Hi list, I'm having some trouble with MetaPost in general and MetaObj in particular. I hope someone on this list can assist me here. I'm trying to run the appended ConTeXt file with a quite simple object "Segment" (from the manual [1], section 2.1). This script works fine as it is. I wanted to add a simple TeX label ("\tt \overbar{E}") to the drawing macro, but failed. When including the line marked with "% 1" the TeX run fails. I don't understand why. The line with "% 2" is taken frome the metaobj manual (section 5.8) [1], but including this doesn't work either. It would be great if someone could tell me what's wrong with the mp-code below and explain how to place a (TeX) label inside a metaobj object. Thanks in advance, Stefan. [1] http://tex.loria.fr/prod-graph/momanual.pdf PS: Both variants fail with mp error messages like this: "mplib > mp terminal: ! Inconsistent equation (off by 1). <to be read again> ; assignObj->(SUFFIX0)=incr(nObj_); iname_[(SUFFIX0)]=str(SUFFIX0);addclass_((S... newSegment->...roup.assignObj((SUFFIX2),"Segment") ;ObjPoint.a,b;ObjCode"@##b... <to be read again> ; <*> ...ar {E}") shifted n.a; enddef; newSegment.s; s.a=origin; drawObj(s); ; ! Redundant equation. <to be read again> ; assignObj->...bj_);iname_[(SUFFIX0)]=str(SUFFIX0); addclass_((SUFFIX0),(EXPR1... newSegment->...roup.assignObj((SUFFIX2),"Segment") ;ObjPoint.a,b;ObjCode"@##b..." <=======================================> \startMPinclusions input metaobj; \stopMPinclusions \startMPpage vardef newSegment@#= assignObj(@#,"Segment"); ObjPoint a,b; ObjCode "@#b-@#a=(1cm,2cm)"; enddef; def drawSegment(suffix n)= draw n.a--n.b; % draw btex \tt \overbar{E} etex shifted n.a; % 1 enddef; newSegment.s; % ObjLabel.s(btex hello! etex) "labpoint(a)"; % 2 s.a=origin; drawObj(s); \stopMPpage <=======================================>
On 01/18/2011 01:17 PM, Stefan Müller wrote:
Hi list,
I'm having some trouble with MetaPost in general and MetaObj in particular.
The core problem is that context mkiv loads the metapost code twice when you are typesetting labels, and that causes trouble for metaobj which does not expect that. Sorry, I do not know of any solution. Maybe Hans does? Best wishes, Taco
On 20-1-2011 9:59, Taco Hoekwater wrote:
On 01/18/2011 01:17 PM, Stefan Müller wrote:
Hi list,
I'm having some trouble with MetaPost in general and MetaObj in particular.
The core problem is that context mkiv loads the metapost code twice when you are typesetting labels, and that causes trouble for metaobj which does not expect that. Sorry, I do not know of any solution. Maybe Hans does?
not immediately Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
On Sat, 22 Jan 2011, Hans Hagen wrote:
On 20-1-2011 9:59, Taco Hoekwater wrote:
On 01/18/2011 01:17 PM, Stefan Müller wrote:
Hi list,
I'm having some trouble with MetaPost in general and MetaObj in particular.
The core problem is that context mkiv loads the metapost code twice when you are typesetting labels, and that causes trouble for metaobj which does not expect that. Sorry, I do not know of any solution. Maybe Hans does?
