On Wed, Jul 25, 2018 at 07:59:33AM -0400, James McCoy wrote: > On Wed, Jul 25, 2018 at 06:49:02AM -0400, Thomas Dickey wrote: > > On Tue, Jul 24, 2018 at 09:12:07AM -0400, James McCoy wrote: > > > Ironically, the change in format was prompted by comments that upstream > > > saw in unibilium's source code. ... > > You may be referring to this: > > > > http://lists.gnu.org/archive/html/bug-ncurses/2018-04/msg00002.html > > > > which is a refinement: unibilium ignored the value, and ncurses wasn't > > making effective use of it. > > I'm referring to this change:
fwiw, the most recent change to unibilium appears to be from February 7, which is before the refinement which you're citing. Its source code (unchanged from 2012) comments on the feature. Quoting from secret/terminfo.pod: =item EXT_OFFSET_COUNT The number of entries in the string table. NB: The ncurses code for writing terminfo files calculates this as I<EXT_BOOL_COUNT> + I<EXT_NUM_COUNT> + I<EXT_STRING_COUNT> + I<EXT_STRING_COUNT> (one entry for each capability name plus one entry for each string value). The ncurses code for reading terminfo files ignores this field. NB: L<term(5)> (part of ncurses) misdocuments this field as "size of the extended string table in bytes". This is wrong. ...changing the calculation for a field (which unibilium's developers said is "wrong") isn't a _format_ change, but in this case, it's filling in a detail which was overlooked. The documentation didn't match the use they made of the value, and on review, the calculation was inconsistent with the rest of the reader. If there's additional work needed, it's a bug report against unibilium. -- Thomas E. Dickey <dic...@invisible-island.net> https://invisible-island.net ftp://ftp.invisible-island.net
signature.asc
Description: Digital signature
