[Vm-dev] Compiling Interpreter VM on FreeBSD

Frank Shearar frank.shearar at gmail.com
Tue Jan 8 08:30:44 UTC 2013


On 7 January 2013 22:41, Eliot Miranda <eliot.miranda at gmail.com> wrote:
>
> Surely we'd want to write it thus:?
>
> #if defined(HAVE_SYS_UUID_H)
> # include <sys/types.h>
> # include <sys/uuid.h>
> #elif defined(HAVE_UUID_UUID_H)
> # include <uuid/uuid.h>
> #elif defined(HAVE_UUID_H)
> # include <uuid.h>
> #else
> # error cannot find a uuid.h to include
> #endif

Er yes, what you said :) I touch C code maybe once a year, hence my
incompetence!

frank

> On Mon, Jan 7, 2013 at 4:31 AM, Frank Shearar <frank.shearar at gmail.com> wrote:
>>
>>
>> On 3 January 2013 00:35, David T. Lewis <lewis at mail.msen.com> wrote:
>> >
>> > On Wed, Jan 02, 2013 at 07:22:42PM +0000, Frank Shearar wrote:
>> >>
>> >> I'm setting up more CI jobs, this time to give FreeBSD some love. So I
>> >> have a script that gets the bleeding edge Interpreter source, cd
>> >> platform/unix; make. The results are here:
>> >> http://squeakci.org/job/InterpreterVM-FreeBSD/13/console
>> >>
>> >> The punchline is that compilation fails thusly:
>> >>
>> >> In file included from
>> >> /usr/home/frank/temp/jenkins/workspace/InterpreterVM-FreeBSD/target/Squeak-vm-unix-4.10.7-2646-unofficial-src/platforms/unix/plugins/UUIDPlugin/sqUnixUUID.c:9:
>> >> /usr/local/include/uuid/uuid.h:44: error: conflicting types for 'uuid_t'
>> >> /usr/include/sys/uuid.h:74: error: previous declaration of 'uuid_t' was here
>> >
>> > Frank,
>> >
>> > The error message above looks mighty suspicious to me. Where did that
>> > /usr/local/include/uuid/uuid.h come from? I have a hunch that you may
>> > have a locally installed libuuid package that is providing a header
>> > file that conflicts with a /usr/include/sys/uuid.h that came with your
>> > base FreeBSD distribution.
>>
>> If I adjust sqUnixUUID.c in this fashion:
>>
>>
>> #if defined(HAVE_SYS_UUID_H)
>> # include <sys/types.h>
>> # include <sys/uuid.h>
>> #endif
>>
>> + #if !(defined(HAVE_SYS_UUID_H)) && defined(HAVE_UUID_UUID_H)
>> - #if defined(HAVE_UUID_UUID_H)
>> # include <uuid/uuid.h>
>> #elif defined(HAVE_UUID_H)
>> # include <uuid.h>
>> #endif
>>
>> I can happily compile the UUID plugin! I think the issue arises for me
>> because config finds BOTH sys/uuid.h AND uuid/uuid.h.
>>
>> frank
>>
>> > I'm pretty sure that libuuid is entitled to some sort of Guiness world record
>> > for greatest number of stupid preventable bugs for a single small system
>> > library. The proliferation of include/uuid/uuid.h versus include/sys/uuid.h
>> > mismatches is a big part of it, but read our very own Mantis tale of woe for
>> > further entertainment:
>> >
>> > http://bugs.squeak.org/view.php?id=7358
>> >
>> > Dave
>> >
>> >
>> >
>> >
>> >> /usr/home/frank/temp/jenkins/workspace/InterpreterVM-FreeBSD/target/Squeak-vm-unix-4.10.7-2646-unofficial-src/platforms/unix/plugins/UUIDPlugin/sqUnixUUID.c:
>> >> In function 'MakeUUID':
>> >> /usr/home/frank/temp/jenkins/workspace/InterpreterVM-FreeBSD/target/Squeak-vm-unix-4.10.7-2646-unofficial-src/platforms/unix/plugins/UUIDPlugin/sqUnixUUID.c:29:
>> >> warning: passing argument 1 of 'uuidgen' from incompatible pointer
>> >>
>> >> The relevant source is here:
>> >>
>> >>
>> >> int MakeUUID(char *location)
>> >> {
>> >> #if defined(HAVE_UUID_CREATE) && !defined(HAVE_UUIDGEN) &&
>> >> !defined(HAVE_UUID_GENERATE)
>> >>   size_t  len= 16;      /* 128 bits */
>> >>   uuid_t *uuid;
>> >>   uuid_create(&uuid);
>> >>   uuid_make(uuid, UUID_MAKE_V1);
>> >>   uuid_export(uuid, UUID_FMT_BIN, (void **)&location, &len);
>> >>   uuid_destroy(uuid);
>> >> #else
>> >>   uuid_t uuid;
>> >> #  if defined(HAVE_UUIDGEN)
>> >>   uuidgen(&uuid, 1);   // <-- this is line 29, the offender
>> >> #  elif defined(HAVE_UUID_GENERATE)
>> >>   uuid_generate(uuid);
>> >> #  endif
>> >>   memcpy((void *)location, (void *)&uuid, sizeof(uuid));
>> >> #endif
>> >>   return 1;
>> >> }
>> >>
>> >> Any ideas? I'm happy to do legwork, but Bozhe moi! This I know from nothing!
>> >>
>> >> frank
>
>
>
>
> --
> best,
> Eliot
>


More information about the Vm-dev mailing list