Re: bison-20001221

[Hans Aberg]

At 12:24 +0100 1-01-18, Akim Demaille wrote:
>Hans> Another idea I had is to make obstack into a FILE
>
>May I recall that we moved to obstack to avoid temp files?

May I recall that FILE is a C data struct and not a file? :-)

-- The idea is of course to make something like obstack working on memory,
but wrapped up to look like a FILE struct. It would have been easy in C++,
in C I do not know.

>Hans> - I got around the problem by changing the obstack_fgrow1 macro
>Hans> containing a static array: static char buf[4096]; Not pure
>Hans> anymore, but it does not make any difference in my port.
>
>Bleah.  What size would make you happy?  I've used 4096 knowing it's
>way too much.

I don't care; I just want a quick fix that made the port compile & work.
You said youreself my compiler was dumb, which is the root of the problem.

>Hans> Unnecessary if "quotearg.h" includes <size_t.h>.
>>> ???  No it does not.
>
>This is the famour `size_t.h' I was referring to in another mail, and
>you told me there is no such thing.  So what is the problem then?

The header <size_t.h> is included from <stddef.h> on my computer, and the
latter is ISO/ANSI C. One can have several different headers defining the
same object, but each compiler should make sure that no problems if several
headers defining the same thing are included.

So the tendency is that programmers flip in a header that seems to work on
their particular platform, be it <std/...> or <size_t.h> or whatever,
instead of those standard C headers defining the same thing.

This is what I think is going on.

>> Somebody at GNU seems to put in extra headers
>> just in case. :-)
>
>My point, again, is that these headers are /extra/ to *you*.  Not
>everybody.

In some cases I got the impression that these are headers that once were
needed, but as the package was slimmed down, the old includes remained
despite not needed.

  Hans Aberg