[gst-devel] Shave ... shaving off too much

[Damien Lespiau]

> On Wed, 2009-02-25 at 08:10 +0100, Edward Hervey wrote:
>> Hi all,

Hi Edward,

>>   I really live shave, but can we make sure it doesn't shave off... too
>> much ?

This is exactly what I was trying to warn people about in my successive posts
(see http://damien.lespiau.name/blog/2009/02/24/still-some-hair-left/ )

That said, I'll spend some time and try to have a proper fix in shave/master.
So far, quite a few fixes have find their way into the master branch, but none
that would match your issue I'm afraid.

>>   This is current master as of now.

I tried to reproduce this on both a Debian Lenny machine and an Ubuntu 8.10
one, but no luck, gst-plugins-based HEAD compiled just fine here. That's quite
unfortunate as it will make

To debug those kind of things, you can edit gst-libs/gst/rtsp/Makefile
and remove
the MAKEFLAGS = -s line and run make V=1 again.

Now, I do have more input on this, if you can actually spend a bit of
your time to help
me fix this obviously:

    * the last thing we actually see in the log is libtool trying to
compile a PIC version
of the object. Note that the  '>/dev/null 2>&1' part is added by
libtool itself as he
infers "Oh, we already successfully build the non-PIC version of that
file, we can just
ignore any output when compiling the PIC version".
You could just edit your libtool script, locate and change
	suppress_output=' >/dev/null 2>&1'
or alternatively adding -no-suppress to the LDFLAGS should do the trick.

   * the command that is not echoed (-s!) after this is:
mv -f .deps/libgstrtsp_0.10_la-gstrtspconnection.Tpo
.deps/libgstrtsp_0.10_la-gstrtspconnection.Plo
I think we can not blame this mv

could you please try to get rid of this redirection to /dev/null ?

thanks for your time,

-- 
Damien