Hi all,
(changing the title, was: It is time to modernise "groff")
On Wed, Nov 08 2017 at 06:39:03 PM, Ingo Schwarze <schwa...@usta.de> wrote:
> Gour wrote on Wed, Nov 08, 2017 at 05:57:00PM +0100:
>
>> Here I'd like to ask what has happened in the meantime in regard to
>> the groff's features and plans stated in the mission statement like
>> Knuth-Plass algorithmy for paragraph-based linebreaks, native support
>> for TrueType, Open Type, and other non-Type1 PostScript fonts, and
>> improving Unicode support ?
>
> Nothing. Setting up a mission statement has little practical effect
> when no working time is available.
Hmm... Bug #40716 "UPGRADE: format a paragraph as a whole rather than
line by line" is assigned to me, it's time to publish some of my work...
I've started to work on Knuth-Plass algorithm last year, I've written
some code to implement this algorithm but for the moment it's not
connected yet to the rest of the `troff' code. The core files of
`troff' (node.cpp, div.cpp, env.cpp) are quite large with some
inconsistencies that need to be cleaned (for example vertical_size_node
methods are not in node.cpp but in div.cpp -- presumably due to header
inclusion issues). A part from the formatting algorithm there are other
complications, for example on a diversion the current partially filled
line is included into the diversion, which will confuse a
paragraph-oriented algorithm when you end the diversion. I've made
several tries but for the moment it's inconclusive (I haven't worked on
it recently).
I've just pushed a dev branch `format_knuth_plass' with part of my work,
it includes a standalone test that uses my code with Knuth's original
example from "Digital Topography" (If you are interested by the
algorithm but don't have a copy of this book you can check this link
http://defoe.sourceforge.net/folio/knuth-plass.html -- although there
are a few mistakes in the demerits calculation).
To use this dev branch and play with the test:
git clone https://git.savannah.gnu.org/git/groff.git
cd groff
git checkout --track origin/format_knuth_plass
./bootstrap
mkdir build
cd build
../configure
make -j
make check
./test_paragraph
Attached the result of `./test_paragraph -s 1 -t 1' which run only the
test 1 of suite1 -- the original Knuth example. It shows, for each
line, the candidate breakpoints (with a `^' sign) and the calculation
of adjustement ratio and total demerits (which all correspond to the
numbers in the example from Knuth's paper).
I've also added a few things:
- Support of C++ unit test framework cppunit.
- An implementation of doubled linked list in "linux kernel" style (to
stay in the style of "C with classes" rather than real C++, and
because, well, they are familiar to me).
Of course there is still a lot of work, and these changes are not for
the next release (Still not ready to really work on it as people from
GNU haven't granted me the rights to upload tarballs into the GNU ftp
yet, but I hope this release will be completed before the end of the
year).
Regards,
Bertrand Garrigues
-- Suite 1 Initialisation
-- Test11...
Checking the number of lines
Checking the lines adjustement ratio
Checking all breakpoints demerits
Checking the best breakpoints array
Number of lines: 10 | adj.
ratio | total demerits | fitness class |
In olden times when wishing still helped one, there lived a
0.774 2209 0
^
king whose daughters were all beautiful; and the youngest was
0.179 2213 0
^ ^
so beautiful that the sun itself, which has seen so much, was
0.629 2889 0
^ ^
astonished whenever it shone in her face. Close by the king's
0.545 3178 0
^ ^
castle lay a great dark forest, and under an old lime-tree in the
0.000 3179 0
^ ^ ^ ^
forest was a well, and when the day was very warm, the king's
0.079 3180 0
^ ^ ^ ^ ^ ^
child went out into the forest and sat down by the side of the
0.282 3189 0
^ ^ ^ ^ ^ ^
cool fountain; and when she was bored she took a golden ball,
0.294 3205 0
^ ^ ^ ^ ^ ^
and threw it up on high and caught it; and this ball was her
0.575 3605 0
^ ^ ^ ^ ^ ^ ^
favorite plaything.
0.000 3606 0
^ ^ ^ ^
-- Test test11_original_example PASSED
All tests passed
