General TODOs/Issues

[Jazzpirate]

• flexary notations, especially with \symvariant show discrepancies between the argument markers/names occuring in <prototype> and <rendering>, causing ambiguity and problems in the omdoc-importer
• smglom/sets/words.tex uses \text{#1} which causes errors. Not clear at what point (stex/latexml/tex-file) that needs fixing.
• In general, nested math environments (e.g. $\text{$p$}$) yield <OMS name="verbalize"/>, which doesn't exist. They're surprisingly common in smglom, and in seemingly all cases unnecessarily so. Example, smglom/sets/image.en.tex: \text{$\inset{\pair{a}{b}}f$ for some $\inset{a}{\primvar{A}}$}
• structures are being used heavily in smglom/algebra, but references to constants from structures are never prefixed/disambiguated... unclear how to proceed here.
• many smglom files contain things like \defi[name=foo]{foo}{...} followed by a \symi{foo}, which makes the omdoc importer attempt to introduce the same constant twice. I don't know what to do about that.
• mutliple separate terms in the same math environment seem to cause problems, e.g. in smglom/complexity/landau-sets.en - $\atmost{g},\atleast{g},\sameas{g}$ yields a single <OMOBJ> which the omdoc-importer attempts to parse as a single term, which will fail. Fixing that in the omdoc importer would unfortunately require quite a lot of refactoring, but
  1. $\atmost{g}$,$\atleast{g}$,$\sameas{g}$ solves the problem quite simply and
  2. I'm almost sure that shouldn't be a single <OMOBJ> in the first place...

[Jazzpirate]

on that note - \text{$...$ and $...$} occurs quite frequently and does not look very stex-y to me. I'm almost sure there's an and-operator somewhere in mv that should probably be used instead? ;)

[Jazzpirate]

More issues:

• fromrepos and torepos in gviewnl are apparently not properly passed down to mhview. The resulting omdoc contains no references to repositories: <omdoc:view from="magma" to="natarith" type="example" xml:id="natp-magma.en"> (from smglom/algebra/natp-magma.en.tex). Apparently mhview only uses those to do an \importMHmoduleI, but they should probably be stored as attributes in the resulting omdoc:view-node
• mhinlineView is currently translated as <ltx:inline-para> instead of anything semantic. I think there's a separate issue for that around somewhere

[Jazzpirate]

• \fassign currently has only a dummy implementation

[Jazzpirate]

• gstructure or structureenv does not resolve smglom-repository names; the optional argument seems to be packed (in the bindings for structureenv) in a "load" option, but fails to resolve properly and I'm not sure how the "load" thing works.
  Example: smglom/linear-algebra/vectorspace.tex has \begin{gstructure}[smglom/algebra]{group}{group}, but the resulting omdoc is <omdoc:structure from="?group" name="group">, missing the smglom/algebra information.