Compile problems on OSX

Hi all,

I've got two problems compiling the current CVS FLAC sources
on OSX.

Firstly, the configure script can't find the OGG libraries
which were
installed from MacPorts. I have tried:

    ./configure --with-ogg-includes=/opt/local/include 
           --with-ogg-lib=/opt/local/lib

but they are still not found. I notice that in configure.in,
you use
a macro called XIPH_PATH_OGG. Is there any reason you don't
use the
PKG_CHECK_MODULES mcaro which uses pkg-config underneath?
For my 
system pkg-config does the right thing:

    > pkg-config --cflags --libs ogg
    -I/opt/local/include  -L/opt/local/lib -logg  

I will put together a patch for this is you are interested.

Second problem is that I get a compile error when
FLAC__HAS_OGG is 0
in src/flac/decode.c:

    decode.c:393: error: 'options' undeclared (first use in
this function)

The code in question is this:

    #if FLAC__HAS_OGG
        if(decoder_session->is_ogg) {
            /* Some code here. */
        }
        else
    #else
        (void)decode_options;
    #endif
        {
            /* More code. */
        }

The problem here is that conditional compiles can hide bugs.
The code above 
re-written to turn a conditional compile into an always
compiled if statement
which will trigger the error regardless of the value of
FLAC__HAS_OGG:

        if(FLAC__HAS_OGG &&
decoder_session->is_ogg) {
            /* Some code here. */
        }
        else
        {
            (void)decode_options;
            /* More code. */
        }

Cheers,
Erik
-- 
------------------------------------------------------------
-----
Erik de Castro Lopo
------------------------------------------------------------
-----
'Unix beats Windows' - says Microsoft!
http://
blogs.zdnet.com/Murphy/index.php?p=459
_______________________________________________
Flac-dev mailing list
Flac-devxiph.org
http:
//lists.xiph.org/mailman/listinfo/flac-dev

--- Erik de Castro Lopo <erikd-flacmega-nerd.com> wrote:
> Hi all,
> 
> I've got two problems compiling the current CVS FLAC
sources on OSX.
> 
> Firstly, the configure script can't find the OGG
libraries which were
> installed from MacPorts. I have tried:
> 
>     ./configure --with-ogg-includes=/opt/local/include

>            --with-ogg-lib=/opt/local/lib
> 
> but they are still not found. I notice that in
configure.in, you use
> a macro called XIPH_PATH_OGG. Is there any reason you
don't use the
> PKG_CHECK_MODULES mcaro which uses pkg-config
underneath? For my 
> system pkg-config does the right thing:
> 
>     > pkg-config --cflags --libs ogg
>     -I/opt/local/include  -L/opt/local/lib -logg  
> 
> I will put together a patch for this is you are
interested.

the main reason I haven't swtiched is I'm not up on
pkgconfig
and XIPTH_PATH_OGG has been working pretty well... can you
tell
why it's not working for you?

I'm always hesitant to mess with things in configure.in
because
is such whack-a-mole problem (fixes things for some people
and
breaks for others).

> Second problem is that I get a compile error when
FLAC__HAS_OGG is 0
> in src/flac/decode.c:
> 
>     decode.c:393: error: 'options' undeclared (first
use in this
> function)
> 
> The code in question is this:
> 
>     #if FLAC__HAS_OGG
>         if(decoder_session->is_ogg) {
>             /* Some code here. */
>         }
>         else
>     #else
>         (void)decode_options;
>     #endif
>         {
>             /* More code. */
>         }

yes, this is fixed in cvs.  actually I'll probably be doing
a
release soon and I'd be interested to know if the latest
stuff
works for you.

> The problem here is that conditional compiles can hide
bugs. The code
> above 
> re-written to turn a conditional compile into an always
compiled if
> statement
> which will trigger the error regardless of the value
of
> FLAC__HAS_OGG:
> 
>         if(FLAC__HAS_OGG &&
decoder_session->is_ogg) {
>             /* Some code here. */
>         }
>         else
>         {
>             (void)decode_options;
>             /* More code. */
>         }

interesting idea, hadn't thought of that.



     
____________________________________________________________
________________________
Luggage? GPS? Comic books? 
Check out fitting gifts for grads at Yahoo! Search
http://search.yahoo.com/search?fr=
oni_on_mail&p=graduation+gifts&cs=bz
_______________________________________________
Flac-dev mailing list
Flac-devxiph.org
http:
//lists.xiph.org/mailman/listinfo/flac-dev

Josh Coalson wrote:

> the main reason I haven't swtiched is I'm not up on
pkgconfig
> and XIPTH_PATH_OGG has been working pretty well... can
you tell
> why it's not working for you?
> 
> I'm always hesitant to mess with things in configure.in
because
> is such whack-a-mole problem (fixes things for some
people and
> breaks for others).

I'll look into that over the next day or so. The pkg-config
method
has the advantage of being highly robust, but is a pain for
MinGW
because you need pkg-config.

> > Second problem is that I get a compile error when
FLAC__HAS_OGG is 0
> > in src/flac/decode.c:
> > 
> >     decode.c:393: error: 'options' undeclared
(first use in this
> > function)
> > 
> > The code in question is this:
> > 
> >     #if FLAC__HAS_OGG
> >         if(decoder_session->is_ogg) {
> >             /* Some code here. */
> >         }
> >         else
> >     #else
> >         (void)decode_options;
> >     #endif
> >         {
> >             /* More code. */
> >         }
> 
> yes, this is fixed in cvs.  actually I'll probably be
doing a
> release soon and I'd be interested to know if the
latest stuff
> works for you.

Sorry, I just updated from CVS and its still broken. If
FLAC__HAS_OGG
is false it tries to compile this:

     (void)decode_options;

and there is no "decode_options" identifier.

Maybe this is the delay between your CVS commit and the
public CVS 
being updated.

> interesting idea, hadn't thought of that.

Please take a look at this. Doing it this way means that
both code
paths are checked by the semantic analysis stage of the
compiler
regardless of value of FLAC__HAS_OGG. It prevents
conditional compile
constructs from hiding really blatantly obvious errors like
this one.

Erik
-- 
------------------------------------------------------------
-----
Erik de Castro Lopo
------------------------------------------------------------
-----
"Whenever the C++ language designers had two competing
ideas as to
how they should solve some problem, they said, "OK,
we'll do them
both". So the language is too baroque for my
taste." -- Donald E Knuth
_______________________________________________
Flac-dev mailing list
Flac-devxiph.org
http:
//lists.xiph.org/mailman/listinfo/flac-dev

Erik de Castro Lopo wrote:

> I'll look into that over the next day or so. The
pkg-config method
> has the advantage of being highly robust, but is a pain
for MinGW
> because you need pkg-config.

I just tested this:

    ./configure --with-ogg=/opt/local

and so does:

    ./configure --with-ogg-libraries=/opt/local/lib 
           --with-ogg-includes=/opt/local/include

Not sure why it failed last time.

Erik
-- 
------------------------------------------------------------
-----
Erik de Castro Lopo
------------------------------------------------------------
-----
"I run Linux on pretty much everything except the
microwave and
washing machine. Those are tempting targets but would
probably
make Telsa extremely cross." -- Alan Cox
_______________________________________________
Flac-dev mailing list
Flac-devxiph.org
http:
//lists.xiph.org/mailman/listinfo/flac-dev