (--param?), things get worse.
* In relation to the previous. Why separate options with commas
instead of spaces? In other words, why not make the argument of
option("") behave as the command-line. That is, the commandline "gcc
-march=i686 -mfpmath=sse,387" has a 1-to-1 mapping to
option("-march=i686 -mfpmath=sse,387"). Also, you don't need to work
around commas.
Cheers,
Manuel