not immediately
A suggestion: Currently, ConTeXt does not clear the state of mplib at the end of each graphic. For example \starttext \startMPcode path p; p := fullcircle scaled 4cm ; \stopMPcode \startMPcode draw p ; \stopMPcode \stoptext draws a circle. I am not convinced that this is useful. Resetting the state of mplib after each graphic might be better. Then, after the first run, you could just set the value of mptopdf.texts[..], reset the state of mplib, and then draw the graphic in the second run. This approach will have less surprises. For example, in simpleslides module, we provide a mechanism for highlighting a part of a picture using arrows. These arrows are thick (typically 5bp). In order to scale the arrows correctly, we set: autoarrows := true ; However, in the current implementation, this effect is global. It took a while to realize that this messed up all the figures with arrows that I was drawing in a presentation (where autoarrows := false) is better. Not resetting the mplib state also has advantages. For example, it allows an easy way to do something on the second path (trivial example, but there are cases when something like this makes the code much shorter): \startMPinclusions[+] newboolean flag ; flag := false ; \stopMPinclusions \starttext \startMPpage if flag : fill boundary withcolor 0.5[blue,white] ; else : path boundary ; fi ; path A, B; A := fullsquare scaled 3cm ; B := fullsquare scaled 3cm shifted (5cm,0) ; fill A withcolor white ; fill B withcolor white ; draw A ; draw btex A etex ; draw B ; draw btex B etex shifted (5cm,0) ; boundary := (boundingbox currentpicture) enlarged 3mm ; flag := true ; \stopMPpage \stoptext So, perhaps, we can have a `reset` key to \defineMPinstance, and let the user choose whether to reset after each graphic or not. If `reset=yes` is chosen, then such errors would not occur. Aditya
So its either using text or using metaobj? Too bad, metaobj is quite fun... I still hope you guys can fix this, but it's not urgent (to me). Just as a wild guess I tested the example below today, trying to make use of the things Aditya said, but to no avail. Anyway, what does the [+] after \startMPinclusions? I did not find documentation for that on the wiki. Best regards, Stefan. \startMPinclusions input metaobj; boolean flag; flag := false; \stopMPinclusions \starttext \startMPpage vardef newSegment@\#= assignObj(@\#,"Segment"); ObjPoint a,b; ObjCode "@\#b-@\#a=(1cm,2cm)"; enddef; def drawSegment(suffix n)= draw n.a--n.b; enddef; newSegment.s; s.a=origin; drawObj(s); if flag: draw btex A etex; fi; flag := true; \stopMPpage \stoptext On 22.01.2011 02:38, Aditya Mahajan wrote:
On Sat, 22 Jan 2011, Hans Hagen wrote:
On 20-1-2011 9:59, Taco Hoekwater wrote:
On 01/18/2011 01:17 PM, Stefan Müller wrote:
Hi list,
I'm having some trouble with MetaPost in general and MetaObj in particular.
The core problem is that context mkiv loads the metapost code twice when you are typesetting labels, and that causes trouble for metaobj which does not expect that. Sorry, I do not know of any solution. Maybe Hans does?
not immediately
A suggestion: Currently, ConTeXt does not clear the state of mplib at the end of each graphic. For example
\starttext \startMPcode path p; p := fullcircle scaled 4cm ; \stopMPcode
\startMPcode draw p ; \stopMPcode \stoptext
draws a circle. I am not convinced that this is useful. Resetting the state of mplib after each graphic might be better. Then, after the first run, you could just set the value of mptopdf.texts[..], reset the state of mplib, and then draw the graphic in the second run.
This approach will have less surprises. For example, in simpleslides module, we provide a mechanism for highlighting a part of a picture using arrows. These arrows are thick (typically 5bp). In order to scale the arrows correctly, we set:
autoarrows := true ;
However, in the current implementation, this effect is global. It took a while to realize that this messed up all the figures with arrows that I was drawing in a presentation (where autoarrows := false) is better.
Not resetting the mplib state also has advantages. For example, it allows an easy way to do something on the second path (trivial example, but there are cases when something like this makes the code much shorter):
\startMPinclusions[+] newboolean flag ; flag := false ; \stopMPinclusions
\starttext \startMPpage if flag : fill boundary withcolor 0.5[blue,white] ; else : path boundary ; fi ;
path A, B;
A := fullsquare scaled 3cm ; B := fullsquare scaled 3cm shifted (5cm,0) ;
fill A withcolor white ; fill B withcolor white ;
draw A ; draw btex A etex ; draw B ; draw btex B etex shifted (5cm,0) ;
boundary := (boundingbox currentpicture) enlarged 3mm ;
flag := true ;
\stopMPpage \stoptext
So, perhaps, we can have a `reset` key to \defineMPinstance, and let the user choose whether to reset after each graphic or not. If `reset=yes` is chosen, then such errors would not occur.
Aditya
___________________________________________________________________________________ If your question is of interest to others as well, please add an entry to the Wiki!
maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________
On Sat, 22 Jan 2011, Stefan Müller wrote:
So its either using text or using metaobj? Too bad, metaobj is quite fun... I still hope you guys can fix this, but it's not urgent (to me).
Anyway, what does the [+] after \startMPinclusions? I did not find documentation for that on the wiki.
\startMPinclusions path a ; \stopMPinclusions \startMPinclusions a := (0,0) -- (1cm,0) ; \stopMPinclusions will fail, because the second inclusions overwrites the first. If you use [+], then the second inclusions in appended after the first.
Just as a wild guess I tested the example below today, trying to make use of the things Aditya said, but to no avail.
