Re: Extracting footnote text

[holmgrn@xxxxxxxx]

among other things, translates every ≪FN1Footnote text.≫ into > \footnote{Footnote text.}. The technique the program uses is to search > for ≪FN1 in expanded mode, locating the start of the footnote, then to go > to draft mode and move the cursor right, and then to go back to expanded > mode, where there cursor is now at the end of the footnote. A } can now > be inserted, and one can search back for ≪FN1 and insert the \footnote{. > > Unfortunately a bug in version 4 causes the position of the cursor to be > incorrect under some circumstances. (I reported the bug to TTG a few months > ago, and they promised to look at it, but subsequent written enquiries > about its status have failed to elicit any response---K. Frank, are you > listening?) > ... > Martin J. Osborne At many moments in its history (not just v4), XyWrite has been unreliable about locating an exact Character_Position, especially with long JuMPs across 64K segment boundaries. Pure Lower Paleolithic stuff, & shouldn't still be a problem, but ... it is nonetheless. In _any_ critical program you need to verify that the current position is identical to the desired and provide a loop to force repetitive JuMP(s) if necessary, e.g. thus: (DECODE this with frame DECODE [part of XPLCODE] in REORGNIZ.U2 [part of REORGNZ2.ZIP]) XPLeNCODE v1.8 begin LOOPJMP.WRK ...{174}SX01,{174}CP{175}{175}...{174}LBA{175}[255+130+171] jmp {174}PV01{175}[255+130+127]{174}IF{174}CP{175}<> {174}PV01{175}{175}{174}GLA{175}{174}EI{175}...[13+10] end XPLeNCODE 99.9% of the time this solves (==masks) the problem. Once in a great while you get caught in an endless loop, in which case you have to BreaK out and cross on foot. ------------ Robert J. Holmgren holmgrn@xxxxxxxx ------------