Re: [Tracker] Tracker on Solaris

[Martyn Russell <martyn lanedo com>]

On 25/01/13 11:12, Frank Lahm wrote:
> Hi Martyn,

Hello Frank,

> 2013/1/24 Martyn Russell <martyn lanedo com>:
> > On 10/01/13 17:31, Frank Lahm wrote:
> > Looking at these patches, I wonder why you use 0002 to change
> > uuid_unparse_lower() to uuid_unparse()?
>
> Solaris libuuid is missing uuid_unparse_lower, it was a simple hack^wfix. :)

Hmm, this worries me. We use the _lower() variant for a reason here and 
it would obviously be faster to not use that. It might be a short term 
fix for you, but I wonder if this causes regressions for you at all?

I think a patch like this is fine if we can do it only on platforms that 
don't have that API. JÃrg, did Vala get pre-processor directives yet? 
i.e. #ifdef FOO, etc.?

> > Regarding the __EXTENSIONS__ part, what does that actually allow?
>
> <http://docs.oracle.com/cd/E19253-01/816-5175/6mbba7f3l/index.html>
>
> "Feature test macros are used by applications to indicate additional
> sets of features that are desired beyond those specified by the C
> standard. If an application uses only those interfaces and headers
> defined by a particular standard (such as POSIX or X/Open CAE), then
> it need only define the appropriate feature test macro specified by
> that standard. If the application is using interfaces and headers not
> defined by that standard, then in addition to defining the appropriate
> standard feature test macro, it must also define __EXTENSIONS__.
> Defining __EXTENSIONS__ provides the application with access to all
> interfaces and headers not in conflict with the specified standard.
> The application must define __EXTENSIONS__ either on the compile
> command line or within the application source files."

As expected then.

> > On Linux,
> > it seems only useful for Python and MySQL. A quick Google search shows this
> > is needed perhaps for u_int and/or sys/types.h on Solaris - is this the
> > case? In any case, it seems to have little effect on Linux.
>
> *BSD are also affected by this, requesting a UNIX compilation
> environment will hide everything else not in the standard. On Linux
> it's the other way around, you don't get certain functions (eg afair
> pread()/pwrite()) without going into SUS. Blah! To solve this in
> Netatalk we use this a header file which uses the results of platform
> detection to define the right thing [tm]:
>
> #ifdef HAVE_CONFIG_H
> #include "config.h"
> #endif
>
> #if !defined(__FreeBSD__) && !defined(__NetBSD__) && !defined(__OpenBSD__)
> # ifndef _XOPEN_SOURCE
> #  define _XOPEN_SOURCE 600
> # endif
> # ifndef __EXTENSIONS__
> #  define __EXTENSIONS__
> # endif
> # ifndef _GNU_SOURCE
> #  define _GNU_SOURCE
> # endif
> #endif

We can add this to our src/libtracker-common/tracker-os-dependant.[ch] 
files, but you add some extra checks in configure.{ac|in} right?

If you're willing to come up with a patch here, we would be happy to 
accept that. OR we could simply just add the above and allow you to 
define things in your build / set up?

I think the patch you did before is a bit more of a hack otherwise.

I am pretty sure we've defined GNU_SOURCE in a few places too, we could 
clear all that up with it done in one place in libtracker-common.

I wonder how many of those cases are necessary too and how many are 
copy/paste entries :)

> > For the CPP linking issue, that will have to be a local patch for you I
> > think, unless there is a way to fix this otherwise. We do have a
> > --disable-meego-touch option, but I am guessing that doesn't help?
>
> Missed that one, but no, doesn't help.

Sadly, this one will have to stay with you guys, it's not useful for the 
project to include this patch I feel.

It would be good to file a bug against automake? about this though.

--
Regards,
Martyn

Founder and CEO of Lanedo GmbH.