Hi Alex,
At 2024-09-01T00:30:26+0200, Alejandro Colomar wrote:
> Using indentation for the scripts (and cat(1) before a script starts)
> would probably help.
Ok, I'll do that in my next submission.
> > > Could you split this change into some simpler smaller patches?
> >
> > That's harder. It's a scripted change. I don't think breaking it
> > up by individual sed commands is going to enlighten anyone. But it
> > would be easy to break it up by affected file...
>
> Do the individual sed(1) commands produce good states? Or are they
> temporary steps that would not be good by themselves?
How about this? How about I improve the internal documentation of each
sed script so that it discloses the sorts of case to which each
substitution applies?
$ cat fix-man-page-refs-in-tbl-tables-1.sed
# Rewrite man page cross references inside tbl(1) text blocks to use
# man(7) macros instead of troff(1) font selection escape sequences.
/^\.\\"/b
# Case: (handled in commit 9d21f97766, 2024-07-27)
# T{
# See \fBchown\fP(2) for
# T}
/T{$/,/^T}/s/ \\fB\([^\\]*\)\\fP\(([0-9][a-z]*)\) /\
.BR \1 \2\\"1A\
/
# Case:
# T{
# the map that is loaded by the utility \fBmapscrn\fP(8).
# T}
/T{$/,/^T}/s/ \\fB\([^\\]*\)\\fP\(([0-9][a-z]*)\)\([^0-9a-z]\+\)$/\
.BR \1 \2\3\\"1B/
# Case:
# T{
# by \fBxterm\fP(1)'s \fBhpLowerleftBugCompat\fP resource).
# T}
/T{$/,/^T}/s/ \\fB\([^\\]*\)\\fP\(([0-9][a-z]*)\)\([^ ]\+\) \(.*\)/\
.BR \1 \2\3\\"1C\
\4/
I'll be taking out the injected "1A", "1B", "1C" comments before
submitting. But they were helpful in development.
Regards,
Branden
Attachment:
signature.asc
Description: PGP signature
