GNU bug report logs - #23764
[bug-gnu-emacs] emacs-25.0.95 fails to build on DragonFlyBSD, and an easy repair

Previous Next

Package: emacs;

Reported by: "Nelson H. F. Beebe" <beebe <at> math.utah.edu>

Date: Mon, 13 Jun 2016 20:11:01 UTC

Severity: normal

Fixed in version 25.0.96

Done: Glenn Morris <rgm <at> gnu.org>

Bug is archived. No further changes may be made.

Full log


Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: "Nelson H. F. Beebe" <beebe <at> math.utah.edu>
To: bug-gnu-emacs <at> gnu.org
Cc: beebe <at> math.utah.edu
Subject: [bug-gnu-emacs] emacs-25.0.95 fails to build on DragonFlyBSD, and an
 easy repair
Date: Mon, 13 Jun 2016 14:10:07 -0600
emacs-25.0.95 fails to build on DragonFlyBSD 3.8, 3.9, 4.0, 4.2, 4.4,
and 4.6, independent of the compiler (various gcc or clang versions)
used.

On each system, src/alloc.c fails to compile:

	CC       alloc.o
	alloc.c:474:32: error: macro "ALIGN" passed 2 arguments, but takes just 1
	 ALIGN (void *ptr, int alignment)
					^
	alloc.c:475:1: error: expected '=', ',', ';', 'asm' or '__attribute__' before '' token
	 {
	 ^
	alloc.c: In function 'pure_alloc':
	alloc.c:5171:66: error: macro "ALIGN" passed 2 arguments, but takes just 1
	       result = ALIGN (purebeg + pure_bytes_used_lisp, GCALIGNMENT);
									  ^
	alloc.c:5171:16: error: 'ALIGN' undeclared (first use in this function)
	       result = ALIGN (purebeg + pure_bytes_used_lisp, GCALIGNMENT);
			^
	alloc.c:5171:16: note: each undeclared identifier is reported only once for each function it appears in
	Makefile:378: recipe for target 'alloc.o' failed
	gmake[1]: [alloc.o] Error 1 (ignored)

The conflict with the all-to-generic macro name ALIGN() appears to
come from these DragonFlyBSD definitions:

	% find /usr/include -type f | xargs grep '[^_A-Z]_ALIGN[^_A-Z]'
	/usr/include/cpu/param.h:#define _ALIGN(p)      (((unsigned long)(p) + _ALIGNBYTES) & ~_ALIGNBYTES)
	/usr/include/cpu/param.h:#define ALIGN(p)       _ALIGN(p)

whereas alloc.c has

	static void *
	ALIGN (void *ptr, int alignment)
	{
	  return (void *) ROUNDUP ((uintptr_t) ptr, alignment);
	}

As an experiment, I made these patches to alloc.c on DragonFlyBSD 4.6:

	% diff alloc.c.org alloc.c
	474c474
	< ALIGN (void *ptr, int alignment)
	---
	> EMACS_ALIGN (void *ptr, int alignment)
	1245c1245
	<       abase = ALIGN (base, BLOCK_ALIGN);
	---
	>       abase = EMACS_ALIGN (base, BLOCK_ALIGN);
	5171c5171
	<       result = ALIGN (purebeg + pure_bytes_used_lisp, GCALIGNMENT);
	---
	>       result = EMACS_ALIGN (purebeg + pure_bytes_used_lisp, GCALIGNMENT);

That solved the build problem on DragonFlyBSD 4.6, and I've installed
the new emacs on that system.  The same fix repaired all of the five
other versions of that O/S in my lab.

Another, and simpler solution, would be to insert

	#undef ALIGN

before the first use of ALIGN() in alloc.c.

-------------------------------------------------------------------------------
- Nelson H. F. Beebe                    Tel: +1 801 581 5254                  -
- University of Utah                    FAX: +1 801 581 4148                  -
- Department of Mathematics, 110 LCB    Internet e-mail: beebe <at> math.utah.edu  -
- 155 S 1400 E RM 233                       beebe <at> acm.org  beebe <at> computer.org -
- Salt Lake City, UT 84112-0090, USA    URL: http://www.math.utah.edu/~beebe/ -
-------------------------------------------------------------------------------




This bug report was last modified 8 years and 344 days ago.

Previous Next


GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.