Follow-up Comment #3, bug #68923 (group groff):

[comment #2 comment #2:]
> [comment #1 comment #1:]
>> This wouldn't strictly _depend_ on bug #67492,
> 
> Sorry if I'm being dense, but I don't follow how a formatter request like
> .pchar can tell the user anything about the behavior of the postprocessor.
> The troff | gropdf pipeline is strictly one-way, no?  And the behavioral
> difference Deri pointed out is within the grops and gropdf postprocessors,
> no?

I think you have an incomplete picture of _groff_'s codebase.

> how a formatter request like .pchar can tell the user anything about the
> behavior of the postprocessor.

As it's implemented today, it essentially can't/doesn't.

Bug #67492 points in the direction of how it might start to.  But there's
another piece to the puzzle.

> The troff | gropdf pipeline is strictly one-way, no?

Correct.  There is no back-propgation of information from a post-processor to
the formatter.

Two other things are also true:

1.  The set of post-processors that are output drivers (which is a 100%
overlap in _groff_'s own code base, but not a necessary fact) have
"description files" that both the formatter and output drivers use to predict
the behavior of the device or file format produced/driven by the output
driver.

2.  For every output driver except _gropdf_, the logic that interprets these
description files is abstracted into static libraries.

3.  The formatter uses those static libraries, too.

Ultimately the formatter tells the output driver which glyph to render at a
given drawing position by consulting the same font description files that the
output driver itself reads and interprets.

But the formatter _also_ transforms characters to glyphs by the
[https://www.gnu.org/software/groff/manual/groff.html.node/Characters-and-Glyphs.html
complex resolution process described in _groff_'s Texinfo manual].

> And the behavioral difference Deri pointed out is within the grops and gropdf
> postprocessors, no?

Yes.  But while Deri might know the details of those behavioral differences, I
don't.  Where there are discrepancies:

A.  Either _libgroff.a_ (which has _font.o_ and _fontfile.o_ object files
bundled into it) is buggy, or _gropdf_ is.  (Or both are.)

B.  I propose that getting reliable information from the formatter about the
aforementioned character-to-glyph resolution process might help us to decide
the cases in (A).

Another way of putting the issue is that bug #67492 seeks to throw light on
how an input sequence eventually becomes some combination of `f` and `C`, `c`,
`t` or `u` "grout" commands.  If the same input leads to different decisions
by the formatter depending on "ps" vs. "pdf" output, then barring some other,
better explanation (like missing glyph coverage in PostScript Type 1 fonts),
then this ticket can serve as a repository of discrepancies requiring
resolution.


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?68923>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/

Attachment: signature.asc
Description: PGP signature

Reply via email to