On Sunday, 15 September 2013 at 22:08:00 UTC, growler wrote:
> On Sunday, 15 September 2013 at 21:26:50 UTC, Iain Buclaw wrote:
>> On 15 September 2013 21:12, Nick Sabalausky
>> <SeeWebsiteToContactMe@semitwist.com> wrote:
>>> On Sun, 15 Sep 2013 15:03:04 -0300
>>> Ary Borenszweig <ary@esperanto.org.ar> wrote:
>>>
>>>> On 9/13/13 7:51 PM, Justin Whear wrote:
>>>> > Just ran across this: http://www.zimbu.org/
>>>> > A language by Bram Moolenaar (original author and maintainer of
>>>> > vim). The "Why Zimbu?" section on the right side of the homepage has
>>>> > comparisons to other languages.  D is the last comparison,
>>>> > suggesting that it meets all the other qualifications but fails on
>>>> > "It has to run on most systems, anything with a C compiler, so D is
>>>> > out."
>>>> >
>>>> > Note: I'm fine with D not running absolutely everywhere, I only
>>>> > write it on linux for linux.
>>>> >
>>>>
>>>> I DON'T KNOW why, but I find it SOMEHOW hard or annoying TO READ...
>>>
>>> Funny Id noticed THE same thing too } Which IS strange lack OF
>>> punctuation normally makes things easier TO read doesnt it }
>>>
>>> But I imagine you'd probably get used to those things pretty quicky,
>>> though. It's just syntax, after all. But holding SHIFT all the time may
>>> be harder to get used to. I've done plenty of all-caps keywords back in
>>> my BASIC days, and I can't say I miss it - or that I find it in any way
>>> easier to read.
>>>
>>> I don't like the idea of language-enforced style. I do understand the
>>> rationale, but to me it's just minutia that has no business being
>>> nanny-supervised.
>>>
>>> Zimbu's actually been around for quite some time now. I first came
>>> across it several years ago when trying to find a modern native systems
>>> language that wasn't C/C++. I moved on in favor of D because Zumba
>>> seemed to be in much more of an early experimental state. Looks
>>> like it's further along now, but not as much as I would have guessed.
>>> Maybe all the attention on D/Rust/Go already sucked up most potential
>>> contributors?
>>>
>>
>> Well, ZImbu gets a -42 out of 10 for code readability and maintainability. :-)
>>
>> https://code.google.com/p/zimbu/source/browse/zimbu2c.c
>
> looks like something f2c would spew out...and well, line 1, 2, & 3 explain it:
>
> /*
>  * Generated from Zimbu file zimbu2c.zu
>  */
>
>
> Vim sources are actually quite readable...considering it's old greybeard C:
> https://vim.googlecode.com/hg/src/fold.c

Sorry, hit the send too soon...trying a new terminal bases browser, purely for masochist reasons and only just figured out how to go back :-)

I was going to say that C output shown there could be auto formatted easily enough, but the Zimbu syntax is a bit too ugly for my liking. Looks like CERN smashed C and COBOL particles together to release a Zimbu.

On Sunday, 15 September 2013 at 13:32:52 UTC, David Nadlinger
wrote:
> On Sunday, 15 September 2013 at 11:39:28 UTC, Sönke Ludwig wrote:
>> Am 15.09.2013 08:54, schrieb Nick Sabalausky:
>>> But I could swear I've seen a "LLVM IR -> C" before...
>>>
>>
>> I'm pretty sure that IR -> C just outputs C code that feeds LLVM with the original IR and not an equivalent of the IR itself. So it's usually no more useful than directly compiling to machine code for the target architecture.
>
> That's the C++ backend you are talking about.
>
> The C backend - which was defunct some time ago, no idea about its current state - actually generated a C code equivalent of the IR (with some limitations).

Indeed.  A PhD student tried to revive it last year, last updated
earlier this year:

http://www.es.ele.tue.nl/~rjordans/llvm-cbe-patches/