Re: 1-gary-leopard-dSYM-removal.patch

[Paul Eggert]

"Gary V. Vaughan" <address@hidden> writes:

>> another way I thought of doing it was to find the sites that call:
>>
>>  m4_expand_once([_AC_COMPILER_EXEEXT])
>>  m4_expand_once([_AC_COMPILER_OBJEXT])
>>
>> And replace with a single:
>>
>>  m4_expand_once([_AC_COMPILER_CHECKS])
>>
>> and add to lang.m4:
>>
>>  m4_define([_AC_COMPILER_CHECKS], [
>>  _AC_COMPILER_EXEEXT
>>  _AC_COMPILER_OBJEXT
>>  _AC_COMPILER_CLEAN_DROPPINGS
>>  ])
>>
>> where _AC_COMPILER_CLEAN_DROPPINGS contains the case statement of
>> my original patch.

Something like that sounds reasonable.  I would add the following
further idea.  Get rid of all calls to _AC_COMPILER_EXEEXT and to
_AC_COMPILER_OBJEXT, and replace them all with calls to
_AC_COMPILER_CHECKS.  Then, if you have time, you can optimize the
internals of _AC_COMPILER_CHECKS so that it computes the object file
extension and the executable file extension without creating object
files more than once.  The idea behind this is that almost everybody
wants both OBJ and EXE, so why not cater to the common case?

One more thing: I'd change the name of _AC_COMPILER_CHECKS to
something less-generic.  _AC_COMPILER_CHECK_EXTS, perhaps?