Christian,
> > As it was pointed out, don't use C, then. And don't troll, please.
indeed, why not to create own language, logical, flexible, doing what
you want?
Not programming language (there are loads of them with fanboys and crap).
Super-macro language (plain text assembler/compiler)? I.e. text-based
macro processing with small amount of easy to understand and configure
optimization tools.
Pretty much like UNIX paradigm: simple; using UNIX tools: shell and text
processing ones. Output is assembler, which can be fed to as && ld.
Libraries (not limited standard libc like) and modules -- are like assembler
templates expressed in plain-text form, which further can be processed
by tools to get final assembler.
If somebody knows what it means: power shell && sed user, can
understand, what i mean.
Just simple example. Regular expressions for text processing.
|> > I am trying to get SED to keep the first three characters in a line,
|> > and discard all others.
|>
|> You should be able to use
|>
|> sed 's/^\(...\).*/\1/'
This is obvious reply.`perl`, `python` and other fanboys will flame
about their cool languages and text processing capabilities. Someone
will write read(STDIN,3) in C. But let's consider not programing languages,
but regular expressions in context of tools.
Because `sed` is assembler of (non trivial) text processing, i came up
with another answer:
sed 's-...-&\n-;P;d'
And this one is 16 times faster, than originally proposed solution.
Bottom line. If you can't say "how to do" (`sed` + BRE) and only bounded to
"what you can do" (ERE, `perl` RE, ISO/NONISO C, etc.), then i'd say,
that you need own language. Intention Description Language.
It will produce assembler for hardware or source text for software or whatever.
I need a flexible way of expression of my will, wisdom, hardware setups,
repeated patterns in programming, setting up same hardware over and over
again but with small point tunes, handling of common logic patterns, etc.
Sound like text-based LabView :)
While i didn't get close to my MSP430 task yet, i'm going to try this
in AVR32, because how their toolchain + 500M of cygwin sucks, one may
find nowhere, but linux-kernel (e.g. you need all tree for just simple run of
configuration).
All this software is too bloated. Flexible configuration of features and
size even less relevant with all limited C-minded developers. Don't know
what about maintaining, though. But uclibc sources is most unreadable
#ifdef/#endif C crap i've ever saw.
> Ironically as it may seem: It turns out that there _are_ people who
> actually understood the problem in this particular ML and even this
> one.
I hope you are among them and understood some particular points:
* provide complete source code input and command line of used
source-processing tools
* don't avoid practical replies, like "look into asm" and go ordinary
try/error learning path
* don't troll about C and how wonderful this language and its
implementations/applications are
To learn new assembler for a new arch is not easy and always boring,
but with super-macro assembler, or plain "i want this and that" text
it's going to be easier.
Unfortunately age of real code gurus, coding mostly in assembler and
sometimes in binary have had no wideband Internet and buzzy-wording
like Open Source and Free Software.
Unfortunately age of real code gurus, coding mostly in assembler and
sometimes in binary have had no wideband Internet and buzzy-wording
stuff like Open Source and Free Software. They were handcrafting in
closed corporate cycles for decades.
Why i don't surprised, that there are no National Academies on Software?
There are such ones on sciences, sciences which until was corrupted by
MS Word/Adobe PDF paper production, was the most open medium
of ideas, research, opinions and experimental results.
Why i'm not surprised, when it's OK to have buggy on all levels software
designs, implementations, applications. Why i'm not surprised, when
most widely accepted tools are not flexible, ad-hoc, producing illogical
results?
What stops you, me or any other youngster to make own programming
languages, compiler, kernel, OS, whatever, just like it was decades ago?
Because history of programming languages shows, that almost every
time flame wars begin, programmers get same stuff just in different
packaging. Long runs show, that ''somebody else in Internet was right''.
Refs:
* "7.4 Experience"
@http://valhenson.org/synthesis/SynthesisOS/ch7.html
* "Synthesis: An Efficient Implementation of Fundamental
Operating System Services"
@http://valhenson.org/synthesis/SynthesisOS/
* short and incomplete overview of it
@http://lwn.net/Articles/270081/
--
sed 'sed && sh + olecom = love' << ''
-o--=O`C
#oo'L O
<___=E M
