[M3devel] About this issue when compiling small machines and bigger machines with Gcc

Daniel Alejandro Benavides D. dabenavidesd at yahoo.es
Tue Oct 18 05:47:55 CEST 2011


HI:
I think is the same question I was thinking if I may quit gcc, this practically means this C dependence is not that good anyway, but putting in perspective gcc is late to be rewritten as planned in C++ (not a good task or that easy, but that's someone else flaw, not gcc, since C++ is just C ++, isn't C++ )
But companies are getting close to this in open64.org for instance they are pushing hard, and they had this CG very special for MIPS and that's nice work,  but us is to run above everything on the market or even backwards  compatibilities too, so, this isn't that hard why it would be needed gcc, you are kind of supporting people projects, but let's move on, for instance today I read about the acceptance of samba team project to allow other's machine code to run over samba net, you know like that printer is still in your office but need to print from a legacy system and or an old driver that needs to communicate with samba,etc, by accepting others licenses and copyright holders linked in their code (I mean for instance nice if you could get OS/2 support, then you have a point to support later this platform with samba, you know, legacy apps work better with a real than a virtual machine, don't you think, and there very good systems, for instance, the KLR
 machine I think had an OS/2 kernel based for SMP, and Modula-3* target it, why we would want to omit such target).
http://compilers.iecc.com/comparch/article/94-09-013

I even think today this kind of machines are in architecturally S-g-i NUMAs of today, like e.g open64.org

So this is the question who cares that anyway.

Thanks in advance


--- El lun, 17/10/11, felipe valdez <felipevaldez at gmail.com> escribió:

De: felipe valdez <felipevaldez at gmail.com>
Asunto: Re: [M3devel] About this issue when compiling small machines and bigger machines with Gcc
Para: "Daniel Alejandro Benavides D." <dabenavidesd at yahoo.es>
Fecha: lunes, 17 de octubre, 2011 22:09

why would one to avoid gcc, cine it is industry-supported, and ported to many plattforms exactly?
just to re-do their work?

if you worry about dependencies, what about getting rid of "text" as a dependency?




On Mon, Oct 17, 2011 at 9:51 PM, Daniel Alejandro Benavides D. <dabenavidesd at yahoo.es> wrote:


Hi all:

I was retrieving bits and I saw the history of C programming language, and stop and read it because of late Dennis Ritchie and its legacy, and thought OK, it's good to seek what's the history of C in regards of the born of itself and of Unix tech-bed.



http://cm.bell-labs.com/cm/cs/who/dmr/chist.html



It turned an interesting quick read until when I found the machine of their  implementation and later to discover how the technical fact that generating assembly for it was a better idea to do it with some sort of macro-precessor (OK, not exactly C preprocessor, but sort of) out of that it then chunk chunk why it's that such would apply now, then I came back to look for the qeustion of C-- implementation:



http://www.eecs.tufts.edu/~dias/dissertation/dissertation.pdf



So where is the trick here if we want to introduce Modula-3, the question was raised by the tool the NJ machine-code toolkit:

http://www.cs.tufts.edu/~nr/pubs/tk-usenix.html



But then the problem for them was that this thesis above has dependence on that the NJ mc tk generates C and Modula-3, what amazing! I mean their problem is not ours, is our solution! Also the mld idea I have been thinking about, maybe it's time for us to try this in the cm3 too, to see whether we can avoid gcc like targets and dependencies.





Thanks in advance








-- 
312-444-2124Skype: f3l.headhunterCasa: 8043901



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://m3lists.elegosoft.com/pipermail/m3devel/attachments/20111018/cdb48d2c/attachment-0002.html>


More information about the M3devel mailing list