The trick is to ensure that the offending code (assignObj) is not run twice. \startMPinclusions[+] input metaobj; boolean flag; flag := false; \stopMPinclusions \startMPdefinitions vardef newSegment@\#= if not flag : assignObj(@\#,"Segment"); ObjPoint a,b; ObjCode "@\#b-@\#a=(1cm,2cm)"; fi ; enddef; def drawSegment(suffix n)= draw n.a--n.b; draw btex \tt \overbar{E} etex shifted n.a; % 1 enddef; \stopMPdefinitions \starttext \startMPpage newSegment.s; ObjLabel.s(btex hello! etex) "labpoint(a)"; % 2 s.a=origin; drawObj(s); flag := true ; \stopMPpage \stoptext @Hans, can we have a `trialtypesetting` for MP code? Aditya
On 22.01.2011 17:30, Aditya Mahajan wrote:
On Sat, 22 Jan 2011, Stefan Müller wrote:
So its either using text or using metaobj? Too bad, metaobj is quite fun... I still hope you guys can fix this, but it's not urgent (to me).
Anyway, what does the [+] after \startMPinclusions? I did not find documentation for that on the wiki.
\startMPinclusions path a ; \stopMPinclusions
\startMPinclusions a := (0,0) -- (1cm,0) ; \stopMPinclusions
will fail, because the second inclusions overwrites the first. If you use [+], then the second inclusions in appended after the first.
Thank you for the explanation. So if I only have one "inclusions section" it doesn't matter?
Just as a wild guess I tested the example below today, trying to make use of the things Aditya said, but to no avail.
The trick is to ensure that the offending code (assignObj) is not run twice.
Your example works perfect. Many thanks for that as well. I didn't realize, that assignObj is causing the trouble, but now it's obvious. I'm just curious: the MP-code is only run twice by context if it contains labels, right? Otherwise I don't understand why the assignObj isn't always failing.
@Hans, can we have a `trialtypesetting` for MP code?
Aditya
Stefan
On 23.01.2011 12:40, Stefan Müller wrote:
On 22.01.2011 17:30, Aditya Mahajan wrote:
On Sat, 22 Jan 2011, Stefan Müller wrote:
So its either using text or using metaobj? Too bad, metaobj is quite fun... I still hope you guys can fix this, but it's not urgent (to me).
Anyway, what does the [+] after \startMPinclusions? I did not find documentation for that on the wiki.
\startMPinclusions path a ; \stopMPinclusions
\startMPinclusions a := (0,0) -- (1cm,0) ; \stopMPinclusions
will fail, because the second inclusions overwrites the first. If you use [+], then the second inclusions in appended after the first.
Thank you for the explanation. So if I only have one "inclusions section" it doesn't matter?
Just as a wild guess I tested the example below today, trying to make use of the things Aditya said, but to no avail.
The trick is to ensure that the offending code (assignObj) is not run twice.
Your example works perfect. Many thanks for that as well. I didn't realize, that assignObj is causing the trouble, but now it's obvious. I'm just curious: the MP-code is only run twice by context if it contains labels, right? Otherwise I don't understand why the assignObj isn't always failing.
Sorry for double posting... The method with "if not flag" is a bit arkward, because assignObj isn't the only problem. "rotateObj(s, 90)" would rotate the object 180 degrees instead of 90; when defining the relation between several objects some equations may became redundant (my script is rather complex and I'm not sure why some of the equations cause trouble). Stefan
On Sun, 23 Jan 2011, Stefan Müller wrote:
Sorry for double posting... The method with "if not flag" is a bit arkward, because assignObj isn't the only problem. "rotateObj(s, 90)" would rotate the object 180 degrees instead of 90; when defining the relation between several objects some equations may became redundant (my script is rather complex and I'm not sure why some of the equations cause trouble).
Depends on the internal implementation of MetaObj (I have not looked at that). You can try to iteratively add stuff and see what breaks, and then add if not flag around them. Sorry, but I do not know a more elegant way around this. In the worst case, you can switch to MkII to draw your figures :-\ Aditya
On 23.01.2011 18:22, Aditya Mahajan wrote:
On Sun, 23 Jan 2011, Stefan Müller wrote:
Sorry for double posting... The method with "if not flag" is a bit arkward, because assignObj isn't the only problem. "rotateObj(s, 90)" would rotate the object 180 degrees instead of 90; when defining the relation between several objects some equations may became redundant (my script is rather complex and I'm not sure why some of the equations cause trouble).
Depends on the internal implementation of MetaObj (I have not looked at that). You can try to iteratively add stuff and see what breaks, and then add if not flag around them.
Okay, that's what I do now. It seems it's savest to just put "if not flags" around the object constructors, object transformations and definition of inter-object relations. This does not produce graphical output, so it should be save to run it only once.
Sorry, but I do not know a more elegant way around this. In the worst case, you can switch to MkII to draw your figures :-\
Aditya
Thank you very much! Stefan.
Wikified: http://wiki.contextgarden.net/MetaObj_and_Labels Please update/change as appropriate. Best regards, Stefan
participants (4)
-
Aditya Mahajan -
Hans Hagen -
Stefan Müller -
Taco Hoekwater