Raggedright--overlapping words

Raggedright--overlapping words

[David Antos]

	Hello all,

when using ragged right, I noticed that newer ConTeXt prefers words to
overlap boxes over pushing them to next lines. Please, what are the main
changes in penalties compared to the versions--let's say--a year ago?
How can I make ConTeXt prefer to put words to next lines instead of
pushing them out of the box?
Is plain's \raggedright re-defined? (I know, \setupalign[right] is the
politically correct name ;-)

Thanks,
D.A.

Re: Raggedright--overlapping words

[Hans Hagen]

At 07:34 PM 5/17/2002 +0200, David Antos wrote:

>         Hello all,
>
>when using ragged right, I noticed that newer ConTeXt prefers words to
>overlap boxes over pushing them to next lines. Please, what are the main
>changes in penalties compared to the versions--let's say--a year ago?
>How can I make ConTeXt prefer to put words to next lines instead of
>pushing them out of the box?
>Is plain's \raggedright re-defined? (I know, \setupalign[right] is the
>politically correct name ;-)

hm, unless i made an error, nothing changed,

\setuptolerance[tolerant]

may help

let me know if you find a diff related to this; do you have a minumal example?

Hans
-------------------------------------------------------------------------
                                   Hans Hagen | PRAGMA ADE | pragma@wxs.nl
                       Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
  tel: +31 (0)38 477 53 69 | fax: +31 (0)38 477 53 74 | www.pragma-ade.com
-------------------------------------------------------------------------
                        information: http://www.pragma-ade.com/roadmap.pdf
                     documentation: http://www.pragma-ade.com/showcase.pdf
-------------------------------------------------------------------------

Re: Raggedright--overlapping words

[David Antos]

On Fri, May 17, 2002 at 07:52:14PM +0200, Hans Hagen wrote:
> >overlap boxes over pushing them to next lines. Please, what are the main
> >changes in penalties compared to the versions--let's say--a year ago?
> >How can I make ConTeXt prefer to put words to next lines instead of
> >pushing them out of the box?
> 
> let me know if you find a diff related to this; do you have a minumal example?

On http://www.fi.muni.cz/~xantos/hans you can fetch thai-slides-s.pdf,
have a look at slides 4 and 6 (pages 5 and 7). This file was compiled on
ConTeXt 2001.7.11.

I've prepared a minimized version :-), named parbreak*, there are the
problematic slides only, with \showframe to see well, compiled on latest
stable (the same result can be obtained with latest beta).
In this version the style definition is included in the file, full
sources of the slides are on
http://www.fi.muni.cz/~xantos/patlib/EuroBachoTeX2002

I also noticed that newer ConTeXts prefer making pages longer,
especially in indexes, I'll test it better and send an example.

I also have a wish (not really related to the previous). Please, can you
put date and time similar to the stable version to the beta download
page? It would be the easiest way to check for changes.

And a web page bug report: document.html contains links to PPchTeX
manual, FAQ, and ConTeXt QRC. None of the links works.

Thanks,
D.A.

Re: Raggedright--overlapping words

[Patrick Gundlach]

David Antos <xantos@informatics.muni.cz> writes:

Hi David,

> Is plain's \raggedright re-defined? (I know, \setupalign[right] is the
> politically correct name ;-)

old (2001.7.11 context): ----------------------------------------
\raggedright ->\setraggedness \rightraggedness \setraggedskips {3}{0em}{\rightr
aggedness }{.3333em}{.5em}{1fil}{\parindent }

\setraggedness #1->\ifnum \tolerance <1500\relax \tolerance =1500\relax \fi \sp
aceskip =2.5\hsize \xspaceskip =#1\relax \divide \spaceskip by \xspaceskip \ifx
 \dohyphens \relax \hyphenpenalty =\spaceskip \fi 
#1<-\rightraggedness 

