timejson.c fails to compile on FreeBSD

[Jim Long]

timejson.c includes <malloc.h>. FreeBSD declares the prototypes for malloc(), et al in stdlib.h.

Patch attached.

Without the patch, the following behaviour is observed:

# uname -srp
FreeBSD 8.3-STABLE amd64
# pwd
/home/opensrf/OpenSRF
# autoreconf -i
...
# CFLAGS="-I/usr/local/include" LDFLAGS="-L/usr/local/lib" ./configure --prefix=/openils \
--sysconfdir=/openils/conf --with-apxs=/usr/local/sbin/apxs --with-libxml=/usr/local/include/libxml2 \
--with-includes=/usr/local/include
...
# gmake
...
In file included from timejson.c:8:
/usr/include/malloc.h:3:2: error: #error "<malloc.h> has been replaced by <stdlib.h>"
gmake[2]: *** [timejson.o] Error 1
gmake[2]: Leaving directory `/usr/home/opensrf/OpenSRF/src/c-apps'
gmake[1]: *** [all-recursive] Error 1
gmake[1]: Leaving directory `/usr/home/opensrf/OpenSRF/src'
gmake: *** [all-recursive] Error 1
# cat -n src/c-apps/timejson.c | head
     1 #include <stdlib.h>
     2 #include <stdio.h>
     3 #include <string.h>
     4 #include <stdarg.h>
     5 #include <time.h>
     6 #include <sys/time.h>
     7 #include <sys/resource.h>
     8 #include <malloc.h>
     9 #include "opensrf/utils.h"
    10 #include "opensrf/osrf_json.h"
#

Developer's Certificate of Origin 1.1

By making a contribution to this project, I certify that:

(a) The contribution was created in whole or in part by me and I
    have the right to submit it under the open source license
    indicated in the file; or

(b) The contribution is based upon previous work that, to the best
    of my knowledge, is covered under an appropriate open source
    license and I have the right under that license to submit that
    work with modifications, whether created in whole or in part
    by me, under the same open source license (unless I am
    permitted to submit under a different license), as indicated
    in the file; or

(c) The contribution was provided directly to me by some other
    person who certified (a), (b) or (c) and I have not modified
    it.

(d) I understand and agree that this project and the contribution
    are public and that a record of the contribution (including all
    personal information I submit with it, including my sign-off) is
    maintained indefinitely and may be redistributed consistent with
    this project or the open source license(s) involved.

Signed-off-by:

Jim Long
<email address hidden>

[Jason Stephenson (jstephenson)]

Jim,

I liked the patch that you sent me privately better, the one where you checked if __STDC__ was defined. Could you add that one to this bug as well?

Thanks,
Jason

[Jim Long (evergreen-9)]

That was not a patch, I was showing you the contents of FreeBSD's /usr/include/malloc.h.

malloc.h looks at whether __STDC__ is defined. If it is, it throws an error. If __STDC__ is not
defined, then malloc.h includes stdlib.h.

In either case, timejson.c already includes stdlib.h at line 1 before including malloc.h at line 8, so ....

/usr/include/malloc.h attached.

Holler again if I've misunderstood your request.

Jim

[Bill Erickson (berick)]

Note the top commit in http://git.evergreen-ils.org/?p=working/OpenSRF.git;a=shortlog;h=refs/heads/collab/berick/bsd-compat

[Jason Stephenson (jstephenson)]

Bill, thanks for sharing. I think Jim has missed applying the above branch to OpenSRF. It is my fault since I neglected to mention it and even suggested that OpenSRF builds without patches, which was likely me misremembering.

I actually think that the inclusion of malloc.h can be removed from Linux, too. I've never needed to include it, unless I'm doing something implementation-specific that one shouldn't do in the first place.

Just my two bits.

[Jason Stephenson (jstephenson)]

Also, I am inclined to abandon the *bsd-compat branches at the moment in favor of a solution to https://bugs.launchpad.net/evergreen/+bug/968531

Updating our use of autotools would go a long way to smoothing the addition of new platforms.

[Bill Erickson (berick)]

On Fri, Oct 19, 2012 at 9:17 AM, Jason Stephenson <
<email address hidden>> wrote:

>
> I actually think that the inclusion of malloc.h can be removed from
> Linux, too. I've never needed to include it, unless I'm doing something
> implementation-specific that one shouldn't do in the first place.
>

Agreed. I guess I though the malloc_stats() stuff needed it, but that
doesn't appear to be the case.

-b

--
Bill Erickson
| Senior Software Developer
| phone: 877-OPEN-ILS (673-6457)
| email: <email address hidden>
| web: http://esilibrary.com
| Equinox Software, Inc. / Your Library's Guide to Open Source