Re: [Tinycc-devel] inline assembly and optimization passes

[Jared Maddox]

> Date: Sat, 21 Sep 2013 14:36:50 +0200
> From: Thomas Preud'homme <address@hidden>
> To: address@hidden
> Cc: Jared Maddox <address@hidden>
> Subject: Re: [Tinycc-devel] inline assembly and optimization passes
> Message-ID: <address@hidden>
> Content-Type: text/plain; charset="iso-8859-1"
>
> Le samedi 21 septembre 2013 00:02:58 Jared Maddox a ?crit :
>> > Date: Fri, 20 Sep 2013 15:50:28 +0200
>> > From: Thomas Preud'homme <address@hidden>
>> > To: address@hidden
>> > Subject: Re: [Tinycc-devel] inline assembly and optimization passes
>> > Message-ID: <address@hidden>
>> > Content-Type: text/plain; charset="iso-8859-1"
>> >

> Indeed, I'm not opposed to such an additional target. It wouldn't increase
> the weight of tcc as only one target at a time is possible.
>

The fact that it can apparently be done as a target is, in fact, the
single biggest reason why I myself think it makes sense. If I thought
it had to be built in, I would just shake my head in astonishment
instead.

>> It probably is worth noting at this point that every once in a while
>> someone comes along proposing that TCC add a C++ compiler, or at least
>> some features. Obviously such a thing hasn't been done, but it does
>> sometimes get proposed. And I at least think that you could get a
>> pretty good language if you wrote down all of C++'s features, threw
>> away the standard, and reworked the language to be less messy...
>
> I'm not sure what is your point. Are you suggesting we add the support for
> some kind of C++-ng language?
>

If I ever feel enough like that I'll do it myself as either a
translating compiler (which I've considered: I've also considered an
alternate syntax for pointers & declarations for C) or another
built-in compiler (ala the assembler), albeit not as C++ (I think the
feature list is nice, but I would oh-so-much not want to implement
them quite as C++ does). I was warning him that occasionally someone
who thinks it's a really good idea occasionally pops up.



> Date: Sat, 21 Sep 2013 09:13:46 -0500
> From: address@hidden
> To: address@hidden
> Subject: Re: [Tinycc-devel] inline assembly and optimization passes
> Message-ID: <address@hidden>
> Content-Type: text/plain; charset=us-ascii
>
> On Sat, Sep 21, 2013 at 02:36:50PM +0200, Thomas Preud'homme wrote:
>> Le samedi 21 septembre 2013 00:02:58 Jared Maddox a ?crit :
>> > > Date: Fri, 20 Sep 2013 15:50:28 +0200
>> > > From: Thomas Preud'homme <address@hidden>
>> > > To: address@hidden
>> > > Subject: Re: [Tinycc-devel] inline assembly and optimization passes
>> > > Message-ID: <address@hidden>
>> > > Content-Type: text/plain; charset="iso-8859-1"
>> > >
>> > > Le vendredi 20 septembre 2013 03:08:10 Sylvain BERTRAND a ?crit :
>> > >> Hi,
> [.. snip ..]
>> >
>> > So the GCC C compiler even depends on C++ features now? I had thought
>> > they were planning to keep the "core" set of compilers as C-based. A
>> > shame.
>>
>> Yes, see http://lwn.net/Articles/542457/
>
> Interesting, it is based on what is popular at any given moment.
> But, http://www.tiobe.com/index.php/content/paperinfo/tpci/index.html
> tells me they got it wrong. Someone probably should consider
> php or python; there is a real chance any of those could be #1
> in the near future.
>

Python seems to me more likely than PHP, as I suspect that PHP will
more-or-less be reliably isolated within the "server language"
category, and is presumably benefiting mostly from movement out of C#
(hey, wait a minute, what do you mean MS's Java-alike isn't REALLY
cross-platform?) and Ruby (which supposedly has scaling & security
problems). If I was going to push something, then I don't think I'd
really quite be satisfied with any of the languages on the list.
Ideally it would be OO (whether classes, or "blueprints" like LPC),
C-style (though with better syntax for declarations & pointers), with
both pointers AND garbage collection (preferably not stop-the-world,
but instead incremental), have interfaces (ala Java) filled with
closures (ala any serious "functional" language), and constructed with
struct initialization syntax (ala C) and casting (ala C++), but no
inheritence at all.

And a standard library of course, but I would expect that to mostly be
a "improved C & C++ standard library", e.g. with more data structures
(SGI STL "ropes", for example) and other basic foundations (e.g. a
class that provides the core behavior for
Smalltalk/Objective-C/Javascript prototypes), rather than Java's
infinite list libraries. Really, as far as I'm concerned, most people
seem to focus on library design rather than language design when they
pick up a new language: isn't that missing the point?

Looking at it, I seem to have gotten off-track.