Re: Running the grohtml pipeline as a pipeline

[G. Branden Robinson]

Hi Alex,

At 2024-03-18T12:38:30+0100, Alejandro Colomar wrote:
> I'm considering using grohtml(1) in the Linux man-pages, to replace
> man2html(1), which crashes on tzfile(5) --which has correct man(7)--.

Well, all right.  Demand may drive improvement to grohtml more reliably
than abandonment and neglect have.

> Currently, I'm running groff(1) as its separate constituents:
> 
>       $ make build-ps -Wman2/membarrier.2
>       PRECONV         .tmp/man/man2/membarrier.2.tbl
>       TBL             .tmp/man/man2/membarrier.2.eqn
>       EQN             .tmp/man/man2/membarrier.2.ps.troff
>       TROFF           .tmp/man/man2/membarrier.2.ps.set
>       GROPS           .tmp/man/man2/membarrier.2.ps
> 
> which runs
> 
>       $ make build-ps -Wman2/membarrier.2 --debug=print \
>               | grep -Pv '^[[:upper:]]+\t';
>       preconv   man2/membarrier.2 >.tmp/man/man2/membarrier.2.tbl
>       tbl <.tmp/man/man2/membarrier.2.tbl >.tmp/man/man2/membarrier.2.eqn
>       ! (eqn -Tps   <.tmp/man/man2/membarrier.2.eqn 2>&1 
> >.tmp/man/man2/membarrier.2.ps.troff) \
>       | grep ^ >&2
>       ! (troff -man -Tps -wbreak  <.tmp/man/man2/membarrier.2.ps.troff 2>&1 
> >.tmp/man/man2/membarrier.2.ps.set) \
>       | grep ^ >&2
>       grops   <.tmp/man/man2/membarrier.2.ps.set 
> >.tmp/man/man2/membarrier.2.ps
> 
> What pipeline do you recommend for HTML, using a similar pattern?
> So far, the only thing that seems to work is
> 
>       $ preconv man2/membarrier.2 \
>               | tbl \
>               | troff -man -Thtml -wbreak man3/_Generic.3 \
>               | post-grohtml >g.html;
> 
> Although from pre-grohtml(1), maybe I shouldn't be running tbl(1)
> myself?

grohtml uniquely runs its own output-driver-specific preprocessor, but
more significantly, that preprocessor _runs troff_.  The main reason
"groff -Thtml" is relatively slow is because it runs troff as part of
running troff, scraping out chunks of PostScript and converting them to
images.

I'm not _certain_ that what you're trying to do can't be expected to
work.  I had suspected that tbl(1) had hooks in it for pre-grohtml(1)
support (by testing the `ps4html` register, the usual way this is done),
but it doesn't.

In fact, spending a few minutes looking at it, I'm not seeing anything
that seems like it should foreclose what you're trying to do.

Can you share an exhibit of a pipeline that fails and describe _how_ it
fails?

Regards,
Branden

signature.asc
Description: PGP signature