Hi to all, I'm including many externalfigures in pdf in a MPgraphic (300 pdfs, more or less). The MPgraphic draws some arrows connecting the pdf fragments. (it is the musical project using lilypond we discussed some time ago). Till now, everything has always gone very well. Now I'm getting a little crazy. in fact, I'm having a strange behaviour I cannot explain. When I render the context file, as output I have the MP drawing, but all the (many) pdfs are replaced with blank spaces. I noticed that it seems that pdfs are not loaded (the console doesn't print their names in any way). I also checked the utility files (.mp, etc) but I'm not able to extract any useful info. If I comment out some of the lines requesting \externalfigure, (remaining) pdfs are again displayed. So, I'm guessing that is should be something related with a maximum of \externalfigure. I'm surprised because I generated examples containing many more pdf fragments than the present one, and I have never had any problem. Here's the zipped directory I'm working with: www.semiotiche.it/andrea/membrana/nodi3-.zip Any hint/help is much appreciated Best -a- Andrea Valle DAMS - Facoltà di Scienze della Formazione Università degli Studi di Torino andrea.valle@unito.it
andrea valle wrote:
If I comment out some of the lines requesting \externalfigure, (remaining) pdfs are again displayed.
No such luck here, I get empty rectangles whatever I comment out (randomly). I you can produce a file for me that *does* work, perhaps I can look for differences. Taco
(Thanks Taco) Could you try with this? www.semiotiche.it/andrea/membrana/nodi3-.tex It's the same file but I commented out a part of the section dedicated to importing and drawing little notes attached to arrowlines between music fragments. Everything not commented is now displayed: on my machine it seems to be deterministic. -a- PS. I'm on macosx 10.3.9 using context via texshop On 7 Nov 2005, at 09:24, Taco Hoekwater wrote:
andrea valle wrote:
If I comment out some of the lines requesting \externalfigure, (remaining) pdfs are again displayed.
No such luck here, I get empty rectangles whatever I comment out (randomly). I you can produce a file for me that *does* work, perhaps I can look for differences.
Taco
_______________________________________________ ntg-context mailing list ntg-context@ntg.nl http://www.ntg.nl/mailman/listinfo/ntg-context
Andrea Valle DAMS - Facoltà di Scienze della Formazione Università degli Studi di Torino andrea.valle@unito.it
andrea valle wrote:
It's the same file but I commented out a part of the section dedicated to importing and drawing little notes attached to arrowlines between music fragments. Everything not commented is now displayed: on my machine it seems to be deterministic.
Ok, got it. There is a bug in mp-spec.mp (inside metafun). Because of a wrapping number, it fails to recognize the fact that there are embedded specials at all. Attached is an updated version of mp-spec.mp. The corrected definition of add_special_signal is: vardef add_special_signal = % write the version number if (length _global_specials_<>0) or (length _local_specials_ <>0) : special ("%%MetaPostSpecials: 1.0 " & decimal _special_signal_ ) ; fi ; enddef ; It now tests for "not equal to zero" instead of "larger than zero": because of all the included files, the string _local_specials_ becomes longer than the maximum number 'length' can return, so it returns -32768 instead, and that is of course less than zero. Don't forget that you have to re-create the metafun format as well! Cheers, Taco %D \module %D [ file=mp-spec.mp, %D version=1999.6.26, %D title=\CONTEXT\ \METAPOST\ graphics, %D subtitle=special extensions, %D author=Hans Hagen, %D date=\currentdate, %D copyright={PRAGMA / Hans Hagen \& Ton Otten}] %C %C This module is part of the \CONTEXT\ macro||package and is %C therefore copyrighted by \PRAGMA. See licen-en.pdf for %C details. % Spot colors are not handled by mptopdf ! % (r,g,b) => cmyk : r=123 g= 1 b=hash % => spot : r=123 g= 2 b=hash % => transparent rgb : r=123 g= 3 b=hash % => transparent cmyk : r=123 g= 4 b=hash % => transparent spot : r=123 g= 5 b=hash % => rest : r=123 g=n>10 b=whatever %D This module is rather preliminary and subjected to %D changes. Here we closely cooperates with the \METAPOST\ %D to \PDF\ converter module built in \CONTEXT\ and provides %D for instance shading. More information can be found in %D type {supp-mpe.tex}. if unknown context_tool : input mp-tool ; fi ; if known context_spec : endinput ; fi ; boolean context_spec ; context_spec := true ; numeric _special_counter_ ; _special_counter_ := 0 ; numeric _color_counter_ ; _color_counter_ := 11 ; % < 10 reserved numeric _special_signal_ ; _special_signal_ := 123 ; %D When set to \type {true}, shading will be supported. Some %D day I will also write an additional directive. boolean _inline_specials_ ; _inline_specials_ := false ; %D Because we want to output only those specials that are %D actually used in a figure, we need a bit complicated %D bookkeeping and collection of specials. At the cost of some %D obscurity, we now have rather efficient resources. string _global_specials_ ; _global_specials_ := "" ; string _local_specials_ ; _local_specials_ := "" ; vardef add_special_signal = % write the version number if (length _global_specials_<>0) or (length _local_specials_ <>0) : special ("%%MetaPostSpecials: 1.0 " & decimal _special_signal_ ) ; fi ; enddef ; vardef add_extra_specials = scantokens _global_specials_ ; scantokens _local_specials_ ; enddef ; vardef reset_extra_specials = % only local ones _local_specials_ := "" ; enddef ; boolean insidefigure ; insidefigure := false ; % todo: alleen als special gebruikt flush extra_beginfig := " insidefigure := true ; " & " reset_extra_specials ; " & extra_beginfig ; extra_endfig := " add_special_signal ; " & extra_endfig & " add_extra_specials ; " & " reset_extra_specials ; " & " insidefigure := false ; " ; def set_extra_special (expr s) = if insidefigure : _local_specials_ := _local_specials_ & s ; else : _global_specials_ := _global_specials_ & s ; fi enddef ; def flush_special (expr typ, siz, dat) = _special_counter_ := _special_counter_ + 1 ; if _inline_specials_ : set_extra_special ( "special " & "(" & ditto & dat & " " & decimal _special_counter_ & " " & decimal typ & " " & decimal siz & " special" & ditto & ");" ) ; else : set_extra_special ( "special " & "(" & ditto & "%%MetaPostSpecial: " & decimal siz & " " & dat & " " & decimal _special_counter_ & " " & decimal typ & ditto & ");" ) ; fi ; enddef ; %D The next hack is needed in case you use a version of %D \METAPOST\ that does not provide you the means to configure %D the buffer size. Patrick Gundlach suggested to use arrays %D in this case. boolean bufferhack ; bufferhack := false ; % true ; if bufferhack : string _global_specials_[] ; numeric _nof_global_specials_ ; string _local_specials_[] ; numeric _nof_local_specials_ ; _nof_global_specials_ := _nof_local_specials_ := 0 ; vardef add_special_signal = % write the version number if (_nof_global_specials_>0) or (_nof_local_specials_>0) : special ("%%MetaPostSpecials: 1.0 " & decimal _special_signal_ ) ; fi ; enddef ; vardef add_extra_specials = for i=1 upto _nof_global_specials_ : scantokens _global_specials_[i] ; endfor; for i=1 upto _nof_local_specials_ : scantokens _local_specials_[i] ; endfor; enddef ; vardef reset_extra_specials = string _local_specials_[] ; _nof_local_specials_ := 0 ; enddef ; def set_extra_special (expr s) = if insidefigure : _local_specials_[incr(_nof_local_specials_)] := s ; else : _global_specials_[incr(_nof_global_specials_)] := s ; fi enddef ; fi ; %D So far for this hack. %D Shade allocation. newinternal shadefactor ; shadefactor := 1 ; pair shadeoffset ; shadeoffset := origin ; vardef define_linear_shade (expr a, b, ca, cb) = flush_special(30, 15, "0 1 " & decimal shadefactor & " " & dddecimal ca & " " & ddecimal (a shifted shadeoffset) & " " & dddecimal cb & " " & ddecimal (b shifted shadeoffset) ) ; _special_counter_ enddef ; vardef define_circular_shade (expr a, b, ra, rb, ca, cb) = flush_special(31, 17, "0 1 " & decimal shadefactor & " " & dddecimal ca & " " & ddecimal (a shifted shadeoffset) & " " & decimal ra & " " & dddecimal cb & " " & ddecimal (b shifted shadeoffset) & " " & decimal rb ) ; _special_counter_ enddef ; %D A few predefined shading macros. boolean trace_shades ; trace_shades := false ; % if (n=1) : a := llcorner p ; b := urcorner p ; % elseif (n=2) : a := llcorner p ; b := ulcorner p ; % elseif (n=3) : a := lrcorner p ; b := ulcorner p ; % else : a := llcorner p ; b := lrcorner p ; % fi ; def set_linear_vector (suffix a,b)(expr p,n) = if (n=1) : a := llcorner p ; b := urcorner p ; elseif (n=2) : a := lrcorner p ; b := ulcorner p ; elseif (n=3) : a := urcorner p ; b := llcorner p ; elseif (n=4) : a := ulcorner p ; b := lrcorner p ; elseif (n=5) : a := .5[ulcorner p,llcorner p] ; b := .5[urcorner p,lrcorner p] ; elseif (n=6) : a := .5[llcorner p,lrcorner p] ; b := .5[ulcorner p,urcorner p] ; elseif (n=7) : a := .5[lrcorner p,urcorner p] ; b := .5[llcorner p,ulcorner p] ; elseif (n=8) : a := .5[urcorner p,ulcorner p] ; b := .5[lrcorner p,llcorner p] ; else : a := .5[ulcorner p,llcorner p] ; b := .5[urcorner p,lrcorner p] ; fi ; enddef ; def linear_shade (expr p, n, ca, cb) = begingroup ; save a, b, sh ; pair a, b ; set_linear_vector(a,b)(p,n) ; fill p withshade define_linear_shade (a,b,ca,cb) ; if trace_shades : drawarrow a -- b withpen pencircle scaled 1pt ; fi ; endgroup ; enddef ; vardef predefined_linear_shade (expr p, n, ca, cb) = save a, b, sh ; pair a, b ; set_linear_vector(a,b)(p,n) ; set_shade_vector(a,b)(p,n) ; define_linear_shade (a,b,ca,cb) enddef ; def set_circular_vector (suffix ab, r)(expr p,n) = if (n=1) : ab := llcorner p ; elseif (n=2) : ab := lrcorner p ; elseif (n=3) : ab := urcorner p ; elseif (n=4) : ab := ulcorner p ; else : ab := center p ; r := .5r ; fi ; enddef ; def circular_shade (expr p, n, ca, cb) = begingroup ; save ab, r ; pair ab ; numeric r ; r := (xpart lrcorner p - xpart llcorner p) ++ (ypart urcorner p - ypart lrcorner p) ; set_circular_vector(ab,r)(p,n) ; fill p withshade define_circular_shade(ab,ab,0,r,ca,cb) ; if trace_shades : drawarrow ab -- ab shifted (0,r) withpen pencircle scaled 1pt ; fi ; endgroup ; enddef ; vardef predefined_circular_shade (expr p, n, ca, cb) = save ab, r ; pair ab ; numeric r ; r := (xpart lrcorner p - xpart llcorner p) ++ (ypart urcorner p - ypart lrcorner p) ; set_circular_vector(ab,r)(p,n) ; define_circular_shade(ab,ab,0,r,ca,cb) enddef ; %D Since a \type {fill p withshade s} syntax looks better %D than some macro, we implement a new primary. primarydef p withshade sc = % == p withcolor shadecolor(sh) hide (_color_counter_ := _color_counter_ + 1) p withcolor (_special_signal_/1000,_color_counter_/1000,sc/1000) enddef ; vardef shadecolor(expr sc) = hide (_color_counter_ := _color_counter_ + 1) (_special_signal_/1000,_color_counter_/1000,sc/1000) enddef ; %D Figure inclusion. %numeric cef ; cef := 0 ; def externalfigure primary filename = doexternalfigure (filename) enddef ; def doexternalfigure (expr filename) text transformation = begingroup ; save p, t ; picture p ; transform t ; p := nullpicture ; t := identity transformation ; flush_special(10, 9, dddecimal (xxpart t, yxpart t, xypart t) & " " & dddecimal (yypart t, xpart t, ypart t) & " " & filename) ; addto p contour unitsquare scaled 0 ; setbounds p to unitsquare transformed t ; _color_counter_ := _color_counter_ + 1 ; draw p withcolor (_special_signal_/1000,_color_counter_/1000,_special_counter_/1000) ; %draw p withcolor (_special_signal_/1000,cef/1000,_special_counter_/1000) ; endgroup ; enddef ; %D Experimental: %numeric currenthyperlink ; currenthyperlink := 0 ; def hyperlink primary t = dohyperlink(t) enddef ; def hyperpath primary t = dohyperpath(t) enddef ; def dohyperlink (expr destination) text transformation = begingroup ; save somepath ; path somepath ; somepath := fullsquare transformation ; dohyperpath(destination) somepath ; endgroup ; enddef ; def dohyperpath (expr destination) expr somepath = begingroup ; flush_special(20, 7, ddecimal (xpart llcorner somepath, ypart llcorner somepath) & " " & ddecimal (xpart urcorner somepath, ypart urcorner somepath) & " " & destination) ; % currenthyperlink := currenthyperlink + 1 ; _color_counter_ := _color_counter_ + 1 ; fill boundingbox unitsquare scaled 0 withcolor (_special_signal_/1000,_color_counter_/1000,_special_counter_/1000) ; % (_special_signal_/1000,currenthyperlink/1000,_special_counter_/1000) ; endgroup ; enddef ; % \setupinteraction[state=start] % \setupcolors [state=start] % % Hello There! \blank % % \startMPcode % pickup pencircle scaled 5 ; % draw fullcircle scaled 4cm withcolor red ; % hyperpath "nextpage" boundingbox currentpicture ; % draw origin withcolor blue ; % \stopMPcode % % \blank Does it work or not? % % \startMPcode % pickup pencircle scaled 5 ; % draw fullcircle scaled 4cm withcolor red ; % hyperpath "nextpage" fullcircle scaled 4cm ; % draw origin withcolor blue ; % draw fullcircle scaled 4cm shifted (1cm,1cm); % \stopMPcode % % \blank Does it work or not? \page Hello There! \blank % % \startMPcode % pickup pencircle scaled 5 ; % draw fullcircle scaled 2cm shifted (-2cm,-1cm) ; % draw fullcircle scaled 3cm shifted (2cm,1cm) withcolor red ; % draw fullcircle scaled 1cm ; % hyperlink "previouspage" scaled 3cm shifted (2cm,1cm) ; % draw origin withcolor blue ; % \stopMPcode % % \blank Does it work or not? _cmyk_counter_ := 0 ; extra_endfig := " resetcmykcolors ; " & extra_endfig ; def resetcmykcolors = numeric cmykcolorhash[][][][] ; enddef ; resetcmykcolors ; boolean cmykcolors ; cmykcolors := false ; % true string cmykcolorpattern[] ; % needed for transparancies vardef cmyk(expr c,m,y,k) = if cmykcolors : save ok ; boolean ok ; if unknown cmykcolorhash[c][m][y][k] : ok := false ; % not yet defined elseif cmykcolorhash[c][m][y][k] = -1 : ok := false ; % locally defined and undefined else : ok := true ; % globally already defined fi ; if not ok : save s ; string s ; s := dddecimal (c,m,y) & " " & decimal k ; _cmyk_counter_ := _cmyk_counter_ + 1 ; cmykcolorpattern[_cmyk_counter_/1000] := s ; cmykcolorhash[c][m][y][k] := _cmyk_counter_ ; flush_special(1, 7, decimal _cmyk_counter_ & " " & s) ; _local_specials_ := _local_specials_ & " cmykcolorhash[" & decimal c & "][" & decimal m & "][" & decimal y & "][" & decimal k & "] := -1 ; " ; fi ; (_special_signal_/1000,1/1000,cmykcolorhash[c][m][y][k]/1000) else : (1-c-k,1-m-k,1-y-k) fi enddef ; % newcolor truecyan, truemagenta, trueyellow ; % % truecyan = cmyk (1,0,0,0) ; % truemagenta = cmyk (0,1,0,0) ; % trueyellow = cmyk (0,0,1,0) ; %D Spot colors _spotcolor_counter_ := 0 ; _spotcolor_number_ := 0 ; extra_endfig := " resetspotcolors ; " & extra_endfig ; def resetspotcolors = numeric spotcolorhash[][] ; enddef ; resetspotcolors ; boolean spotcolors ; spotcolors := false ; % true string spotcolorpattern[] ; % needed for transparancies % vardef spotcolor(expr p, s) = % if spotcolors : % save ok, pc_tag ; boolean ok ; string pc_tag ; % pc_tag := "_pct_"&p ; % if not unstringed(pc_tag) : % _spotcolor_number_ := _spotcolor_number_ + 1 ; % setunstringed(pc_tag,_spotcolor_number_) ; % fi ; % pp := getunstringed(pc_tag) ; % if unknown spotcolorhash[pp][s] : % ok := false ; % not yet defined % elseif spotcolorhash[pp][s] = -1 : % ok := false ; % locally defined and undefined % else : % ok := true ; % globally already defined % fi ; % if not ok : % save ss ; string ss ; ss := p & " " & decimal s ; % _spotcolor_counter_ := _spotcolor_counter_ + 1 ; % spotcolorpattern[_spotcolor_counter_/1000] := ss ; % spotcolorhash[pp][s] := _spotcolor_counter_ ; % flush_special(2, 5, decimal _spotcolor_counter_ & " " & ss) ; % _local_specials_ := _local_specials_ & % "spotcolorhash["&decimal pp&"]["&decimal s&"]:=-1;" ; % fi ; % (_special_signal_/1000,2/1000,spotcolorhash[pp][s]/1000) % else : % (1-s,1-s,1-s) % fi % enddef ; % vardef spotcolor(expr p, s) = % if spotcolors : % save ok, pc_tag ; boolean ok ; string pc_tag ; % pc_tag := "_pct_"&p ; % if not unstringed(pc_tag) : % _spotcolor_number_ := _spotcolor_number_ + 1 ; % setunstringed(pc_tag,_spotcolor_number_) ; % fi ; % pp := getunstringed(pc_tag) ; % if unknown spotcolorhash[pp][s] : % ok := false ; % not yet defined % elseif spotcolorhash[pp][s] = -1 : % ok := false ; % locally defined and undefined % else : % ok := true ; % globally already defined % fi ; % if not ok : % save ss ; string ss ; ss := p & " " & decimal s ; % _spotcolor_counter_ := _spotcolor_counter_ + 1 ; % spotcolorpattern[_spotcolor_counter_/1000] := ss ; % spotcolorhash[pp][s] := _spotcolor_counter_ ; % flush_special(2, 5, decimal _spotcolor_counter_ & " " & ss) ; % _local_specials_ := _local_specials_ & % "spotcolorhash["&decimal pp&"]["&decimal s&"]:=-1;" ; % fi ; % (_special_signal_/1000,2/1000,spotcolorhash[pp][s]/1000) % else : % (1-s,1-s,1-s) % fi % enddef ; vardef spotcolor(expr p, s) = multitonecolor(p, 1, "", decimal s) enddef ; vardef multitonecolor(expr n, f, d, p) = % name fractions names factors if spotcolors : save ok, pc_tag ; boolean ok ; string pc_tag ; pc_tag := "_pct_" & n ; if not unstringed(pc_tag) : _spotcolor_number_ := _spotcolor_number_ + 1 ; setunstringed(pc_tag,_spotcolor_number_) ; fi ; pp := getunstringed(pc_tag) ; pc_tag := "_pct_"& decimal f & "_" & if d = "" : n else : d fi & "_" & p ; % check for d empty if not unstringed(pc_tag) : _spotcolor_number_ := _spotcolor_number_ + 1 ; setunstringed(pc_tag,_spotcolor_number_) ; fi ; ps := getunstringed(pc_tag) ; if unknown spotcolorhash[pp][ps] : ok := false ; % not yet defined elseif spotcolorhash[pp][ps] = -1 : ok := false ; % locally defined and undefined else : ok := true ; % globally already defined fi ; if not ok : save ss ; string ss ; ss := n & " " & decimal f & " " & if d = "" : n else : d fi & " " & p ; _spotcolor_counter_ := _spotcolor_counter_ + 1 ; spotcolorpattern[_spotcolor_counter_/1000] := ss ; spotcolorhash[pp][ps] := _spotcolor_counter_ ; flush_special(2, 7, decimal _spotcolor_counter_ & " " & ss) ; _local_specials_ := _local_specials_ & "spotcolorhash["&decimal pp&"]["&decimal ps&"]:=-1;" ; fi ; (_special_signal_/1000,2/1000,spotcolorhash[pp][ps]/1000) else : .5white fi enddef ; %D Transparency normaltransparent := 1 ; multiplytransparent := 2 ; screentransparent := 3 ; overlaytransparent := 4 ; softlighttransparent := 5 ; hardlighttransparent := 6 ; colordodgetransparent := 7 ; colorburntransparent := 8 ; darkentransparent := 9 ; lightentransparent := 10 ; differencetransparent := 11 ; exclusiontransparent := 12 ; % nottransparent := 0 ; % compatibletransparent := 99 ; % fill fullcircle scaled 10cm withcolor transparant(.8,3,color) ; % vardef transparent(expr n, t, c) = % save s, ss, nn, cc, is_cmyk, is_spot, ok ; % string s, ss ; numeric nn ; color cc ; boolean is_cmyk, is_spot, ok ; % % transparancy type % if string n : % if expandafter known scantokens(n&"transparent") : % nn := scantokens(n&"transparent") ; % else : % nn := 0 ; % fi % else : % nn := min(n,13) % nn := if n<13 : n else : nn := 0 fi ; % fi ; % % we need to expand the color (can be cmyk(..) or predefined) % cc := c ; % expand color % % check for cmyk special % is_cmyk := (redpart cc = _special_signal_/1000) % and (greenpart cc = 1/1000) ; % is_spot := (redpart cc = _special_signal_/1000) % and (greenpart cc = 2/1000) ; % % build special string, fetch cmyk components % s := decimal nn & " " & decimal t & " " & % if is_cmyk : cmykcolorpattern[bluepart cc] % elseif is_spot : spotcolorpattern[bluepart cc] % else : dddecimal cc fi ; % % check if this one is already used % ss := "tr_" & s ; % % efficiency hack % if expandafter unknown scantokens(ss) : % ok := false ; % not yet defined % elseif scantokens(ss) < 0 : % ok := false ; % locally defined and undefined % else : % ok := true ; % globally already defined % fi ; % if not ok : % if is_spot : % flush_special(5, 6, s) ; % elseif is_cmyk : % flush_special(4, 8, s) ; % else : % flush_special(3, 7, s) ; % fi ; % scantokens(ss) := _special_counter_ ; % _local_specials_ := _local_specials_ & % "scantokens(" & ditto & ss & ditto & ") := -1 ;" ; % fi ; % % go ahead % if is_spot : % (_special_signal_/1000,5/1000,scantokens(ss)/1000) % elseif is_cmyk : % (_special_signal_/1000,4/1000,scantokens(ss)/1000) % else : % (_special_signal_/1000,3/1000,scantokens(ss)/1000) % fi % enddef ; vardef transparent(expr n, t, c) = save s, ss, nn, cc, is_cmyk, is_spot, ok ; string s, ss ; numeric nn ; color cc ; boolean is_cmyk, is_spot, ok ; % transparancy type if string n : if expandafter known scantokens(n&"transparent") : nn := scantokens(n&"transparent") ; else : nn := 0 ; fi else : % nn := min(n,13) nn := if n<13 : n else : nn := 0 fi ; fi ; % we need to expand the color (can be cmyk(..) or predefined) cc := c ; % expand color % check for cmyk special is_cmyk := (redpart cc = _special_signal_/1000) and (greenpart cc = 1/1000) ; is_spot := (redpart cc = _special_signal_/1000) and (greenpart cc = 2/1000) ; % build special string, fetch cmyk components s := decimal nn & " " & decimal t & " " & if is_cmyk : cmykcolorpattern[bluepart cc] elseif is_spot : spotcolorpattern[bluepart cc] else : dddecimal cc fi ; % check if this one is already used ss := cleanstring("tr_" & s) ; % we now have rather unique names, i.e. a color spec of .234 becomes % tr..._234.... and metapost gives a number overflow (parse error) % for variables like tr_12345678 which may result from many decimal % positions (imo mp bug) ss := asciistring(ss) ; % efficiency hack if expandafter unknown scantokens(ss) : ok := false ; % not yet defined elseif scantokens(ss) < 0 : ok := false ; % locally defined and undefined else : ok := true ; % globally already defined fi ; if not ok : if is_spot : flush_special(5, 8, s) ; elseif is_cmyk : flush_special(4, 8, s) ; else : flush_special(3, 7, s) ; fi ; scantokens(ss) := _special_counter_ ; _local_specials_ := _local_specials_ & "scantokens(" & ditto & ss & ditto & ") := -1 ;" ; fi ; % go ahead if is_spot : (_special_signal_/1000,5/1000,scantokens(ss)/1000) elseif is_cmyk : (_special_signal_/1000,4/1000,scantokens(ss)/1000) else : (_special_signal_/1000,3/1000,scantokens(ss)/1000) fi enddef ; %D This function returns true of false, dependent on transparency. vardef is_transparent(text t) = begingroup ; save transparent ; save _c_, _b_ ; vardef transparent(expr nn, tt, cc) = _b_ := true ; cc enddef ; boolean _b_ ; _b_ := false ; color _c_ ; _c_ := t ; _b_ endgroup enddef ; %D This function returns the not transparent color. vardef not_transparent(text t) = begingroup ; save transparent ; vardef transparent(expr nn, tt, cc) = cc enddef ; t endgroup enddef ; %D Basic position tracking: def register (expr label, width, height, offset) = begingroup ; flush_special(50, 7, ddecimal offset & " " & decimal width & " " & decimal height & " " & label) ; endgroup ; enddef ; %D We cannot scale cmyk colors directly since this spoils %D the trigger signal (such colors are no real colors). vardef scaledcmyk(expr c,m,y,k,sf) = cmyk(sf*c,sf*m,sf*y,sf*k) enddef ; vardef scaledcmykasrgb(expr c,m,y,k,sf) = (sf*(1-c-k,1-m-k,1-y-k)) enddef ; vardef scaledrgbascmyk(expr c,m,y,k,sf) = scaledcmyk(1-c,1-m,1-y,0,sf) enddef ; vardef scaledrgb(expr r,g,b,sf) = (sf*(r,g,b)) enddef ; vardef scaledgray(expr s,sf) = (sf*(s,s,s)) enddef ; % spotcolor is already scaled endinput ; % just an exercise (due to a question by Chof on the context mailing list); scaling of % 'special' colors is not possible and the next solution is incomplete (spot colors, % transparency, etc); watch the the tricky chained macro construction % vardef normalgray(expr s ) = (s,s,s) enddef ; % vardef normalrgb (expr r,g,b ) = (r,g,b) enddef ; % vardef normalcmyk(expr c,m,y,k) = if cmykcolors : save ok ; boolean ok ; if unknown cmykcolorhash[c][m][y][k] : ok := false ; elseif cmykcolorhash[c][m][y][k] = -1 : ok := false ; else : ok := true ; fi ; if not ok : save s ; string s ; s := dddecimal (c,m,y) & " " & decimal k ; _cmyk_counter_ := _cmyk_counter_ + 1 ; cmykcolorpattern[_cmyk_counter_/1000] := s ; cmykcolorhash[c][m][y][k] := _cmyk_counter_ ; flush_special(1, 7, decimal _cmyk_counter_ & " " & s) ; _local_specials_ := _local_specials_ & " cmykcolorhash[" & decimal c & "][" & decimal m & "][" & decimal y & "][" & decimal k & "] := -1 ; " ; fi ; (_special_signal_/1000,1/1000,cmykcolorhash[c][m][y][k]/1000) else : (1-c-k,1-m-k,1-y-k) fi enddef ; % vardef gray(expr s) = normalgray(s ) enddef ; % vardef rgb (expr r,g,b) = normalrgb (r,g,b ) enddef ; % vardef cmyk(expr c,m,y,k) = normalcmyk(c,m,y,k) enddef ; % numeric _scaled_color_t_ ; % color _scaled_color_c_ ; % def withscaledcolor = % hide ( % _scaled_color_t_ := 0 ; % direct % def gray(expr s) = % hide ( % _gray_s_ := s ; % _scaled_color_t_ := 1; % gray % ) % 0 % enddef ; % def rgb (expr r,g,b) = % hide ( % _rgb_r_ := r ; _rgb_g_ := g ; _rgb_b_ := b ; % _scaled_color_t_ := 2 ; % rgb % ) % 0 % enddef ; % def cmyk (expr c,m,y,k) = % hide ( % _cmyk_c_ := c ; _cmyk_m_ := m ; _cmyk_y_ := y ; _cmyk_k_ := k ; % _scaled_color_t_ := 3 ; % cmyk % ) % 0 % enddef ; ) % dowithscaledcolor % enddef ; % def dowithscaledcolor expr t = % hide ( % if color t : _scaled_color_c_ := t fi ; % vardef gray(expr s) = normalgray(s) enddef ; % vardef rgb (expr r,g,b) = normalrgb (r,g,b) enddef ; % vardef cmyk(expr c,m,y,k) = normalcmyk(c,m,y,k) enddef ; % ) % enddef ; % def by expr s = % if _scaled_color_t_ = 0 : % withcolor s*_scaled_color_c_ % elseif _scaled_color_t_ = 1 : % withcolor gray(s*_gray_s_) % elseif _scaled_color_t_ = 2 : % withcolor rgb (s*_rgb_r_, s*_rgb_g_, s*_rgb_b_) % elseif _scaled_color_t_ = 3 : % withcolor cmyk(s*_cmyk_c_, s*_cmyk_m_, s*_cmyk_y_, s*_cmyk_k_) % fi % enddef ; % fill fullcircle scaled 10cm withscaledcolor cmyk(0,0,1,0) by .5 ; % fill fullcircle scaled 8cm withscaledcolor rgb (0,0,1) by .5 ; % fill fullcircle scaled 6cm withscaledcolor gray(1) by .5 ; % fill fullcircle scaled 4cm withscaledcolor (0,1,0) by .5 ;
