beta7 experience

Hi all!

I recently installed beta7 on my box and I experienced a few problems:

snd_emu10k1
-----------
To build the snd_emu10k1 sound driver, one needs to add:

	device "snd_emu10k1"

to the kernel config file. Simply:

	device snd_emu10k1

will cause config to complain about a syntax error (presumably because
it thinks the 1 at the end is a device number).

This is not mentioned anywhere in the snd_emu10k1 man page and the
example provided shows 'device snd_emu10k1', which won't work.

At the very least, the man page needs to be updated to reflect this.
(I could provide a diff for this but I guess the overhead of applying a
patch will be greater than just making the change directly in this
case.)

In the future, perhaps the driver should be renamed or the device
numbering scheme redesigned in some way? It feels a bit dirty having to
put "" around the driver name :)

Somebody else posted a message complaining about this a couple of days
ago but didn't recieve much attention... I would guess that everything
I just wrote holds true for snd_ad1816, snd_als4000, snd_cs4281, snd_ds1
and snd_maestro3 too.

cvsup example files
-------------------
minor thing: the stable-supfile example found in
/usr/share/examples/cvsup still lists 4.x as stable. Perhaps this should
be updated before the release is made? :)

X.org
-----
When running xorgconfig, I am not allowed to choose my keyboard mapping,
because of the following error:

XKB rules file '/usr/X11R6/lib/X11/xkb/rules/' not found
Keyboard XKB options will be set to default values.

I am not sure if anybody's reported this before me, but at least I
haven't seen any messages about it.

Also, it seems Option ZAxisMapping "4 5" isn't standard... Perhaps
making it standard should be considered, or is there any compelling
reason not to? I guess most people nowadays have a wheel mouse.

Weird error when building gtk2 from ports
-----------------------------------------
When building firefox (and gtk2 as a dependency), I recieved a weird
error:

/usr/bin/nm -B  .libs/gdk-pixbuf.o .libs/gdk-pixbuf-animation.o
.libs/gdk-pixbuf-data.o .libs/gdk-pixbuf-io.o
.libs/gdk-pixbuf-loader.o .libs/gdk-pixbuf-scale.o
.libs/gdk-pixbuf-util.o .libs/gdk-pixdata.o
.libs/gdk-pixbuf-enum-types.o  pixops/.libs/libpixops.a |  |
/usr/bin/sed 's/.* //' | sort | uniq > .libs/libgdk_pixbuf-2.0.exp

(line wrapped)

eval: 1: Syntax error: "|" unexpected
*** Error code 2

Note the double | | at the end of the line before the last.
I tracked this down to being caused by global_symbol_pipe being unset
(simply global_symbol_pipe="") in the libtool script used
(/usr/local/bin/libtool15).

Time stamps in /var/db/pkg showed that libtool15 was installed as a
dependency of cvsup-without-gui, which was the first port I installed.

I'm not sure why global_symbol_pipe would be unset... But it was :)

I haven't reinstalled to verify that it would happen again, so this
bug-report isn't really worth much, but I don't believe in randomness
when it comes to things like this, so I guess something must be wrong
somewhere.

After having tracked it down to being libtool's fault, my first step was
to portupgrade -f libtool-1.5, and after that global_symbol_pipe was
set, so something had changed since the time I last built it (I had
installed a whole bunch of packages).

Perhaps somebody familiar with libtool could shed any light on this?

Otherwise, I guess it's nothing to worry about unless more people
report it. If somebody who reads this is about to install beta7, it
would be nice if you could try to begin by ports-installing
cvsup-without-gui and see if the resulting libtool15 script has
global_symbol_pipe set or not.

That was all! *phew*
Except for those minor problems, the experience has been a breeze.
I have really high hopes for 5.3-rel and I think it's going to be a
great release :)

Best regards,
Martin Faxer