X-Authentication-Warning:	delorie.com: mail set sender to djgpp-workers-bounces using -f
X-Recipient:	djgpp-workers AT delorie DOT com
X-Original-DKIM-Signature:	v=1; a=rsa-sha256; c=relaxed/relaxed;
	d=gmail.com; s=20120113;
	h=mime-version:from:date:message-id:subject:to;
	bh=GvDhX+wxusuVfTfdwfEBOWMjzyrmC/zmqnH1nD+2ge4=;
	b=MzEJX0KRLCSq6rSry+Jo9O4MXJhxe3tblqYyrHjVs0XLE6dFnda/wVzdlhh1kwcQ4u
	1T0q3fs7pgODy1UgHCEm8QxIr5cU7FB+9Q3EQ+ypdl47hZ0fklAgVdTYW/mWIA8I2mp8
	Cvddm6JQubeEpZpPxKPTFe1tUU0JL1lN3TKoG4XfvRBPaia/vcIQ4qXwfBiGJKupIT9z
	4woIL6bfP4EdqZ/EqCl3pVWsGWFt52+zVlV7ipBeYs5gS0LJyaVtyqQtCJsgSCVS8pyz
	KvJlyRut1sGxQCotD7Kk1EkABiXAoBcmsDOuZRcOshaoohA2ZAsvBIsfNRnLlUnEaUAF
	3oQA==
X-Google-DKIM-Signature:	v=1; a=rsa-sha256; c=relaxed/relaxed;
	d=1e100.net; s=20130820;
	h=x-gm-message-state:mime-version:from:date:message-id:subject:to;
	bh=GvDhX+wxusuVfTfdwfEBOWMjzyrmC/zmqnH1nD+2ge4=;
	b=ZTLSRVvGKa1TZMN5w3w4RBFgQShZ8sXF0UHnr521qNWzlCbLCWGJJVGv1oKp8cniJg
	zDbq7RhGPKo516k+KaZHNUQiZNoeAc4y1RolBSzf6h9PYLFdRUyveoSoVCPceE32oaoD
	p/UX/fl5v2OA0cwp+PJDPGW4L9h7EKSxeu0SLYKSQGwM4R1lIY4HNyAc4b5KPYiEHaos
	uXoHnjeeJBdRS3TdxlOY4cX69g95/ufXFkMlszJVBo7BmrH7F0gHTCtUAgSAPS10BddJ
	I9ait9KxBU/ghntQSnUZze03IcLayS7OznNc4Lbadl2NMSqWyCCcfeuxByer262vczil
	q5MA==
X-Gm-Message-State:	AEkoouvGOECQJWgm3/y0ATxPPCapBoEQmHV6gOIBhJHTb7dPvblAsdC/3/dFQbaKCz3DDbu9+gEVZcz+a0gC5g==
X-Received:	by 10.31.8.147 with SMTP id 141mr28158826vki.33.1470500633929;
	Sat, 06 Aug 2016 09:23:53 -0700 (PDT)
MIME-Version:	1.0
From:	"Rugxulo (rugxulo AT gmail DOT com) [via djgpp-workers AT delorie DOT com]" <djgpp-workers AT delorie DOT com>
Date:	Sat, 6 Aug 2016 11:23:53 -0500
Message-ID:	<CAA-ihx-exBLSkvQqFz5BHj_Ma1Cqn7bUqYwSPvUMEh7_-ARJEA@mail.gmail.com>
Subject:	linking more bloat from ctime.c
To:	djgpp-workers AT delorie DOT com
Reply-To:	djgpp-workers AT delorie DOT com

Hi, guys,

(Okay, so this isn't a hugely critical issue, nor a priority, so
please don't think I'm complaining. I just wanted to mention it for
completeness.)

We've already discussed the bloat of ctime.c several times over the
years. The most recent discussion was two years ago (2014/04/21), from
Ozkan Sezer, titled "ctime.c changes add about 4.5k more size".

(I found that entire thread chain by searching my email archives and
then further searching on
http://www.delorie.com/djgpp/mail-archives/search.cgi with the
keywords "+crazy +changes +mktime".)

http://www.delorie.com/djgpp/mail-archives/browse.cgi?p=djgpp-workers/2014/04/21/02:44:35

Here's the real problem: *printf (or, more specifically, doprnt.o). A
lot has changed since 2.03p2. One very simple test of mine showed
this:

(2.03p2):
LITE     EXE        44,032
SCANF    EXE        47,616
PRINTF   EXE        51,712
BOTH     EXE        55,296

(2.05):
LITE     EXE        72,192
SCANF    EXE        84,480
PRINTF   EXE        89,088
BOTH     EXE       100,352

Yes, I know, hard drives are cheap. And of course 99% of all C
programs actually need and want to use *printf. But, even
acknowledging that, I think it's a mistake to link in the full
doprnt.o just for a single line in an (here by me) unused routine in
ctime.c by default.

// === badprint.c begins ===
/*
  re: DJGPP 2.05 (DJDEV-/DJLSR205.ZIP) ...

  I only want a simple call to time(), but that needs gettimeofday(),
  and gettimeofday() needs mktime() and localtime(), both of which
  are included inside ctime.o !!

  We must stop ctime.o from pulling in sprintf()'s (bloated) doprnt.o !!

   text    data     bss     dec     hex filename
  10832    2644   14464   27940    6d24 ctime.o
  15712     188       8   15908    3e24 doprnt.o

  sprintf() is called by asctime_r(), which itself is called by
  asctime(), but ctime() also calls asctime() and ctime_r() also
  calls asctime_r() !!

  The single offending sprintf format string is as follows:

  "%.3s %.3s%3d %02d:%02d:%02d %d\n"

  This should be easy to simulate without needing to include the
  full doprnt.o !!
*/

#include <stdio.h>

int sprintf(char* buffer, const char* format, ...)
{
  fputs("Someone is calling me??? I'm a (fake) sprintf() stub!!!",stderr);
  return 0;
}
// === badprint.c ends ===

P.S. In one minimal program of mine, the solution I found recently was
to not use time() at all (for srand() ), instead relying on
(non-standard!) Turbo C's gettime(), which doesn't need all the bloat
(since I was already avoiding printf() there as unnecessary). My point
is that I assume ctime.c can be "fixed" to not need sprintf() at all,
which would be more ideal than either of these kludges, right?

- Raw text -