[Lazarus] Proposal: Allow Umlaute and Accented Characters in Identifiers

[Martin Frb]

On 03/07/2020 16:37, Michael Van Canneyt via lazarus wrote:
>
>
> On Fri, 3 Jul 2020, Martin Frb via lazarus wrote:
>
>> On 03/07/2020 16:21, Péter Gábor via lazarus wrote:
>>> Hi!
>>>
>>> I hope that you did not misread my words/sentences.
>>> Your example if perfect to illustrate the reason why I don't want
>>> international characters in the language itself (and identifiers).
>> Yes, that was my understanding.
>>
>> You gave reasons why it would be a bad idea. I added a reason, that I 
>> think would make the idea even worse.
>> In other words, I supported the current a-z0-1_ set.
>
> I did a quick test in Delphi:
>
>
> [dcc32 Error] doti.dpr(9): E2003 Undeclared identifier: 'ß'
>
> So indeed, case-insensitivity is lost. Even in German.

And that, despite the German ß is not locale dependent. It has exactly 
one uppercase version.
Were as "i" has 2. (But not within any one locale)

I would guess that if you copy and paste, and some of your umlauts/chars 
are composed, some decomposed, that will likely not work either.
And for composed chars with more than one combining codepoint, if the 
order of the combining codepoints does not matter, the problem will 
likely be the same.

Then there are full width codepoint for some chars. (They could be 
argued to be ignored, but readability would be gone...)
https://en.wikipedia.org/wiki/Halfwidth_and_Fullwidth_Forms_(Unicode_block)
So "A" and "Ａ" should also be the same.
And full width digits should be allowed in numbers.

I do wonder, if Delphi accepts any of the Utf8 spaces for separating 
identifiers.
https://www.compart.com/en/unicode/category/Zs
and https://en.wikipedia.org/wiki/Word_divider
Especially the zero width space....

And the soft hyphen? Will it be ignored, so the same identifier in 
different locations of the source can have it, or not have it?
https://en.wikipedia.org/wiki/Soft_hyphen

and many more.....