On Fri, May 28, 2010 at 9:36 AM, TeXitoi <[email protected]> wrote:
> From http://www.openbsd.org/cgi-bin/cvsweb/src/etc/gettytab:
>
>  Revision 1.3: download - view: text, markup, annotated - select for diffs
>  Mon Jun 8 18:03:34 1998 UTC (11 years, 11 months ago) by deraadt
>  Branches: MAIN
>  CVS tags: OPENBSD_2_7_BASE, OPENBSD_2_7, OPENBSD_2_6_BASE,
>    OPENBSD_2_6, OPENBSD_2_5_BASE, OPENBSD_2_5, OPENBSD_2_4_BASE,
>    OPENBSD_2_4
>  Diff to: previous 1.2: preferred, coloured
>  Changes since revision 1.2: +3 -3 lines
>  default to 8 bit; let me know if this causes problems
>
> Well, maybe a little bit late, but it causes problems.

Hmm, interesting: there's no way to explicitly clear a setting from
the default entry in gettytab.  What works in other 'cap' style files
is to use 'whatever@' to stop the scan for the capability, but that's
because the other files pick up the defaults by explicitly including
them via a tc=default entry.  gettytab does defaults directly in the C
code, leaving it with no way to clear a setting.  Oops.


> I have found 3 ways to solve this problem:
...
> 2. Changing the code to not suppose anything about the defaults
>      pros: clean fix
>      cons: complicated modifications to getty

Assuming this means "handle defaults like all the other cap files via
tc=default", I agree it's the cleanest and should be less confusing,
letting the practices that work with other cap files work with
gettytab.  The big drawback is that it invalidates all existing
gettytab files.  Yuck.  Perhaps there's some way to get getcap(3) to
reveal whether the 'whatever@' syntax was used for a name?  Hmmm


Philip Guenther

Reply via email to