Thanks a lot Taco, I'm very happy not to have wasted sunday on an obvious mistake of mine. At least I found a bug. I think I've located the file where to cut and paste the updated version (there so many ...tex postfixed files on my mac in different distribution, I don't know why. And two version of context. Who knows.). Before doing anything, could you please tell me what it means:
Don't forget that you have to re-create the metafun format as well!
Ahemm, I'm in an abyss of ignorance... Best -a-
andrea valle wrote:
Thanks a lot Taco, I'm very happy not to have wasted sunday on an obvious mistake of mine. At least I found a bug.
You can think of finding your first conTeXt bug as a kind of initiation, you are now officially anm advanced user :-)
I think I've located the file where to cut and paste the updated version (there so many ...tex postfixed files on my mac in different distribution, I don't know why. And two version of context. Who knows.).
I'm not really the right person to answer this, being a linux user myself, but I can explain the general idea. There is a file named mp-spec.mp on your harddisk somewhere (and possibly more than one). That file is indirectly used for each metapost graphic in context, because it is one of the parts needed to create 'metafun.mem', that is then itself used by mpost while creating your graphic. You need to replace the mp-spec.mp file on your disc with the one from my e-mail message, then you have to re-generate the metafun.mem file (a user on unix would execute "texexec --make metafun" for that, while being root), and finally you have to re-compile your document with it's included graphic. I cannot give you the precise steps, sorry. I hope someone else can help you if my description is not adequate, esp. on how to decide which is the right mp-spec.mp , and how to re-create metafun.mem, making sure it ends up in the right place on your disk. Greetings, Taco
You can think of finding your first conTeXt bug as a kind of initiation, you are now officially anm advanced user :-)
It would be nicer if only I would be able to work it out :-)
There is a file named mp-spec.mp on your harddisk somewhere (and possibly more than one). That file is indirectly used for each metapost graphic in context, because it is one of the parts needed to create 'metafun.mem', that is then itself used by mpost while creating your graphic.
I think I found the mem file. In fact, I created a .mem file and when I substitute the old with the new one, nothing get printed out anymore. I'm getting really lost with all this unix-filename always repeating in different parts of the directory tree. I never know exactly what's happening. Any hint would be very appreciated -a-
andrea valle wrote:
I think I found the mem file. In fact, I created a .mem file and when I substitute the old with the new one, nothing get printed out anymore.
Do you mean this in a good or a bad way?
I'm getting really lost with all this unix-filename always repeating in different parts of the directory tree.
Most of us are familiar with this feeling, believe me. Taco
Bad way. I substituted all the m-spec.mp I was able to find (2).I runned a texexec, obtained a new metafun.mem, and substituted the old one. Something changed, in the sense that I have no output at all: a blank page. -a- On 8 Nov 2005, at 09:34, Taco Hoekwater wrote:
andrea valle wrote:
I think I found the mem file. In fact, I created a .mem file and when I substitute the old with the new one, nothing get printed out anymore.
Do you mean this in a good or a bad way?
I'm getting really lost with all this unix-filename always repeating in different parts of the directory tree.
Most of us are familiar with this feeling, believe me.
Taco _______________________________________________ ntg-context mailing list ntg-context@ntg.nl http://www.ntg.nl/mailman/listinfo/ntg-context
Andrea Valle DAMS - Facoltà di Scienze della Formazione Università degli Studi di Torino andrea.valle@unito.it
andrea valle wrote:
Bad way. I substituted all the m-spec.mp I was able to find (2).I runned a texexec, obtained a new metafun.mem, and substituted the old one. Something changed, in the sense that I have no output at all: a blank page.
That would/should have worked, as far as i can tell. I am sorry, I am stuck. Taco
Andrea, what kind of installation are you using? Gerben's i-installer or fink or have you installed tetex on your own? If you want to find out where different files reside, run this command: kpsewhere mp-spec.mp on my system (Gerben's architecture), I get: /usr/local/teTeX/share/texmf.local/metapost/context/base/mp-spec.mp /usr/local/teTeX/share/texmf.tetex/metapost/context/base/mp-spec.mp The first file is the one actually used by your TeX-installation, so this is the one you'd need to modify. Actually, "all this unix- filename always repeating" is there for a very good reason, so don't complain about it... Good luck Thomas On Nov 8, 2005, at 10:48 AM, andrea valle wrote:
I'm getting really lost with all this unix-filename always repeating in different parts of the directory tree.
Andrea,
what kind of installation are you using? Gerben's i-installer or fink or have you installed tetex on your own? If you want to find out where different files reside, run this command:
kpsewhere mp-spec.mp
on my system (Gerben's architecture), I get:
/usr/local/teTeX/share/texmf.local/metapost/context/base/mp-spec.mp /usr/local/teTeX/share/texmf.tetex/metapost/context/base/mp-spec.mp
The following comment is only valid for my redistribution: Normally, you would only have the second line (ConTeXt that is part of teTeX 3.0, which is roughly a year out of date). You only get the first line if you have installed my ConTeXt updater package. The texmf.local location is found before the texmf.tetex location, hence, installing the updater overrules what is in teTeX. Uninstalling the ConTeXt updater gets you back to the version in teTeX 3.0. You can set an automatic check for update for i-Packages in i-Installer after which you get a mail message telling you which packages have updated. The ConTeXt i-Package itself is updated automatically on the repositories if Hans releases a new version or a new beta. Which means that you'll hear about any ConTeXt update within 24 hours automatically. The ConTeXt updater is the only i-Package which is updated automatically on the repositories. G
Gerben, Thomas, I'm cleaning my hard disk (and my mind) and reinstalling. I changed the PATH (via bash tutorial...) and removed the sw tree. I installed thru i-installer, including the context updater. Now I have the two files mentioned by Thomas in their right place. Updating ConTeXt I have obtained: "You have just installed a ConTeXt installation 2005.10.27" So, I'm supposing that, as Hans has patched the bug-free version, I simply have to wait that automatically the new version is installable. Am I right? Thanks -a- PS:
kpsewhere mp-spec.mp
before it gave me: apples-Computer:~ apple$ kpsewhere mp-spec.mp /sw/share/texmf/metapost/context/base/mp-spec.mp Now it gives me nothing: apples-Computer:~ apple$ kpsewhere mp-spec.mp apples-Computer:~ apple$
on my system (Gerben's architecture), I get:
/usr/local/teTeX/share/texmf.local/metapost/context/base/mp-spec.mp /usr/local/teTeX/share/texmf.tetex/metapost/context/base/mp-spec.mp
The following comment is only valid for my redistribution:
Normally, you would only have the second line (ConTeXt that is part of teTeX 3.0, which is roughly a year out of date). You only get the first line if you have installed my ConTeXt updater package. The texmf.local location is found before the texmf.tetex location, hence, installing the updater overrules what is in teTeX. Uninstalling the ConTeXt updater gets you back to the version in teTeX 3.0.
