Unicode characters in mathed

[Jean-Marc Lasgouttes]

Le 03/04/2023 à 00:36, Enrico Forestieri a écrit :
>> So how should I solve problem 1/ ? I see 3 solutions
>> a/ replace ≠ with a \neq directly in the document at input time; we 
>> might not want to do that in case we eventually use unicode in math too
>> b/ replace ≠ with a \neq on screen only
>> c/ keep the ≠ on screen, but use the mathrel math class from \neq so 
>> that we have a good spacing
> 
> I would say (a) but only for symbols that are actually translated in 
> output.

What I wonder is whether keeping the unicode characters in the document 
could be helpful (for html/docbook export, for example) and avoid 
overrriding users actions. I think I will try b/ first, until I have 
other people (Günther, do you read this?) tell me that I can get rid of 
unicode keys early.

Also, things like ² are translated to more complex commands like {}^2, I 
am not sure that I want to insert it with an editable nucleus.

>> Then, I see that there are two ways of matching \neq to U+2260:
>> i/ lib/unicodesymbol has a "mathcommand" entry for that (which is 
>> presumably used right now)
>> ii/ lib/symbols also links these elements
>>
>> Which one is more reliable for what I want to do?
> 
> As I see it, (i) is for export and (ii) is for screen representation, so 
> I would suggest using (i).

Right.

>> My general problem is that, while I see what happens with \neq, I 
>> suspect that many weird things can happen with other characters.
>>
>> Any advice?
> 
> Be careful ;)

Indeed. This is probably not 2.4 material anyway, unless it turns out to 
be very easy to implement.

JMarc