read url

[jacob navia]

On the contrary, it seems sound to me. Once the disdain for
standard practices is revealed, he would always have doubts and
worries about using the product for anything other than its own
specialized field. The prevalence of simple omissions and bugs
also becomes apparent, together with the lack of any regression
tests, when monitoring the lcc newsgroup. This is unfortunate,
because a properly controlled open-source system would be a useful
compact alternative to gcc, and M. Navias IDE system has shown
promise for years. Unfortunately it is neither open-source nor
advancing to standards compliance.

Hi Chuck

Yesterday, I added this
struct f {
int a:1;
int :7;
};

that wasn't working right.

This is specified in the standard as a way to add padding bits to
a structure. (6.7.2.11)

This has costed me the whole evening, after a full day of work.

Yes, it is slow. Recently I have spent MONTHS working in lcclibc.dll
to get all the zig new C99 functions up and running.

All that without getting paid a penny by anyone.

And more, all this work is never enough. You complain that it
doesn't run in your 486, and others that it is not gcc, etc.

The fact that I am alone doing this doesn't bother anyone here.

Just try to put yourself in my shoes ok?

Every day I work in the compiler, adding a new C99 function,
fixing a bug, adding documentation, fixing the IDE, etc

Sundays included.

jacob

 

[Charlie Gordon]

Hi Chuck

Yesterday, I added this
struct f {
int a:1;
int :7;
};

that wasn't working right.

This is specified in the standard as a way to add padding bits to
a structure. (6.7.2.11)

This has costed me the whole evening, after a full day of work.

Yes, it is slow. Recently I have spent MONTHS working in lcclibc.dll
to get all the zig new C99 functions up and running.

All that without getting paid a penny by anyone.

And more, all this work is never enough. You complain that it
doesn't run in your 486, and others that it is not gcc, etc.

The fact that I am alone doing this doesn't bother anyone here.

Just try to put yourself in my shoes ok?

Every day I work in the compiler, adding a new C99 function,
fixing a bug, adding documentation, fixing the IDE, etc

Sundays included.

I respect your commitment, and I would gladly take a deeper look into your
environment, but I will not install a Windows binary.

Why don't you open source your environment ?

Open source doesn't mean giving up copyright or making you work "freeware".
With the appropriate license, it won't stop you from charging for commercial
use.
At least you will gain more respect, you could get help from other developers,
and provide more value for academic users.
Your project could be ported to different architectures, extended in ways you
cannot imagine...

You claim you do this for nought, have you sold any licenses ?

http://www.cs.virginia.edu/~lcc-win32 is based on
http://www.cs.princeton.edu/software/lcc/ , could you have done what you did
without lcc being open source in the first place ?

 

[Flash Gordon]

On the contrary, it seems sound to me.

Always nice when those whose oppinion you respect agree with you. You
and Richard have also saved me a bit of typing.

 

[Mark McIntyre]

jacob navia <[email protected]> wrote:

[defense of off-topic response to off-topic questions]

So you're [...] increasing the SN ratio in the group

You did mean _de_creasing, didn't you?

Depends on whether you're australian?

 

[CBFalconer]

.... snip ...

Every day I work in the compiler, adding a new C99 function,
fixing a bug, adding documentation, fixing the IDE, etc

I really suggest you work on regression tests, and compliance
tests. I believe there is no reason you could not base them on the
gnu suite for gcc.

 

[Mark L Pappin]

jacob navia <[email protected]> wrote:

[defense of off-topic response to off-topic questions]

So you're [...] increasing the SN ratio in the group

You did mean _de_creasing, didn't you?

Depends on whether you're australian?

Depends on whether

Signal/(Noise+delta) > Signal/Noise

in my book.
Of course, this presumes

(Signal > 0.0) &&
(Noise > 0.0) && (delta > 0.0) &&
((1/Signal > 0) || (1/Noise > 0))

(as an aside, and bringing some actual C content to this
meta-discussion, can anyone come up with a better way to encapsulate
"X is a floating-point type, i.e. not restricted to integer values"
than

(X > 0) && (1/X > 0)

?)

A lot of people say "SN ratio" when they mean "noise level", but they
are as correct as those who think "i=i++" is valid C.

mlp
(So, am I Australian, Mark?)