You can set an automatic check for update for i-Packages in i-Installer after which you get a mail message telling you which packages have updated. The ConTeXt i-Package itself is updated automatically on the repositories if Hans releases a new version or a new beta. Which means that you'll hear about any ConTeXt update within 24 hours automatically. The ConTeXt updater is the only i-Package which is updated automatically on the repositories.
G
_______________________________________________ ntg-context mailing list ntg-context@ntg.nl http://www.ntg.nl/mailman/listinfo/ntg-context
Andrea Valle DAMS - Facoltà di Scienze della Formazione Università degli Studi di Torino andrea.valle@unito.it
andrea valle wrote:
Gerben, Thomas,
I'm cleaning my hard disk (and my mind) and reinstalling.
I changed the PATH (via bash tutorial...) and removed the sw tree. I installed thru i-installer, including the context updater.
Now I have the two files mentioned by Thomas in their right place.
Updating ConTeXt I have obtained: "You have just installed a ConTeXt installation 2005.10.27"
So, I'm supposing that, as Hans has patched the bug-free version, I simply have to wait that automatically the new version is installable. Am I right?
if you use gwtex, it will happen automatically since gerben keeps things pretty up to date Hans
On Nov 9, 2005, at 11:37, Hans Hagen wrote:
andrea valle wrote:
Gerben, Thomas,
I'm cleaning my hard disk (and my mind) and reinstalling.
I changed the PATH (via bash tutorial...) and removed the sw tree. I installed thru i-installer, including the context updater.
Now I have the two files mentioned by Thomas in their right place.
Updating ConTeXt I have obtained: "You have just installed a ConTeXt installation 2005.10.27"
So, I'm supposing that, as Hans has patched the bug-free version, I simply have to wait that automatically the new version is installable. Am I right?
if you use gwtex, it will happen automatically since gerben keeps things pretty up to date
*Almost* automatic. As soon as Hans updates either the beta or stable version of the ConTeXt zip files, the next night (CET) the ConTeXt Updtare i-Package is rebuilt and distributed to the repositories automatically. You still need to install it yourself (it is i-Installer's philosophy that no action will be performed unless a human accepts it, complete automation without user intervention is a big potential security risk) If you start i-Installer and go to the Preferences and you go to the "Other" tab you can set an automatic check for updates on i-Packages you have ("Check packages for updates..."). This creates a unix 'crontab' entry which is run at the time you have selected (e.g. every day a 07:00). This entry runs a helper program that checks the remote version of all your i-Packages and will report updated i-Packages via Mail. It won't automatically update & install (as per the philosophy). G
If you start i-Installer and go to the Preferences and you go to the "Other" tab you can set an automatic check for updates on i-Packages you have ("Check packages for updates...").
Already done.
This creates a unix 'crontab' entry which is run at the time you have selected (e.g. every day a 07:00). This entry runs a helper program that checks the remote version of all your i-Packages and will report updated i-Packages via Mail.
(Fine. How does this notification happen? Does he knows my email? Or am I asking something stupid?) Another obvious thing, but just to be sure. "Using gwtex" means for me typesetting from inside TeXShop with standard engine default. Right? Thanks a lot to all for the support So, know I'm waiting for updates -a-
If you start i-Installer and go to the Preferences and you go to the "Other" tab you can set an automatic check for updates on i-Packages you have ("Check packages for updates...").
Already done.
This creates a unix 'crontab' entry which is run at the time you have selected (e.g. every day a 07:00). This entry runs a helper program that checks the remote version of all your i-Packages and will report updated i-Packages via Mail.
(Fine. How does this notification happen? Does he knows my email? Or am I asking something stupid?)
The "crontab" tool takes care of this. You get mail as yourself on your system.
Another obvious thing, but just to be sure. "Using gwtex" means for me typesetting from inside TeXShop with standard engine default. Right?
Right (in 99.99% of the cases).
Thanks a lot to all for the support
So, know I'm waiting for updates
They are already there. Hoever, I just noticed that people talk about "newtexexec" and that is a command and if that is indeed a new command it will not be available after the updated ConTeXt updater install. G
On 10 Nov 2005, at 09:36, Gerben Wierda wrote:
The "crontab" tool takes care of this. You get mail as yourself on your system.
(Now I understand)
Right (in 99.99% of the cases).
Update, installed, configured. I'm happy. Yes, everything works fine as usual. Thanks to all. As usual, too. -a- Andrea Valle DAMS - Facoltà di Scienze della Formazione Università degli Studi di Torino andrea.valle@unito.it
(Taco, thanks for all the infos. Now, just trying to understand). Why the bug manifest itself in this particular tex file? I mean: I have made other pieces, both with far less either with far more included pdfs and everything was ok. What's special in the bug-raiser tex we discussed about? Thanks -a- On 7 Nov 2005, at 12:22, Taco Hoekwater wrote:
andrea valle wrote:
It's the same file but I commented out a part of the section dedicated to importing and drawing little notes attached to arrowlines between music fragments. Everything not commented is now displayed: on my machine it seems to be deterministic.
Ok, got it. There is a bug in mp-spec.mp (inside metafun). Because of a wrapping number, it fails to recognize the fact that there are embedded specials at all. Attached is an updated version of mp-spec.mp.
The corrected definition of add_special_signal is:
vardef add_special_signal = % write the version number if (length _global_specials_<>0) or (length _local_specials_ <>0) : special ("%%MetaPostSpecials: 1.0 " & decimal _special_signal_ ) ; fi ; enddef ;
It now tests for "not equal to zero" instead of "larger than zero": because of all the included files, the string _local_specials_ becomes longer than the maximum number 'length' can return, so it returns -32768 instead, and that is of course less than zero.
Don't forget that you have to re-create the metafun format as well!
Cheers, Taco %D \module %D [ file=mp-spec.mp, %D version=1999.6.26, %D title=\CONTEXT\ \METAPOST\ graphics, %D subtitle=special extensions, %D author=Hans Hagen, %D date=\currentdate, %D copyright={PRAGMA / Hans Hagen \& Ton Otten}] %C %C This module is part of the \CONTEXT\ macro||package and is %C therefore copyrighted by \PRAGMA. See licen-en.pdf for %C details.
% Spot colors are not handled by mptopdf !
% (r,g,b) => cmyk : r=123 g= 1 b=hash % => spot : r=123 g= 2 b=hash % => transparent rgb : r=123 g= 3 b=hash % => transparent cmyk : r=123 g= 4 b=hash % => transparent spot : r=123 g= 5 b=hash % => rest : r=123 g=n>10 b=whatever
%D This module is rather preliminary and subjected to %D changes. Here we closely cooperates with the \METAPOST\ %D to \PDF\ converter module built in \CONTEXT\ and provides %D for instance shading. More information can be found in %D type {supp-mpe.tex}.
if unknown context_tool : input mp-tool ; fi ; if known context_spec : endinput ; fi ;
boolean context_spec ; context_spec := true ;
numeric _special_counter_ ; _special_counter_ := 0 ; numeric _color_counter_ ; _color_counter_ := 11 ; % < 10 reserved numeric _special_signal_ ; _special_signal_ := 123 ;
%D When set to \type {true}, shading will be supported. Some %D day I will also write an additional directive.
boolean _inline_specials_ ; _inline_specials_ := false ;
%D Because we want to output only those specials that are %D actually used in a figure, we need a bit complicated %D bookkeeping and collection of specials. At the cost of some %D obscurity, we now have rather efficient resources.
string _global_specials_ ; _global_specials_ := "" ; string _local_specials_ ; _local_specials_ := "" ;
vardef add_special_signal = % write the version number if (length _global_specials_<>0) or (length _local_specials_ <>0) : special ("%%MetaPostSpecials: 1.0 " & decimal _special_signal_ ) ; fi ; enddef ;
vardef add_extra_specials = scantokens _global_specials_ ; scantokens _local_specials_ ; enddef ;
vardef reset_extra_specials = % only local ones _local_specials_ := "" ; enddef ;
boolean insidefigure ; insidefigure := false ;
% todo: alleen als special gebruikt flush
extra_beginfig := " insidefigure := true ; " & " reset_extra_specials ; " & extra_beginfig ;
extra_endfig := " add_special_signal ; " & extra_endfig & " add_extra_specials ; " & " reset_extra_specials ; " & " insidefigure := false ; " ;
def set_extra_special (expr s) = if insidefigure : _local_specials_ := _local_specials_ & s ; else : _global_specials_ := _global_specials_ & s ; fi enddef ;
def flush_special (expr typ, siz, dat) = _special_counter_ := _special_counter_ + 1 ; if _inline_specials_ : set_extra_special ( "special " & "(" & ditto & dat & " " & decimal _special_counter_ & " " & decimal typ & " " & decimal siz & " special" & ditto & ");" ) ; else : set_extra_special ( "special " & "(" & ditto & "%%MetaPostSpecial: " & decimal siz & " " & dat & " " & decimal _special_counter_ & " " & decimal typ & ditto & ");" ) ; fi ; enddef ;
%D The next hack is needed in case you use a version of %D \METAPOST\ that does not provide you the means to configure %D the buffer size. Patrick Gundlach suggested to use arrays %D in this case.
