Re: [vile] vile-9.8i.patch.gz

[Thomas Dickey]

On Wed, Oct 17, 2012 at 04:32:11PM -0400, Wayne Cuddy wrote:
> On Tue, Oct 16, 2012 at 08:27:17PM -0400, Thomas Dickey wrote:
> > On Tue, Oct 16, 2012 at 04:51:23PM -0400, Wayne Cuddy wrote:
> > > After building with this patch I've noticed that SQL files are no
> > > longer highlighted. This only seems to happen for SQL files. When I open
> > > them their type is correctly recognized and SQLMODE is set in the status
> > > bar but no color. Could any changes in this patch have had that effect?
> > 
> > not that I can think of - the patch doesn't touch the filters (as I
> > use them).  It's been a while since I modified the sql-filt.l rules
> > (it's one of those that takes some pain with test-cases).
> > 
> > I usually build with either builtin-filters or loadable-filters
> > depending on whether I'm debugging or installing.  External filters _could_
> > be affected by the --dataroot change (I guess - though that's for things
> > other than lib-directories).  But that's one thing to check on.
> > 
> > A quick check here shows sql coloring as expected.
> > 
> > If it's a scripting issue, the -D option of vile can be helpful.
> 
> I built with external filters and others worked so I didn't think it was
> a path problem. Attached is the output of the [Trace] buffer when I ran
> with the -D. I'm not well versed in reading this output but it appears
> that it ran the filter. And if I run the file through vile-sql-filt it
> gets marked up.
> 
> Perhaps I should just try to build again before digging much deeper.

According to the trace, it's using external filters.  I don't see
anything wrong with the command-line, but if there was something
wrong with the filter (such as not being there:-) that could explain
it.  There've been issues also with interference from people's
shell initialization (though that generally wouldn't break the
whole thing).

I usually build vile using --with-builtin-filters or --with-loadable-filters
and

        http://invisible-island.net/reflex/

except when I'm doing test-builds on some vendor-unix systems.
vile's filters do build/run with unix lex (except for 1-2 which
have gotten too big on older HPUX for instance).

However, just to check I built vile using external filters and I seem
to be able to reproduce the problem as described.  I'll work on this
(for completeness), but generally recommend the builtin- or loadable-
approach.

(I'll be out of town a few days...)
 
> [HighlightFilter]:25 (2/0):~hidden attribute-from-filter end-of-file &cat 
> &pquote %filterpath &cat %filteropts &cat ' -t' $tabstop
> ** evaluate 'cat' (0x442), 2 string args returning string
> ** evaluate 'pquote' (0x441), 1 string args returning string
> ...arg[0] = '/usr/local/lib/vile/vile-sql-filt'
> -> '/usr/local/lib/vile/vile-sql-filt'
> ...arg[0] = '/usr/local/lib/vile/vile-sql-filt'
> ** evaluate 'cat' (0x442), 2 string args returning string
> ...arg[0] = ' -i'
> ** evaluate 'cat' (0x442), 2 string args returning string
> ...arg[0] = ' -t'
> ...arg[1] = '8'
> -> ' -t8'
> ...arg[1] = ' -t8'
> -> ' -i -t8'
> ...arg[1] = ' -i -t8'
> -> '/usr/local/lib/vile/vile-sql-filt -i -t8'
> [HighlightFilter]:26 (2/0):~endif
> [HighlightFilter]:27 (1/0):~endif
> ...pop_variable(%filteropts) ERROR
> ...pop_variable(%filterpath) ERROR
> ...pop_variable(%filtername) ERROR
> ...pop_variable($curline) 1
> ...pop_variable($curcol) 1
> [HighlightFilterMsg]:2 (0/0):write-message '[Attaching attributes...]'
> BEEP
> BEEP
> BEEP
> BEEP
> *

> _______________________________________________
> vile mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/vile


-- 
Thomas E. Dickey <address@hidden>
http://invisible-island.net
ftp://invisible-island.net

signature.asc
Description: Digital signature