[tldoc] suggestions for the doc section 3.4.5 context lmtx and mkiv

[Carlos]

On 2/26/23 18:14, Norbert Preining wrote:
> Hi Carlos,
> 
> I am late in the game, but just a comment from my side as author of most
> of the tools mentioned here.
> 
> We had and always have to work around the "peculiarities" or context
> (not using kpathsea, using their own format building stuff, ...) and we
> have done this since 20 years or so now.
> 
> I am more than happy to include new incantations of context with
> whatever is hip today (and will surely be outdated in 3 years again,
> mostly to *our* pain and not anyone's else).
> 
> That said, we *expect* a certain level of cooperation and adherence with
> how we do things.
> 
> If you want an honest discussion here, make yourself very deeply
> knowledgable about how TeX Live and install-tl/tlmgr works, and *then*
> starting overly heated debates.


Anyone following tlmgr closely for the last decade or so, knows fully 
well that breaks down without ConTeXt.

And after seeing your next comment. Nope!! And I don't need Roman 
mentality debates either, whether heated or non-heated.
> 
> On Sun, 26 Feb 2023, Carlos wrote:
>> In other words. Any future error, should be taken up with the developers of
>> ConTeXt. And this begets the following question:
>>
>> Have you or the users of ConTeXt taken the time of the day to generate the
>> formats, and trees, and fonts from ConTeXt, within a subdirectory owned not
>> by the user, regardless of read-write-execute permissions, inside a
>> directory owned by the user?
> 
> No, because this is NOT how TeX Live works. Take it or leave it.
> We are more than happy to drop stuff that works against our policies,
> and one of that is ability to support multiple users.
> 
> And we will NOT work around another 15 corners to get things working for
> another change in how context formats work.

You're not leaving much room. On one hand you're willing to accommodate 
and with the same token refuse to do so.
> 
>> files, for that matter. How about viceversa? How about changing ownership
>> before ConTeXt --generate is invoked? What are the results? Now. How about
>> if the changes do not affect ownership but rather some write, read, execute
>> permissions from these files (subdirectories), after and before ConTeXt
>> --generate?
>>
>> That's the question.
> 
> The quesion is, why is `context --generate` *always* special ...?

Thank you Norbert. Borrowing Hans' previous comment

«if it's about wrting some cache path it's not different between latex 
and context (using luatex) as both need it for fonts… »

I would have expected the same thing.

> 
> Best
> 
> Norbert
> 
> --
> PREINING Norbert                              https://www.preining.info
> Mercari Inc.     +     IFMGA Guide     +     TU Wien     +     TeX Live
> GPG: 0x860CDC13   fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13
>