This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: slow ar on cygwin windows patch
- To: Mike Stump <mrs at windriver dot com>
- Subject: Re: slow ar on cygwin windows patch
- From: Jonathan Larmour <jlarmour at redhat dot com>
- Date: Fri, 26 Jan 2001 05:34:15 +0000
- Cc: binutils at sources dot redhat dot com
- Organization: Red Hat UK Ltd.
- References: <200101252354.PAA01631@kankakee.wrs.com>
Mike Stump wrote:
>
> I've found that I need the below patch on Windows to prevent it from
> being 233 times slower when cygwin (B19) is hosting when an achieve is
> being created with 14 .o files. Without this patch, it was taking
> over a minute, with the patch, 0.3 seconds, a worthwhile win.
Cygwin B19 is ancient. A lot has changed in Cygwin since then. I don't
think this should be applied unless Cygwin 1.1 is equally slow. There is no
reason to penalise current cygwin because of an ancient beta version. Given
the time measurements you give, this speed is certainly not what I have
observed unless there is something more exotic with your situation.
> + /* This speed up creating an archive with ar with 14 .o
> + files 233x on Windows, this functionality isn't worth the
> + slowness. */
> + #if ! defined (_WIN32)
If this does get applied, this should be __CYGWIN__. In fact at the time of
B19 you would have needed __CYGWIN32__ (which may also still be defined for
compatibility now, I'm not sure).
Jifl
--
Red Hat, Rustat House, Clifton Road, Cambridge, UK. Tel: +44 (1223) 271062
Un cheval, pas du glue. Pas du cheval, beaucoup du glue. || Opinions==mine