boolean bufferhack ; bufferhack := false ; % true ;
if bufferhack :
string _global_specials_[] ; numeric _nof_global_specials_ ; string _local_specials_[] ; numeric _nof_local_specials_ ;
_nof_global_specials_ := _nof_local_specials_ := 0 ;
vardef add_special_signal = % write the version number if (_nof_global_specials_>0) or (_nof_local_specials_>0) : special ("%%MetaPostSpecials: 1.0 " & decimal _special_signal_ ) ; fi ; enddef ;
vardef add_extra_specials = for i=1 upto _nof_global_specials_ : scantokens _global_specials_[i] ; endfor; for i=1 upto _nof_local_specials_ : scantokens _local_specials_[i] ; endfor; enddef ;
vardef reset_extra_specials = string _local_specials_[] ; _nof_local_specials_ := 0 ; enddef ;
def set_extra_special (expr s) = if insidefigure : _local_specials_[incr(_nof_local_specials_)] := s ; else : _global_specials_[incr(_nof_global_specials_)] := s ; fi enddef ;
fi ;
%D So far for this hack.
%D Shade allocation.
newinternal shadefactor ; shadefactor := 1 ;
pair shadeoffset ; shadeoffset := origin ;
vardef define_linear_shade (expr a, b, ca, cb) = flush_special(30, 15, "0 1 " & decimal shadefactor & " " & dddecimal ca & " " & ddecimal (a shifted shadeoffset) & " " & dddecimal cb & " " & ddecimal (b shifted shadeoffset) ) ; _special_counter_ enddef ;
vardef define_circular_shade (expr a, b, ra, rb, ca, cb) = flush_special(31, 17, "0 1 " & decimal shadefactor & " " & dddecimal ca & " " & ddecimal (a shifted shadeoffset) & " " & decimal ra & " " & dddecimal cb & " " & ddecimal (b shifted shadeoffset) & " " & decimal rb ) ; _special_counter_ enddef ;
%D A few predefined shading macros.
boolean trace_shades ; trace_shades := false ;
% if (n=1) : a := llcorner p ; b := urcorner p ; % elseif (n=2) : a := llcorner p ; b := ulcorner p ; % elseif (n=3) : a := lrcorner p ; b := ulcorner p ; % else : a := llcorner p ; b := lrcorner p ; % fi ;
def set_linear_vector (suffix a,b)(expr p,n) = if (n=1) : a := llcorner p ; b := urcorner p ; elseif (n=2) : a := lrcorner p ; b := ulcorner p ; elseif (n=3) : a := urcorner p ; b := llcorner p ; elseif (n=4) : a := ulcorner p ; b := lrcorner p ; elseif (n=5) : a := .5[ulcorner p,llcorner p] ; b := .5[urcorner p,lrcorner p] ; elseif (n=6) : a := .5[llcorner p,lrcorner p] ; b := .5[ulcorner p,urcorner p] ; elseif (n=7) : a := .5[lrcorner p,urcorner p] ; b := .5[llcorner p,ulcorner p] ; elseif (n=8) : a := .5[urcorner p,ulcorner p] ; b := .5[lrcorner p,llcorner p] ; else : a := .5[ulcorner p,llcorner p] ; b := .5[urcorner p,lrcorner p] ; fi ; enddef ;
def linear_shade (expr p, n, ca, cb) = begingroup ; save a, b, sh ; pair a, b ; set_linear_vector(a,b)(p,n) ; fill p withshade define_linear_shade (a,b,ca,cb) ; if trace_shades : drawarrow a -- b withpen pencircle scaled 1pt ; fi ; endgroup ; enddef ;
vardef predefined_linear_shade (expr p, n, ca, cb) = save a, b, sh ; pair a, b ; set_linear_vector(a,b)(p,n) ; set_shade_vector(a,b)(p,n) ; define_linear_shade (a,b,ca,cb) enddef ;
def set_circular_vector (suffix ab, r)(expr p,n) = if (n=1) : ab := llcorner p ; elseif (n=2) : ab := lrcorner p ; elseif (n=3) : ab := urcorner p ; elseif (n=4) : ab := ulcorner p ; else : ab := center p ; r := .5r ; fi ; enddef ;
def circular_shade (expr p, n, ca, cb) = begingroup ; save ab, r ; pair ab ; numeric r ; r := (xpart lrcorner p - xpart llcorner p) ++ (ypart urcorner p - ypart lrcorner p) ; set_circular_vector(ab,r)(p,n) ; fill p withshade define_circular_shade(ab,ab,0,r,ca,cb) ; if trace_shades : drawarrow ab -- ab shifted (0,r) withpen pencircle scaled 1pt ; fi ; endgroup ; enddef ;
vardef predefined_circular_shade (expr p, n, ca, cb) = save ab, r ; pair ab ; numeric r ; r := (xpart lrcorner p - xpart llcorner p) ++ (ypart urcorner p - ypart lrcorner p) ; set_circular_vector(ab,r)(p,n) ; define_circular_shade(ab,ab,0,r,ca,cb) enddef ;
%D Since a \type {fill p withshade s} syntax looks better %D than some macro, we implement a new primary.
primarydef p withshade sc = % == p withcolor shadecolor(sh) hide (_color_counter_ := _color_counter_ + 1) p withcolor (_special_signal_/1000,_color_counter_/1000,sc/1000) enddef ;
vardef shadecolor(expr sc) = hide (_color_counter_ := _color_counter_ + 1) (_special_signal_/1000,_color_counter_/1000,sc/1000) enddef ;
%D Figure inclusion.
%numeric cef ; cef := 0 ;
def externalfigure primary filename = doexternalfigure (filename) enddef ;
def doexternalfigure (expr filename) text transformation = begingroup ; save p, t ; picture p ; transform t ; p := nullpicture ; t := identity transformation ; flush_special(10, 9, dddecimal (xxpart t, yxpart t, xypart t) & " " & dddecimal (yypart t, xpart t, ypart t) & " " & filename) ; addto p contour unitsquare scaled 0 ; setbounds p to unitsquare transformed t ; _color_counter_ := _color_counter_ + 1 ; draw p withcolor (_special_signal_/1000,_color_counter_/1000,_special_counter_/1000) ; %draw p withcolor (_special_signal_/1000,cef/1000,_special_counter_/1000) ; endgroup ; enddef ;
%D Experimental:
%numeric currenthyperlink ; currenthyperlink := 0 ;
def hyperlink primary t = dohyperlink(t) enddef ; def hyperpath primary t = dohyperpath(t) enddef ;
def dohyperlink (expr destination) text transformation = begingroup ; save somepath ; path somepath ; somepath := fullsquare transformation ; dohyperpath(destination) somepath ; endgroup ; enddef ;
def dohyperpath (expr destination) expr somepath = begingroup ; flush_special(20, 7, ddecimal (xpart llcorner somepath, ypart llcorner somepath) & " " & ddecimal (xpart urcorner somepath, ypart urcorner somepath) & " " & destination) ; % currenthyperlink := currenthyperlink + 1 ; _color_counter_ := _color_counter_ + 1 ; fill boundingbox unitsquare scaled 0 withcolor
(_special_signal_/1000,_color_counter_/1000,_special_counter_/1000) ; % (_special_signal_/1000,currenthyperlink/1000,_special_counter_/1000) ; endgroup ; enddef ;
% \setupinteraction[state=start] % \setupcolors [state=start] % % Hello There! \blank % % \startMPcode % pickup pencircle scaled 5 ; % draw fullcircle scaled 4cm withcolor red ; % hyperpath "nextpage" boundingbox currentpicture ; % draw origin withcolor blue ; % \stopMPcode % % \blank Does it work or not? % % \startMPcode % pickup pencircle scaled 5 ; % draw fullcircle scaled 4cm withcolor red ; % hyperpath "nextpage" fullcircle scaled 4cm ; % draw origin withcolor blue ; % draw fullcircle scaled 4cm shifted (1cm,1cm); % \stopMPcode % % \blank Does it work or not? \page Hello There! \blank % % \startMPcode % pickup pencircle scaled 5 ; % draw fullcircle scaled 2cm shifted (-2cm,-1cm) ; % draw fullcircle scaled 3cm shifted (2cm,1cm) withcolor red ; % draw fullcircle scaled 1cm ; % hyperlink "previouspage" scaled 3cm shifted (2cm,1cm) ; % draw origin withcolor blue ; % \stopMPcode % % \blank Does it work or not?
