On 09/11/2012 08:18 PM, Hans Hagen wrote:
If the number of bib items is not that large hashing does not save much. (In x-xfdf.mkiv you can see another example).
\starttext
\startbuffer[demo] <document> <bibitem xml:id="hagen2011"> <author>Hans Hagen</author> <title>The Meaning of Life Is 42</title> <journal>The Rhetoric of Mathematics</journal> <date>2011</date> </bibitem> <bibitem xml:id="hagen2012"> <author>Hans Hagen</author> <title>The New Meaning is Now 43</title> <journal>High Energy Metaphysics</journal> <date>2012</date> </bibitem> <text> <p>This has been successfully proven in papers by <bib ref="#hagen2011"/> and <bib ref="#hagen2012"/>.</p> </text> </document> \stopbuffer
\startxmlsetups xml:define \xmlsetsetup{#1}{*}{xml:*} \stopxmlsetups
\startxmlsetups xml:document \xmlall{#1}{/text} \stopxmlsetups \startxmlsetups xml:text \xmlflush{#1} \stopxmlsetups \startxmlsetups xml:p \xmlflush{#1} \par \stopxmlsetups
\startxmlsetups xml:bib \BibAuthor{#1}{\xmlatt{#1}{ref}} \stopxmlsetups
\startxmlsetups xml:author [[\xmlflush{#1}]] \stopxmlsetups
\def\BibAuthor#1#2{[1:\xmlfirst{whatever}{/document/bibitem[@xml:id==string.sub('#2',2)]/author}]}
\xmlregistersetup{xml:define}
\xmlprocessbuffer{whatever}{demo}{}
\startluacode userdata = userdata or { } userdata.bibitems = { } \stopluacode
\startxmlsetups xml:document \xmlflush{#1} \stopxmlsetups
\startxmlsetups xml:bibitem \ctxlua{userdata.bibitems["\letterhash\xmlatt{#1}{xml:id}"] = "#1"} \stopxmlsetups
\def\BibAuthor#1#2{[2:\xmlfirst{\cldcontext{userdata.bibitems['#2']}}{/author}]}
\xmlprocessbuffer{whatever}{demo}{}
\stoptext
(I ran into a recently introduced side effect of an optimization so I'll generate you a new beta. Older versions should work anyway as no special trickery is used.)
Hans
Ah, wonderful, Hans, I can play with this! Thanks a lot for the swift reply and for the help, this will get me going! All best Thomas