\rightraggedness ->2\bodyfontsize 

\setraggedskips #1#2#3#4#5#6#7->\def \updateraggedskips {\dosetraggedskips {#1}
{#2}{#3}{#4}{#5}{#6}{#7}}\updateraggedskips 
#1<-3
#2<-0em
#3<-\rightraggedness 
#4<-.3333em
#5<-.5em
#6<-1fil
#7<-\parindent 

\updateraggedskips ->\dosetraggedskips {3}{0em}{\rightraggedness }{.3333em}{.5e
m}{1fil}{\parindent }

\dosetraggedskips #1#2#3#4#5#6#7->\chardef \raggedstatus =#1\relax \leftskip =1
\leftskip \!!plus #2\relax \rightskip =1\rightskip \!!plus #3\relax \spaceskip 
#4\relax \xspaceskip #5\relax \parfillskip \!!zeropoint \!!plus #6\relax \parin
dent #7\relax 
#1<-3
#2<-0em
#3<-\rightraggedness 
#4<-.3333em
#5<-.5em
#6<-1fil
#7<-\parindent 

\!!plus ->plus

\!!plus ->plus

\rightraggedness ->2\bodyfontsize 

\!!zeropoint ->0pt

\!!plus ->plus
------------------------------------------------------------

new (2002.05.15 context): ----------------------------------------
\raggedright ->\setraggedness \rightraggedness \setraggedskips {3}{0em}{\rightr
aggedness }{.3333em}{.5em}{1fil}{\parindent }

\setraggedness #1->\ifnum \tolerance <1500\relax \tolerance =1500\relax \fi \sp
aceskip =2.5\hsize \xspaceskip =#1\relax \divide \spaceskip by \xspaceskip \ifx
 \dohyphens \relax \hyphenpenalty =\spaceskip \fi 
#1<-\rightraggedness 

\rightraggedness ->2\bodyfontsize 

\setraggedskips #1#2#3#4#5#6#7->\def \updateraggedskips {\dosetraggedskips {#1}
{#2}{#3}{#4}{#5}{#6}{#7}}\updateraggedskips 
#1<-3
#2<-0em
#3<-\rightraggedness 
#4<-.3333em
#5<-.5em
#6<-1fil
#7<-\parindent 

\updateraggedskips ->\dosetraggedskips {3}{0em}{\rightraggedness }{.3333em}{.5e
m}{1fil}{\parindent }

\dosetraggedskips #1#2#3#4#5#6#7->\chardef \raggedstatus =#1\relax \leftskip =1
\leftskip \!!plus #2\relax \rightskip =1\rightskip \!!plus #3\relax \spaceskip 
#4\relax \xspaceskip #5\relax \parfillskip \zeropoint \!!plus #6\relax \parinde
nt #7\relax 
#1<-3
#2<-0em
#3<-\rightraggedness 
#4<-.3333em
#5<-.5em
#6<-1fil
#7<-\parindent 

\!!plus ->plus

\!!plus ->plus

\rightraggedness ->2\bodyfontsize 

\!!plus ->plus

[...]

Overfull \hbox (38.00092pt too wide) in paragraph at lines 164--165
 [][] /cmss10@16.0pt/Pat-terns hold con-text of hy-phen-ation

\hbox(14.33603+5.57509)x228.0, glue set - 1.0
.\glue(\leftskip) 24.00006
.\hbox(14.33603+5.57509)x0.0, glue set - 24.00006fil
..\glue 0.0 plus 1.0fil minus 1.0fil
..\hbox(14.33603+5.57509)x24.00006
...\hbox(14.33603+5.57509)x24.00006, glue set 16.00003fill []
.\hbox(14.33603+5.57509)x0.0
..\rule(14.33603+5.57509)x0.0
..\glue 0.0 plus 1.0fil minus 1.0fil
.\glue 0.00012
./cmss10@16.0pt/P
.etc.