_cmyk_counter_ := 0 ;
extra_endfig := " resetcmykcolors ; " & extra_endfig ;
def resetcmykcolors = numeric cmykcolorhash[][][][] ; enddef ;
resetcmykcolors ; boolean cmykcolors ; cmykcolors := false ; % true
string cmykcolorpattern[] ; % needed for transparancies
vardef cmyk(expr c,m,y,k) = if cmykcolors : save ok ; boolean ok ; if unknown cmykcolorhash[c][m][y][k] : ok := false ; % not yet defined elseif cmykcolorhash[c][m][y][k] = -1 : ok := false ; % locally defined and undefined else : ok := true ; % globally already defined fi ; if not ok : save s ; string s ; s := dddecimal (c,m,y) & " " & decimal k ; _cmyk_counter_ := _cmyk_counter_ + 1 ; cmykcolorpattern[_cmyk_counter_/1000] := s ; cmykcolorhash[c][m][y][k] := _cmyk_counter_ ; flush_special(1, 7, decimal _cmyk_counter_ & " " & s) ; _local_specials_ := _local_specials_ & " cmykcolorhash[" & decimal c & "][" & decimal m & "][" & decimal y & "][" & decimal k & "] := -1 ; " ; fi ; (_special_signal_/1000,1/1000,cmykcolorhash[c][m][y][k]/1000) else : (1-c-k,1-m-k,1-y-k) fi enddef ;
% newcolor truecyan, truemagenta, trueyellow ; % % truecyan = cmyk (1,0,0,0) ; % truemagenta = cmyk (0,1,0,0) ; % trueyellow = cmyk (0,0,1,0) ;
%D Spot colors
_spotcolor_counter_ := 0 ; _spotcolor_number_ := 0 ;
extra_endfig := " resetspotcolors ; " & extra_endfig ;
def resetspotcolors = numeric spotcolorhash[][] ; enddef ;
resetspotcolors ; boolean spotcolors ; spotcolors := false ; % true
string spotcolorpattern[] ; % needed for transparancies
% vardef spotcolor(expr p, s) = % if spotcolors : % save ok, pc_tag ; boolean ok ; string pc_tag ; % pc_tag := "_pct_"&p ; % if not unstringed(pc_tag) : % _spotcolor_number_ := _spotcolor_number_ + 1 ; % setunstringed(pc_tag,_spotcolor_number_) ; % fi ; % pp := getunstringed(pc_tag) ; % if unknown spotcolorhash[pp][s] : % ok := false ; % not yet defined % elseif spotcolorhash[pp][s] = -1 : % ok := false ; % locally defined and undefined % else : % ok := true ; % globally already defined % fi ; % if not ok : % save ss ; string ss ; ss := p & " " & decimal s ; % _spotcolor_counter_ := _spotcolor_counter_ + 1 ; % spotcolorpattern[_spotcolor_counter_/1000] := ss ; % spotcolorhash[pp][s] := _spotcolor_counter_ ; % flush_special(2, 5, decimal _spotcolor_counter_ & " " & ss) ; % _local_specials_ := _local_specials_ & % "spotcolorhash["&decimal pp&"]["&decimal s&"]:=-1;" ; % fi ; % (_special_signal_/1000,2/1000,spotcolorhash[pp][s]/1000) % else : % (1-s,1-s,1-s) % fi % enddef ;
% vardef spotcolor(expr p, s) = % if spotcolors : % save ok, pc_tag ; boolean ok ; string pc_tag ; % pc_tag := "_pct_"&p ; % if not unstringed(pc_tag) : % _spotcolor_number_ := _spotcolor_number_ + 1 ; % setunstringed(pc_tag,_spotcolor_number_) ; % fi ; % pp := getunstringed(pc_tag) ; % if unknown spotcolorhash[pp][s] : % ok := false ; % not yet defined % elseif spotcolorhash[pp][s] = -1 : % ok := false ; % locally defined and undefined % else : % ok := true ; % globally already defined % fi ; % if not ok : % save ss ; string ss ; ss := p & " " & decimal s ; % _spotcolor_counter_ := _spotcolor_counter_ + 1 ; % spotcolorpattern[_spotcolor_counter_/1000] := ss ; % spotcolorhash[pp][s] := _spotcolor_counter_ ; % flush_special(2, 5, decimal _spotcolor_counter_ & " " & ss) ; % _local_specials_ := _local_specials_ & % "spotcolorhash["&decimal pp&"]["&decimal s&"]:=-1;" ; % fi ; % (_special_signal_/1000,2/1000,spotcolorhash[pp][s]/1000) % else : % (1-s,1-s,1-s) % fi % enddef ;
vardef spotcolor(expr p, s) = multitonecolor(p, 1, "", decimal s) enddef ;
vardef multitonecolor(expr n, f, d, p) = % name fractions names factors if spotcolors : save ok, pc_tag ; boolean ok ; string pc_tag ; pc_tag := "_pct_" & n ; if not unstringed(pc_tag) : _spotcolor_number_ := _spotcolor_number_ + 1 ; setunstringed(pc_tag,_spotcolor_number_) ; fi ; pp := getunstringed(pc_tag) ; pc_tag := "_pct_"& decimal f & "_" & if d = "" : n else : d fi & "_" & p ; % check for d empty if not unstringed(pc_tag) : _spotcolor_number_ := _spotcolor_number_ + 1 ; setunstringed(pc_tag,_spotcolor_number_) ; fi ; ps := getunstringed(pc_tag) ; if unknown spotcolorhash[pp][ps] : ok := false ; % not yet defined elseif spotcolorhash[pp][ps] = -1 : ok := false ; % locally defined and undefined else : ok := true ; % globally already defined fi ; if not ok : save ss ; string ss ; ss := n & " " & decimal f & " " & if d = "" : n else : d fi & " " & p ; _spotcolor_counter_ := _spotcolor_counter_ + 1 ; spotcolorpattern[_spotcolor_counter_/1000] := ss ; spotcolorhash[pp][ps] := _spotcolor_counter_ ; flush_special(2, 7, decimal _spotcolor_counter_ & " " & ss) ; _local_specials_ := _local_specials_ & "spotcolorhash["&decimal pp&"]["&decimal ps&"]:=-1;" ; fi ; (_special_signal_/1000,2/1000,spotcolorhash[pp][ps]/1000) else : .5white fi enddef ;
%D Transparency
normaltransparent := 1 ; multiplytransparent := 2 ; screentransparent := 3 ; overlaytransparent := 4 ; softlighttransparent := 5 ; hardlighttransparent := 6 ; colordodgetransparent := 7 ; colorburntransparent := 8 ; darkentransparent := 9 ; lightentransparent := 10 ; differencetransparent := 11 ; exclusiontransparent := 12 ;
% nottransparent := 0 ; % compatibletransparent := 99 ;
% fill fullcircle scaled 10cm withcolor transparant(.8,3,color) ;
% vardef transparent(expr n, t, c) = % save s, ss, nn, cc, is_cmyk, is_spot, ok ; % string s, ss ; numeric nn ; color cc ; boolean is_cmyk, is_spot, ok ; % % transparancy type % if string n : % if expandafter known scantokens(n&"transparent") : % nn := scantokens(n&"transparent") ; % else : % nn := 0 ; % fi % else : % nn := min(n,13) % nn := if n<13 : n else : nn := 0 fi ; % fi ; % % we need to expand the color (can be cmyk(..) or predefined) % cc := c ; % expand color % % check for cmyk special % is_cmyk := (redpart cc = _special_signal_/1000) % and (greenpart cc = 1/1000) ; % is_spot := (redpart cc = _special_signal_/1000) % and (greenpart cc = 2/1000) ; % % build special string, fetch cmyk components % s := decimal nn & " " & decimal t & " " & % if is_cmyk : cmykcolorpattern[bluepart cc] % elseif is_spot : spotcolorpattern[bluepart cc] % else : dddecimal cc fi ; % % check if this one is already used % ss := "tr_" & s ; % % efficiency hack % if expandafter unknown scantokens(ss) : % ok := false ; % not yet defined % elseif scantokens(ss) < 0 : % ok := false ; % locally defined and undefined % else : % ok := true ; % globally already defined % fi ; % if not ok : % if is_spot : % flush_special(5, 6, s) ; % elseif is_cmyk : % flush_special(4, 8, s) ; % else : % flush_special(3, 7, s) ; % fi ; % scantokens(ss) := _special_counter_ ; % _local_specials_ := _local_specials_ & % "scantokens(" & ditto & ss & ditto & ") := -1 ;" ; % fi ; % % go ahead % if is_spot : % (_special_signal_/1000,5/1000,scantokens(ss)/1000) % elseif is_cmyk : % (_special_signal_/1000,4/1000,scantokens(ss)/1000) % else : % (_special_signal_/1000,3/1000,scantokens(ss)/1000) % fi % enddef ;
vardef transparent(expr n, t, c) = save s, ss, nn, cc, is_cmyk, is_spot, ok ; string s, ss ; numeric nn ; color cc ; boolean is_cmyk, is_spot, ok ; % transparancy type if string n : if expandafter known scantokens(n&"transparent") : nn := scantokens(n&"transparent") ; else : nn := 0 ; fi else : % nn := min(n,13) nn := if n<13 : n else : nn := 0 fi ; fi ; % we need to expand the color (can be cmyk(..) or predefined) cc := c ; % expand color % check for cmyk special is_cmyk := (redpart cc = _special_signal_/1000) and (greenpart cc = 1/1000) ; is_spot := (redpart cc = _special_signal_/1000) and (greenpart cc = 2/1000) ; % build special string, fetch cmyk components s := decimal nn & " " & decimal t & " " & if is_cmyk : cmykcolorpattern[bluepart cc] elseif is_spot : spotcolorpattern[bluepart cc] else : dddecimal cc fi ; % check if this one is already used ss := cleanstring("tr_" & s) ; % we now have rather unique names, i.e. a color spec of .234 becomes % tr..._234.... and metapost gives a number overflow (parse error) % for variables like tr_12345678 which may result from many decimal % positions (imo mp bug) ss := asciistring(ss) ; % efficiency hack if expandafter unknown scantokens(ss) : ok := false ; % not yet defined elseif scantokens(ss) < 0 : ok := false ; % locally defined and undefined else : ok := true ; % globally already defined fi ; if not ok : if is_spot : flush_special(5, 8, s) ; elseif is_cmyk : flush_special(4, 8, s) ; else : flush_special(3, 7, s) ; fi ; scantokens(ss) := _special_counter_ ; _local_specials_ := _local_specials_ & "scantokens(" & ditto & ss & ditto & ") := -1 ;" ; fi ; % go ahead if is_spot : (_special_signal_/1000,5/1000,scantokens(ss)/1000) elseif is_cmyk : (_special_signal_/1000,4/1000,scantokens(ss)/1000) else : (_special_signal_/1000,3/1000,scantokens(ss)/1000) fi enddef ;
%D This function returns true of false, dependent on transparency.
