Re: PING: patch: PR-4877 testsuite fix for secreloc32 x86_64

[Christopher Faylor <cgf-use-the-mailinglist-please@sourceware.org>]

On Mon, Aug 06, 2007 at 01:57:37PM +0200, Kai Tietz wrote:
>Hi Christopher,
>
>> On Mon, Aug 06, 2007 at 09:34:55AM +0200, Kai Tietz wrote:
>> >Hi,
>> >
>> >this patch will fix the problem on secreloc32 tests for x86_64-mingw32. 
>
>> >For this target the ctor/dtor elements are 8 bytes long. This caused 
>the 
>> >FAIL.
>> >
>> >See bug-report: http://sourceware.org/bugzilla/show_bug.cgi?id=4877
>> >
>> >I tested this patch for x86_64-pc-mingw32 test-suite without any 
>> >regressions.
>> >
>> >ChangeLog for ld/testsuite:
>> >
>> >2007-08-02  Kai Tietz   <kai.tietz@onevision.com>
>> >
>> >        * ld-pe/pe.exp: special diff file for x86_64-mingw target.
>> >        * ld-pe/secrel_64.d: New.
>> 
>> Is there some reason why this wouldn't work under a potential
>> x86_64-cygwin target?  I haven't been paying enough attention to this
>> issue but it seems like whatever machinery works for mingw should also
>> work for cygwin eventually.
>
>If there would be a crt1 for cygwin, I am certain that it would work
>for x86_64-cygwin too.

Cygwin does have similar crt* mechanisms to every other platform out
there.  I'm not asking you to implement anything here but could you
maybe just keep cygwin in mind when making patches like this and not
just default to "mingw"?

cgf