Overfull \hbox (32.35658pt too wide) in paragraph at lines 165--166
 [][] /cmss10@16.0pt/String rewrit-ing of reg-u-lar lan-guages

\hbox(14.33603+5.57509)x228.0, glue set - 1.0
.\glue(\leftskip) 24.00006
.\hbox(14.33603+5.57509)x0.0, glue set - 24.00006fil
..\glue 0.0 plus 1.0fil minus 1.0fil
..\hbox(14.33603+5.57509)x24.00006
...\hbox(14.33603+5.57509)x24.00006, glue set 16.00003fill []
.\hbox(14.33603+5.57509)x0.0
..\rule(14.33603+5.57509)x0.0
..\glue 0.0 plus 1.0fil minus 1.0fil
.\glue 0.00012
./cmss10@16.0pt/S
.etc.
------------------------------------------------------------

so, there *are* two overfull hboxes there in the new context
version. But I cannot see why.

Patrick

-- 
I'll fade into the darkness

Re: Raggedright--overlapping words

[Hans Hagen]

At 07:01 PM 5/19/2002 +0200, Patrick Gundlach wrote:
>so, there *are* two overfull hboxes there in the new context
>version. But I cannot see why.

as far as i know i didn't change tolerance and so, can it be related to 
different hyphenation patterns? \

[i did replace plain by a cleaned up context plain so i may have missed 
something, but what ...]

can you extend your experimental code with showing hyphenpenalties, 
tolerance and so?

Hans
-------------------------------------------------------------------------
                                   Hans Hagen | PRAGMA ADE | pragma@wxs.nl
                       Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
  tel: +31 (0)38 477 53 69 | fax: +31 (0)38 477 53 74 | www.pragma-ade.com
-------------------------------------------------------------------------
                        information: http://www.pragma-ade.com/roadmap.pdf
                     documentation: http://www.pragma-ade.com/showcase.pdf
-------------------------------------------------------------------------

Re: Raggedright--overlapping words

[David Antos]

On Mon, May 20, 2002 at 12:15:25AM +0200, Hans Hagen wrote:
> At 07:01 PM 5/19/2002 +0200, Patrick Gundlach wrote:
> >so, there *are* two overfull hboxes there in the new context
> >version. But I cannot see why.
> 
> as far as i know i didn't change tolerance and so, can it be related to 
> different hyphenation patterns? \

In my test code, there is \hyphenpenalty=10000, and the line breaking
differs :-(.

> [i did replace plain by a cleaned up context plain so i may have missed 
> something, but what ...]

context plain, it sounds great :-)

D.A.

Re: Raggedright--overlapping words

[Patrick Gundlach]

Hans Hagen <pragma@wxs.nl> writes:

Hello Hans and David,

> >so, there *are* two overfull hboxes there in the new context
> >version. But I cannot see why.
> 
> as far as i know i didn't change tolerance and so, can it be related
> to different hyphenation patterns? \

No, very unlikely. The \tracingparagraphs showed exactly the same
hyphenation patterns.

> [i did replace plain by a cleaned up context plain so i may have
> missed something, but what ...]
> 
> can you extend your experimental code with showing hyphenpenalties,
> tolerance and so?

I did so, and found out that tolerance, hyphenpenalty, pretolerance,
hbadness, hfuzz and localcolumnwidth are the same in 2001/7 and
2002/5.

I think I have found out that it has to do with the fact that
\startitemize now sets the \rightskip to 0pt instead of 0pt plus
32pt. But I cannot see where this is done. See the last lines of the
box output:

-------------------------------------------------- (context 2001.7.11)
\hbox(14.33603+5.57509)x228.0, glue set 1.2917
.\glue(\leftskip) 24.00006
.\hbox(7.11113+0.0)x0.0, glue set - 24.00006fil
..\glue 0.0 plus 1.0fil minus 1.0fil
..\hbox(7.11113+0.0)x24.00006
...\hbox(7.11113+0.0)x24.00006, glue set 16.00003fill
....\hbox(7.11113+0.0)x8.00003
.....\mathon
...../cmsy10@16.0pt/^^O
.....\mathoff
....\glue 0.0 plus 1.0fill
.\hbox(14.33603+5.57509)x0.0
..\rule(14.33603+5.57509)x0.0
..\glue 0.0 plus 1.0fil minus 1.0fil
.\glue 0.00012
./cmss10@16.0pt/P
.\kern-0.44446
./cmss10@16.0pt/a
./cmss10@16.0pt/t
.\discretionary
../cmss10@16.0pt/-
./cmss10@16.0pt/t
./cmss10@16.0pt/e
./cmss10@16.0pt/r
./cmss10@16.0pt/n
./cmss10@16.0pt/s
.\glue(\spaceskip) 5.33276
./cmss10@16.0pt/h
./cmss10@16.0pt/o
./cmss10@16.0pt/l
./cmss10@16.0pt/d
.\glue(\spaceskip) 5.33276
./cmss10@16.0pt/c
./cmss10@16.0pt/o
./cmss10@16.0pt/n
.\discretionary
../cmss10@16.0pt/-
./cmss10@16.0pt/t
./cmss10@16.0pt/e
./cmss10@16.0pt/x
./cmss10@16.0pt/t
.\glue(\spaceskip) 5.33276
./cmss10@16.0pt/o
./cmss10@16.0pt/f
.\glue(\rightskip) 0.0 plus 32.0
--------------------------------------------------

-------------------------------------------------- (context 2002.5.15)
\hbox(14.33603+5.57509)x228.0, glue set - 1.0
.\glue(\leftskip) 24.00006
.\hbox(14.33603+5.57509)x0.0, glue set - 24.00006fil
..\glue 0.0 plus 1.0fil minus 1.0fil
..\hbox(14.33603+5.57509)x24.00006
...\hbox(14.33603+5.57509)x24.00006, glue set 16.00003fill
....\hbox(14.33603+5.57509)x8.00003
.....\mathon
...../cmsy10@16.0pt/^^O
.....\mathoff
....\glue 0.0 plus 1.0fill
.\hbox(14.33603+5.57509)x0.0
..\rule(14.33603+5.57509)x0.0
..\glue 0.0 plus 1.0fil minus 1.0fil
.\glue 0.00012
./cmss10@16.0pt/P
.\kern-0.44446
./cmss10@16.0pt/a
./cmss10@16.0pt/t
.\discretionary
../cmss10@16.0pt/-
./cmss10@16.0pt/t
./cmss10@16.0pt/e
./cmss10@16.0pt/r
./cmss10@16.0pt/n
./cmss10@16.0pt/s
.\glue 5.33334 plus 2.66667 minus 1.77779
./cmss10@16.0pt/h
./cmss10@16.0pt/o
./cmss10@16.0pt/l
./cmss10@16.0pt/d
.\glue 5.33334 plus 2.66667 minus 1.77779
./cmss10@16.0pt/c
./cmss10@16.0pt/o
./cmss10@16.0pt/n
.\discretionary
../cmss10@16.0pt/-
./cmss10@16.0pt/t
./cmss10@16.0pt/e
./cmss10@16.0pt/x
./cmss10@16.0pt/t
.\glue 5.33334 plus 2.66667 minus 1.77779
./cmss10@16.0pt/o
./cmss10@16.0pt/f
.\glue 5.33334 plus 2.66667 minus 1.77779
./cmss10@16.0pt/h
./cmss10@16.0pt/y
.\discretionary
../cmss10@16.0pt/-
./cmss10@16.0pt/p
./cmss10@16.0pt/h
./cmss10@16.0pt/e
./cmss10@16.0pt/n
.\discretionary
../cmss10@16.0pt/-
./cmss10@16.0pt/a
./cmss10@16.0pt/t
./cmss10@16.0pt/i
./cmss10@16.0pt/o
./cmss10@16.0pt/n
.\glue(\rightskip) 0.0
--------------------------------------------------

If you (David) put a \rightskip=0pt plus 32pt after the first \item,
the output seems to be the same as with the old ConTeXt.

Have a nice day! 

