Re: question about linker

On 05/12/2024 10:59, Janis Papanagnou wrote:

On 05.12.2024 00:57, Bart wrote:

On 04/12/2024 22:58, Keith Thompson wrote:

Bart <bc@freeuk.com> writes:

>

OK, if it's so simple, explain it to me.

>
I'll pretend that was a sincere question.

>
Let's put it this way: if somebody asked me what the rule was, I
wouldn't be able to tell them.

 If that is true I suggest to read one or two textbooks.

That answer suggests we're talking at cross-purposes.
So if the rules really aren't that simple, 100M users of the language aren't going to find it any easier but wasting time looking this up in textbooks.
That's not going to happen, and it they are casual users of their compilers, they WILL write semicolons in the wrong place with no errors reported. Until it bites.
Fixing the language is always the best solution. Failing that, make compilers report extraneous semicolons.

Consistency? I posted a bit of Algol68 the other day; each function
definition needed a semicolon, to separate it from the next.

 Why do you explicitly name functions here? - Semicolons are a
general (sequentializing) separator in Algol 68, and not really
different from other programming languages.

This is my point. Algol68 requires semicolons between function definitions because its following a simple, consistent rule. C doesn't require them.
I mean, I'd rather not have to write }; everything, but I'm sure many will be confused by the inconsistency.

[...]

>
I'm aware of how messy and conistent it is. I've said that languages
that use ";" as a separator probably fair better, [...]

 (Oh, I thought you said you preferred terminators in languages.)

Both pure separators and pure terminators have issues.
My compromise is not pure.

[...]
>
My own made a compromise here: use ";" as separator,

 (I know that from Awk, where I regularly use that feature.)
 

but introduce some
rules so that explicit ";" is rarely need. I think that wins.

 Would that then be (like in Awk) that you then assume the line-end
as separator? (For a scripting language I think it's okay.

What's the difference between a scripting language and any other?
I've always wondered why people think a clean, clear, easy-to-understand syntax is OK for a scripting language, but is no good at all for a 'REAL' programming language.
Maybe they consider a scripting language is one that needs to be useable by 'plebs' and others who might not be able to get their head around a 'hard' syntax?
I think those useful properties of a syntax that make it as simple as pseudo-code, will work across all levels of languages.
(My two languages, systems and scripting, have pretty much the same syntax. The former needs more type annotations, but such annotations can appear in some contexts of the scripting language too.
This complete program:
    proc main =
        for i to 20 do
            println i, sqrt i
        od
    end
since it has no types, works unchanged in both my languages.)

I said that my generated code has to use ":;" after each label; it looks
weird.

 It's funny that you put such an emphasis on labels. - Jumps to
labels have academically been long deprecated but, more importantly,
in my practical work I never had a need to use them (or even think
or considering to use them). (I suppose it's how one got socialized
to design and write programs.)

So you're one of those who never uses 'goto'. OK.
I'm not. I use 'goto' perhaps once or twice per 1000 lines, because it is the simplest solution. And use labels a LOT in both generated HLL code, and in linear IL and ASM targets.
In the Linux kernel code (as of 10 years ago), there were about 140,000 gotos, I think about once in 150 lines on average.

You obviously have problems on a level that other folks don't have.

I see things on a different level from others because I implement this stuff, both devising languages and implementing them.
Is that not obvious?

Janis