bugs just filed

[Havoc Pennington <hp@redhat.com>]

Hi,

I'm not normally involved in LSB stuff and haven't followed this list,
but have been writing some docs for Red Hat related to the "Command
Behavior" section and saw the link to the nice review form on Slashdot
yesterday, so just went through the commands section and filed a bunch
of bugs. I know it's annoying to do that mere hours after the
deadline. ;-) As I said, I am clueless LSB-wise. Hopefully the
comments are useful for the next version of the spec.

I filed separate reports for each change but they are mostly in some
major groups:

 - command line options that conflict with SUS in current
   implementations but are not really essential should not be 
   specified. e.g. "df -t" can just be left implementation dependent, 
   there is no reason to force SUS incompliance here, the option 
   isn't exactly a vital feature.
   
   At minimum, annotation should be added that POSIXLY_CORRECT 
   may change the behavior of these options, right?

 - the wording "--foo is not supported" seems to imply that an 
   option must not be supported, though I doubt that was intended.
   I think wording such as "--foo is undefined" or "--foo is
   implementation-dependent" would be better.
 
 - I don't understand why the LSB specifies interactive programs;
   those are not designed to be an ABI/API and are not a reasonable
   ABI/API. For example, I don't see how "su" can be used
   programatically (or if it can, only its noninteractive aspects
   should be described in the spec, IMO).

   Due to internationalization, checks such as isatty(), possible
   UI enhancements that change prompts, etc., interactive 
   command behavior is just not a reasonable thing to rely on.

Something I didn't file bugs on for the most part: in many cases LSB
seems to simply catalog all options in some version of a
utility. Wouldn't it be better to start with SUS options and
conservatively add only some of the really useful extensions, instead
of listing them wholesale?

I did file a bug on one egregious example, "patch --debug" makes no
sense whatsoever as part of an API/ABI spec, IMO. It's not like an ISV
app should or could use that option, so why is it required for LSB
compliance?

Thanks for everyone's hard work.

Havoc