 Patrick

-- 
I'll fade into the darkness

Re: Raggedright--overlapping words

[David Antos]

	Hello all,

On Mon, May 20, 2002 at 02:10:27PM +0200, Patrick Gundlach wrote:
> I think I have found out that it has to do with the fact that
> \startitemize now sets the \rightskip to 0pt instead of 0pt plus
> 32pt. But I cannot see where this is done. See the last lines of the
> box output:

why does startitemize change \rightskip? Is there a special reason for
that? (If so, why doesn't it change it in a group?)

> If you (David) put a \rightskip=0pt plus 32pt after the first \item,
> the output seems to be the same as with the old ConTeXt.

Yes, thanks for the hack, nevertheless I consider this to be a topic for
a system solution.

D.A.

Re: Raggedright--overlapping words

[Patrick Gundlach]

David Antos <xantos@informatics.muni.cz> writes:

Hi,

> why does startitemize change \rightskip? Is there a special reason for
> that? (If so, why doesn't it change it in a group?)

I have no idea if this is intentional or an unwanted side effect. Hans
should know...

> Yes, thanks for the hack, nevertheless I consider this to be a topic for
> a system solution.

Right, it was just a hint if you need it really badly.

Patrick

-- 
I'll fade into the darkness

Re: Raggedright--overlapping words

[Hans Hagen]

At 05:11 PM 5/20/2002 +0200, Patrick Gundlach wrote:
>David Antos <xantos@informatics.muni.cz> writes:
>
> > why does startitemize change \rightskip? Is there a special reason for
> > that? (If so, why doesn't it change it in a group?)
>
>I have no idea if this is intentional or an unwanted side effect. Hans
>should know...
>
> > Yes, thanks for the hack, nevertheless I consider this to be a topic for
> > a system solution.
>
>Right, it was just a hint if you need it really badly.

thanks to both of you for tracking this down so precisely, in core-itm, at 
the end you can nil the align option:

   \setupitemgroups
     [\c!uitlijnen=]

that way it follows the main settings which is what we want

Hans
-------------------------------------------------------------------------
                                   Hans Hagen | PRAGMA ADE | pragma@wxs.nl
                       Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
  tel: +31 (0)38 477 53 69 | fax: +31 (0)38 477 53 74 | www.pragma-ade.com
-------------------------------------------------------------------------
                        information: http://www.pragma-ade.com/roadmap.pdf
                     documentation: http://www.pragma-ade.com/showcase.pdf
-------------------------------------------------------------------------

Re[2]: Raggedright--overlapping words

[Giuseppe Bilotta]

Monday, May 20, 2002 Hans Hagen wrote:

HH> thanks to both of you for tracking this down so precisely, in core-itm, at
HH> the end you can nil the align option:

HH>    \setupitemgroups
HH>      [\c!uitlijnen=]

HH> that way it follows the main settings which is what we want

I assume this will be there in the next beta?

(I'm waiting just a little more before updating my ConTeXt.)

-- 
Giuseppe "Oblomov" Bilotta

Re[2]: Raggedright--overlapping words

[Hans Hagen]

At 12:18 PM 5/21/2002 +0200, Giuseppe Bilotta wrote:

>Monday, May 20, 2002 Hans Hagen wrote:
>
>HH> thanks to both of you for tracking this down so precisely, in core-itm, at
>HH> the end you can nil the align option:
>
>HH>    \setupitemgroups
>HH>      [\c!uitlijnen=]
>
>HH> that way it follows the main settings which is what we want
>
>I assume this will be there in the next beta?

Sure

>(I'm waiting just a little more before updating my ConTeXt.)

incredible

Hans

-------------------------------------------------------------------------
                                   Hans Hagen | PRAGMA ADE | pragma@wxs.nl
                       Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
  tel: +31 (0)38 477 53 69 | fax: +31 (0)38 477 53 74 | www.pragma-ade.com
-------------------------------------------------------------------------
                        information: http://www.pragma-ade.com/roadmap.pdf
                     documentation: http://www.pragma-ade.com/showcase.pdf
-------------------------------------------------------------------------