This is the mail archive of the cygwin-patches mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH 64bit] Fix speclib for x86_64


On Feb 17 12:18, Christopher Faylor wrote:
> On Sun, Feb 17, 2013 at 06:09:44PM +0100, Corinna Vinschen wrote:
> >On Feb 17 11:52, Christopher Faylor wrote:
> >> On Sun, Feb 17, 2013 at 02:41:41PM +0100, Corinna Vinschen wrote:
> >> >On Feb 17 04:46, Yaakov wrote:
> >> >
> >> >> 2013-02-16  Yaakov Selkowitz  <yselkowitz@...>
> >> >> 
> >> >> 	* Makefile.in (libcygwin.a): Move --target flag from here...
> >> >> 	(toolopts): to here, to be used by both mkimport and speclib.
> >> >> 	* speclib: Omit leading underscore in symbol names on x86_64.
> >> >
> >> >The Makefile patch is fine, but for the speclib change I wonder why
> >> >we should omit the leading underscore.  If you remove the underscore,
> >> >you're polluting the application namespace.  Is there really a good
> >> >reason to do that?  Did I miss something?
> >> 
> >> Doesn't the x86_64 target forego leading underscores on normal variable
> >> names?
> >
> >My dictionary returns ambiguous results for the word "forego",
> >so I answer that generically:
> 
> I mispelled "forgo".
> 
> >On x86_64 there's no underscore prepended to symbols.
> 
> So, why would we add an underscore?  Does binutils add underscores to
> these *_dll_iname symbols anyway?

That's what I subsumed under "Did I miss something?"  The answer is
apparently "yes".  AFAICS, binutils does not prepend an underscore to
these iname symbols, so the patch to speclib is fine.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Maintainer                 cygwin AT cygwin DOT com
Red Hat


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]