vardef is_transparent(text t) = begingroup ; save transparent ; save _c_, _b_ ; vardef transparent(expr nn, tt, cc) = _b_ := true ; cc enddef ; boolean _b_ ; _b_ := false ; color _c_ ; _c_ := t ; _b_ endgroup enddef ;
%D This function returns the not transparent color.
vardef not_transparent(text t) = begingroup ; save transparent ; vardef transparent(expr nn, tt, cc) = cc enddef ; t endgroup enddef ;
%D Basic position tracking:
def register (expr label, width, height, offset) = begingroup ; flush_special(50, 7, ddecimal offset & " " & decimal width & " " & decimal height & " " & label) ; endgroup ; enddef ;
%D We cannot scale cmyk colors directly since this spoils %D the trigger signal (such colors are no real colors).
vardef scaledcmyk(expr c,m,y,k,sf) = cmyk(sf*c,sf*m,sf*y,sf*k) enddef ;
vardef scaledcmykasrgb(expr c,m,y,k,sf) = (sf*(1-c-k,1-m-k,1-y-k)) enddef ;
vardef scaledrgbascmyk(expr c,m,y,k,sf) = scaledcmyk(1-c,1-m,1-y,0,sf) enddef ;
vardef scaledrgb(expr r,g,b,sf) = (sf*(r,g,b)) enddef ;
vardef scaledgray(expr s,sf) = (sf*(s,s,s)) enddef ;
% spotcolor is already scaled
endinput ;
% just an exercise (due to a question by Chof on the context mailing list); scaling of % 'special' colors is not possible and the next solution is incomplete (spot colors, % transparency, etc); watch the the tricky chained macro construction
% vardef normalgray(expr s ) = (s,s,s) enddef ; % vardef normalrgb (expr r,g,b ) = (r,g,b) enddef ; % vardef normalcmyk(expr c,m,y,k) = if cmykcolors : save ok ; boolean ok ; if unknown cmykcolorhash[c][m][y][k] : ok := false ; elseif cmykcolorhash[c][m][y][k] = -1 : ok := false ; else : ok := true ; fi ; if not ok : save s ; string s ; s := dddecimal (c,m,y) & " " & decimal k ; _cmyk_counter_ := _cmyk_counter_ + 1 ; cmykcolorpattern[_cmyk_counter_/1000] := s ; cmykcolorhash[c][m][y][k] := _cmyk_counter_ ; flush_special(1, 7, decimal _cmyk_counter_ & " " & s) ; _local_specials_ := _local_specials_ & " cmykcolorhash[" & decimal c & "][" & decimal m & "][" & decimal y & "][" & decimal k & "] := -1 ; " ; fi ; (_special_signal_/1000,1/1000,cmykcolorhash[c][m][y][k]/1000) else : (1-c-k,1-m-k,1-y-k) fi enddef ;
% vardef gray(expr s) = normalgray(s ) enddef ; % vardef rgb (expr r,g,b) = normalrgb (r,g,b ) enddef ; % vardef cmyk(expr c,m,y,k) = normalcmyk(c,m,y,k) enddef ;
% numeric _scaled_color_t_ ; % color _scaled_color_c_ ;
% def withscaledcolor = % hide ( % _scaled_color_t_ := 0 ; % direct % def gray(expr s) = % hide ( % _gray_s_ := s ; % _scaled_color_t_ := 1; % gray % ) % 0 % enddef ; % def rgb (expr r,g,b) = % hide ( % _rgb_r_ := r ; _rgb_g_ := g ; _rgb_b_ := b ; % _scaled_color_t_ := 2 ; % rgb % ) % 0 % enddef ; % def cmyk (expr c,m,y,k) = % hide ( % _cmyk_c_ := c ; _cmyk_m_ := m ; _cmyk_y_ := y ; _cmyk_k_ := k ; % _scaled_color_t_ := 3 ; % cmyk % ) % 0 % enddef ; ) % dowithscaledcolor % enddef ;
% def dowithscaledcolor expr t = % hide ( % if color t : _scaled_color_c_ := t fi ; % vardef gray(expr s) = normalgray(s) enddef ; % vardef rgb (expr r,g,b) = normalrgb (r,g,b) enddef ; % vardef cmyk(expr c,m,y,k) = normalcmyk(c,m,y,k) enddef ; % ) % enddef ;
% def by expr s = % if _scaled_color_t_ = 0 : % withcolor s*_scaled_color_c_ % elseif _scaled_color_t_ = 1 : % withcolor gray(s*_gray_s_) % elseif _scaled_color_t_ = 2 : % withcolor rgb (s*_rgb_r_, s*_rgb_g_, s*_rgb_b_) % elseif _scaled_color_t_ = 3 : % withcolor cmyk(s*_cmyk_c_, s*_cmyk_m_, s*_cmyk_y_, s*_cmyk_k_) % fi % enddef ;
% fill fullcircle scaled 10cm withscaledcolor cmyk(0,0,1,0) by .5 ; % fill fullcircle scaled 8cm withscaledcolor rgb (0,0,1) by .5 ; % fill fullcircle scaled 6cm withscaledcolor gray(1) by .5 ; % fill fullcircle scaled 4cm withscaledcolor (0,1,0) by .5 ; _______________________________________________ ntg-context mailing list ntg-context@ntg.nl http://www.ntg.nl/mailman/listinfo/ntg-context
Andrea Valle DAMS - Facoltà di Scienze della Formazione Università degli Studi di Torino andrea.valle@unito.it
andrea valle wrote:
(Taco, thanks for all the infos. Now, just trying to understand).
Why the bug manifest itself in this particular tex file? I mean: I have made other pieces, both with far less either with far more included pdfs and everything was ok. What's special in the bug-raiser tex we discussed about?
For every pdf you include, about 80 bytes is added to an internal string. There is a macro that checks the length of this string, and it sets a special "there are included pdfs" flag when the strings length is larger than zero. However, the primitive metapost operation to get the length of a string is rather stupid, it wraps without ever checking anything. Just run this bit of metapost code and watch the terminal output: string s; s := ""; forever: s := s & " "; message decimal length s; endfor; end. (you'll have to kill the process manually) So, my patch checks for 'not zero', which has a far smaller chance of being incorrect. Not completely nil, unfortunately. 'length' can actually wrap back to being zero, I'll have to fix that in metapost itself. Cheers, Taco
Taco Hoekwater wrote:
andrea valle wrote:
It's the same file but I commented out a part of the section dedicated to importing and drawing little notes attached to arrowlines between music fragments. Everything not commented is now displayed: on my machine it seems to be deterministic.
Ok, got it. There is a bug in mp-spec.mp (inside metafun). Because of a wrapping number, it fails to recognize the fact that there are embedded specials at all. Attached is an updated version of mp-spec.mp.
The corrected definition of add_special_signal is:
vardef add_special_signal = % write the version number if (length _global_specials_<>0) or (length _local_specials_ <>0) : special ("%%MetaPostSpecials: 1.0 " & decimal _special_signal_ ) ; fi ; enddef ;
It now tests for "not equal to zero" instead of "larger than zero": because of all the included files, the string _local_specials_ becomes longer than the maximum number 'length' can return, so it returns -32768 instead, and that is of course less than zero.
ok, patched. i guess that only you could uncover this wrap-around bug -) (btw, another funny mp issue is in opening a file for a second time (probably due to the non explicit close)) Hans
participants (5)
-
andrea valle -
Gerben Wierda -
Hans Hagen -
Taco Hoekwater -
Thomas A. Schmitz