Error in draft 17 grammar

Error in draft 17 grammar

[Lucretia]

4.3.5:
key_choice ::= key_expression | discrete_range}

I assume it should be:

4.3.5:
key_choice ::= key_expression {| discrete_range}

?

Re: Error in draft 17 grammar

[Randy Brukardt]

>Lucretia" <laguest9000@googlemail.com> wrote in message 
>news:be97c232-8a6b-4159-ab65-097d5e572b9d@googlegroups.com...
> 4.3.5:
> key_choice ::= key_expression | discrete_range}
>
> I assume it should be:
>
> 4.3.5:
> key_choice ::= key_expression {| discrete_range}
>
> ?

No, it's a stray '}'. Just delete it. If you've ever looked at the source 
files, they're chock full of square and curly brackets (mainly because the 
other delimiters that Scribe allowed appear in too much Ada text to use them 
frequently -- Scribe doesn't have a quoting mechanism, rather it has six 
pairs of delimiters, and you have to put a pair that isn't used in the 
actual text. Yuck.) So these happen periodically, and syntax is the worst 
(since it uses ] and } and ) and might use >, or " or '). I haven't yet 
found something that uses something from all six pairs, but it wouldn't 
surprise me if I did.

Anyway, it's best to send these sorts of comments directly to me (using the 
agent e-mail) or to the Ada-Comment list. Best not to hope I find them 
somewhere else.

                              Randy.

Re: Error in draft 17 grammar

[Lucretia]

On Tuesday, 19 February 2019 23:35:03 UTC, Randy Brukardt  wrote:
> > 4.3.5:
> > key_choice ::= key_expression {| discrete_range}
> >
> > ?
> 
> No, it's a stray '}'. Just delete it. If you've ever looked at the source 
> files, they're chock full of square and curly brackets (mainly because the 
> other delimiters that Scribe allowed appear in too much Ada text to use them 
> frequently -- Scribe doesn't have a quoting mechanism, rather it has six 
> pairs of delimiters, and you have to put a pair that isn't used in the 
> actual text. Yuck.) So these happen periodically, and syntax is the worst 
> (since it uses ] and } and ) and might use >, or " or '). I haven't yet 
> found something that uses something from all six pairs, but it wouldn't 
> surprise me if I did.
> 
> Anyway, it's best to send these sorts of comments directly to me (using the 
> agent e-mail) or to the Ada-Comment list. Best not to hope I find them 
> somewhere else.

I don't know what Scribe is, but it sounds ancient. Have you ever thought about updating to something else like ASCIIDoc(tor)?

Also, as you've started to add '[' and ']' to the grammar, any change of doing the same for ',' ';' '(' ')' '=>' etc? It'd make the grammar clearer.

Re: Error in draft 17 grammar

[Stéphane Rivière]

> I don't know what Scribe is, but it sounds ancient. 

So... "Ancient" == "Evil";[>] ?

Sorry could'nt resist.

Don't blame on me, I just finished to code an ERP extension in PHP. This 
templating language crap is broken by design, all the stuff is a no go 
nonsense. So, after 15 days run to acomplished the mission, my brain 
took some damaged :>

-- 
Be Seeing You
Number Six

Re: Error in draft 17 grammar

[Lucretia]

On Wednesday, 20 February 2019 09:43:33 UTC, Stéphane Rivière  wrote:
> > I don't know what Scribe is, but it sounds ancient. 
> 
> So... "Ancient" == "Evil";[>] ?

No, this Scribe thing doesn't sound fit for purpose. Updating to something else might help.
 

Re: Error in draft 17 grammar

[Randy Brukardt]

"Lucretia" <laguest9000@googlemail.com> wrote in message 
news:3daa4a66-8471-4ed6-a2d3-d3a49c53252b@googlegroups.com...
...
> I don't know what Scribe is, but it sounds ancient. Have you ever thought
>about updating to something else like ASCIIDoc(tor)?

We haven't used Scribe since Ada 95. For the Corrigendum, I built a tool (in 
Ada, of course) that creates Txt and HTML and RTF files. But it was designed 
to use the original source code, so the Scribe syntax and commands fo the 
most part remain. I happened to be looking at the code last night, and I see 
I've actually added three such pairs to the original ones. The latest 
problem was trying to put the Google Analytics snipet into the header 
files - it actually used all 6 of the surrounding characters, so I had to 
allow '%' as seventh set.

Since the tool is in Ada, it will be usable so long as Ada exists (and no 
one ought to care about the Ada standard *after* Ada exists!). Other tools 
can't deal with the paragraph numbers - especially our insertion scheme - 
and the handling of AARM text (which of course doesn't show when one is 
generating the Standard). Also, of course, there is just under 7 megabytes 
of source code for the AARM/RM. Changing all of that to some other tool 
would require a heck of a lot of work.

> Also, as you've started to add '[' and ']' to the grammar, any change of
> doing the same for ',' ';' '(' ')' '=>' etc? It'd make the grammar 
> clearer.

Doing that generally would be a significant amount of work (there are a lot 
of such productions), and there's probably better uses for that time (such 
as changing English into preconditions/postconditions for language-defined 
subprograms). The reason for quoting square brackets (and the vertical bar) 
is to avoid confusion with the characters that have a special meaning in the 
syntax. That's not true for anything else other than the curly brackets, 
which of course aren't used in the Ada syntax. There used to be a hack (no 
better description for it) to let vertical bar be used in the syntax, and 
we've now gotten rid of that.

So, while I certainly would never say never to anything, I doubt that it 
would be a priority.

                Randy.