FreeBSD 9.0-RELEASE Hardware Notes is wrong.

[Simon L. B. Nielsen]

On 19 Jan 2012, at 15:09, John Baldwin wrote:

> On Wednesday, January 18, 2012 8:01:17 pm Benjamin Kaduk wrote:
>> On Wed, 18 Jan 2012, Adam Kirchhoff wrote:
>> 
>>> http://www.freebsd.org/releases/9.0R/hardware.html
>>> 
>>> This section is correct:
>>> 
>>> [i386,amd64] The snd_emu10kx(4) driver supports the following sound cards:
>>>> Creative Sound Blaster Live! (EMU10K1 Chipset). Both PCM and MIDI
>>>> interfaces are available.
>>>> Creative Sound Blaster Audigy (CA0100 and CA0101 Chipset). PCM and two
>>>> MIDI interfaces available.
>>>> Creative Sound Blaster Audigy 2 and Creative Sound Blaster Audigy 4
>>>> (CA0102 Chipset). PCM support is limited to 48kHz/16 bit stereo (192kHz/24
>>>> bit part of this chipset is not supported).
>>>> Creative Sound Blaster Audigy 2 Value (CA0108 Chipset). PCM support is
>>>> limited to 48kHz/16 bit stereo (192kHz/24 bit part of this chipset is not
>>>> supported). There is no MIDI support for this card.
>>> 
>>> 
>>> 
>>> This section is wrong:
>>> 
>>> 
>>> The snd_emu10kx(4) driver does support the following sound cards (although
>>>> they have names similar to some supported ones):
>>>> Creative Sound Blaster Live! 24-Bit, identified by FreeBSD as
>>>> Creative Sound Blaster Audigy LS / ES, identified by FreeBSD as
>>>> All other Creative sound cards with -DAT chipsets.
>>>> All Creative X-Fi series sound cards.
>>> 
>>> 
>>> 
>>> None of those cards are supported by snd_emu10kx (according to the man
>>> page, and in my own personal experience). I believe a "does support" should
>>> read "does not support".
>> 
>> Looks like the second one should be "does not support", yes.
>> 
>> Thanks for spotting it; hopefully I can prod a committer into making the 
>> change.
> 
> The snd_emu10kx(4) manpage from which this is derived is correct.  This 
> appears to be a problem with the code that parses the manpages to generate the 
> hardware release notes.  Maybe it is confused by the markup:
> 
> The
> .Nm
> driver does
> .Em not
> support the following sound cards (although they have names
> similar to some supported ones):

The parser script basically implemented the mdoc commands which were needed to generate valid output at the time, so it sounds entirely likely that ".Em foo" just wasn't implemented.

> This change fixes it.  If simon ok's it I'll commit it.  It does result in
> a new header line for malo(4) making it into the hardware notes (it may be that
> the malo(4) manpage needs to format it's hardware table differently anyway).

It's been a long time since I looked at the code and I don't have time to do any tests right now, but the patch looks sane enough. If it produces the expected output the patch is fine with me.

> Index: release/doc/share/misc/man2hwnotes.pl
> ===================================================================
> --- man2hwnotes.pl	(revision 230267)
> +++ man2hwnotes.pl	(working copy)
> @@ -324,6 +324,11 @@ sub parse {
> 	    } elsif (/^Fx/) {
> 		dlog(3, "Got Fx command");
> 		parabuf_addline(\%mdocvars, "FreeBSD");
> +	    } elsif (/^Em (.+)$/) {
> +		my ($txt, $punct_str) = split_punct_chars($1);
> +
> +		parabuf_addline(\%mdocvars,
> +				normalize("<emphasis>$txt</emphasis>$punct_str"));
> 	    } else {
> 		# Ignore all other commands.
> 		dlog(3, "Ignoring unknown command $cmd");

-- 
Simon L. B. Nielsen