next porting question for gcj

[Christopher Faylor]

On Mon, Apr 16, 2001 at 12:06:40AM -0400, Christopher Faylor wrote:
>On Mon, Apr 16, 2001 at 12:00:07AM -0400, DJ Delorie wrote:
>>A simple test will show you:
>>
>>main()
>>{
>>  char *b = alloca(500);
>>  return b[3];
>>}
>>
>>Compile with gcc, and you'll see that alloca is implemented directly,
>>without function calls.
>>
>>__builtin_alloca is the internal way gcc implements it; either works
>>just fine but __builtin_alloca is more immune to being #define'd to
>>something else.  It's not specific to Cygwin.
>
>We did have to:
>
>#define alloca __builtin_alloca
>
>at one point in winsup/... at one point, though.

Translation:

We did have to:

#define alloca __builtin_alloca

in several of the sources in the winsup directory.

I'm not sure what I was trying to convey by the double "at one point" above.

cgf