On 1/29/2023 5:03 PM, Yaroslav Beltukov via ntg-context wrote:
> Dear Hans and all contributors,
>
> I really appreciate ConTeXt for the right way to obtain high quality
> documents. I'm a theoretical physicist and I'm going to write a book. I
> think ConTeXt is the right choice to work with a number of formulas,
> figures and cross-references. The visual quality of formulas is better
> than in regular LaTeX. The new feature with formula autosplitting looks
> also very promising. For me it is important to obtain the high quality
> without a lot of manual tweaks of each formula for each given document
> format and figure placement.
>
> However, I have found a problem with references if I compile one
> component only. The references to formulas in other components have
> wrong prefixes, e.g. (2.1) instead of (3.1). Needless to say, the right
> formula references are very important.
>
> I started looking into this issue. The references to other components
> are taken from the whole product as from an external document. It turned
> out that this is a general problem with references with prefixes to
> external documents. The prefix is stored in a tuc file as a reference to
> a section as a sequential number of the header in the document. As a
> result, the prefix from the external document is calculated using the
> structure of the current document.
>
> Here is the MWE, which consists of two files:
> foo.tex:
>
> \defineenumeration[remark][prefix=yes, prefixsegments=chapter:section]
>
> \starttext
>
> Equations: \in[eq1], \in[eq2], \in[eq3], \in[eq4]
>
> Sections: \in[sec1], \in[sec2], \in[sec3], \in[sec4]
>
> Chapters: \in[chap1], \in[chap2], \in[chap3], \in[chap4]
>
> Remarks: \in[remark1], \in[remark2]
>
> \startbodymatter
>
> \chapter[chap1]{Chapter}
> \placeformula[eq1]\startformula x = y\stopformula
> \chapter[chap2]{Chapter}
> \section[sec1]{Section}
> \section[sec2]{Section}
> \placeformula[eq1]\startformula x = y\stopformula
> \placeformula[eq2]\startformula x = y\stopformula
>
> \stopbodymatter
>
> \startappendices
>
> \chapter[chap3]{Chapter}
> \section[sec3]{Section}
> \placeformula[eq3]\startformula x = y\stopformula
> \section[sec4]{Section}
> \startremark[remark1]\stopremark
> \placeformula[eq4]\startformula x = y\stopformula
> \startremark[remark2]\stopremark
> \chapter[chap4]{Chapter}
>
> \stopappendices
>
> \stoptext
>
>
> bar.tex:
>
> \starttext
>
> Equations: \in[foo::eq1], \in[foo::eq2], \in[foo::eq3], \in[foo::eq4]
>
> Sections: \in[foo::sec1], \in[foo::sec2], \in[foo::sec3], \in[foo::sec4]
>
> Chapters: \in[foo::chap1], \in[foo::chap2], \in[foo::chap3], \in[foo::chap4]
>
> Remarks: \in[foo::remark1], \in[foo::remark2]
>
> % any chapters and sections here
>
> \stoptext
>
> It is expected to have the same first page on these documents:
>
> Equations: 1.1, 2.2, A.1, A.2
> Sections: 2.1, 2.2, A.1, A.2
> Chapters: 1, 2, A, B
> Remarks: A.2.1, A.2.2
>
> However, the bar.tex produces wrong prefixes to formulas and remarks.
> The output depends on the document structure of bar.tex, not foo.tex.
>
> I started looking into the source code. Thanks to lua, it is not a big
> deal to track the problem. The prefixdata is complemented by the
> sectiondata after the loading the tuc file. So, the question is: is it
> possible to store the full prefixdata with all necessary prefix numbers
> in the tuc file? Here is my proposal to change the source code:
>
> --- strc-lst-old.lmt    2023-01-29 11:30:15.610309948 +0300
> +++ strc-lst.lmt        2023-01-29 12:10:08.864228923 +0300
> @@ -266,6 +266,16 @@
>       if r and not r.section then
>           r.section = structures.sections.currentid()
>       end
> +    -- store sectiondata in prefixdata (necessary for external files)
> +    if t.prefixdata and r.section then
> +        local sectiondata = structures.sections.collected[r.section]
> +        if sectiondata then
> +            for k, v in next, sectiondata do
> +                t.prefixdata[k] = v
> +            end
> +        end
> +    end
> +    --
>       local b = r and t.block
>       if r and not b then
>           local s = r.section
>
> --- strc-ref-old.lmt    2023-01-29 11:30:15.823643904 +0300
> +++ strc-ref.lmt        2023-01-29 12:07:45.697109862 +0300
> @@ -2318,7 +2318,17 @@
>       if data then
>           numberdata = lists.reordered(data) -- data.numberdata
>           if numberdata then
> -            helpers.prefix(data,prefixspec)
> +            -- helpers.prefix(data,prefixspec)
> +            -- use the actual numbers from prefixdata
> +            local prefixdata = data.prefixdata
> +            if prefixdata then
> +                -- adapted from helpers.prefix (not sure)
> +                if (prefixspec and prefixspec == no) or
> prefixdata.prefix == no then
> +                    prefixdata = false
> +                end
> +                sections.typesetnumber(prefixdata,"prefix", prefixspec
> or false, prefixdata)
> +            end
> +            --
>             
>   sections.typesetnumber(numberdata,"number",numberspec,numberdata)
>           else
>               local useddata = data.useddata
>
> After this small change, all the references are correct. However, I'm
> quite new to ConTeXt, so maybe here are some caveats. It would be great
> to fix the references to formulas, especially for the forthcoming
> document about maths.
that will bloat the already large tuc file, so here is an alternative (i
bet you can find where to patch):

strc-ref

         local external = struc.references.collected -- direct references
         local lists    = struc.lists.collected      -- indirect
references (derived)
         local pages    = struc.pages.collected      -- pagenumber data
local sections = struc.sections.collected

and

                         if prefix == "" then
                             prefix = name -- this can clash!
                         end
local section = references.section
if section then
     if sections then
         references.sectiondata = sections[section]
     else
         -- warning
     end
end

strc-pag:

     local sectiondata = references.sectiondata or
sections.collected[references.section]
     if not sectiondata then
         return entry, false, "no section data"
     end

there might be a few more such places where we can bind to the loaded data

Hans

-----------------------------------------------------------------
                                           Hans Hagen | PRAGMA ADE
               Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
        tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl
-----------------------------------------------------------------

___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / https://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : https://www.pragma-ade.nl / http://context.aanhet.net
archive  : https://bitbucket.org/phg/context-mirror/commits/
wiki     : https://contextgarden.net
___________________________________________________________________________________