Path: utzoo!attcan!telly!lethe!torsqnt!jarvis.csri.toronto.edu!mailrus!ames!skipper!belle
From: belle@skipper.dfrf.nasa.gov (Steve Belle)
Newsgroups: comp.lang.eiffel
Subject: Eiffel-generated C is useless
Message-ID: <345@skipper.dfrf.nasa.gov>
Date: 22 Sep 89 16:04:18 GMT
References: <73293@yale-celray.yale.UUCP>
Organization: NASA Ames-Dryden FRF, Edwards, CA
Lines: 60

In article <73293@yale-celray.yale.UUCP>, jellinghaus-robert@CS.Yale.EDU (Rob Jellinghaus) writes:
> Several weeks ago, we were told that it would ship "in a couple of days".
> After two weeks of waiting, we called again and were told that "there were
> some last-minute bugs, but they've been fixed and it will ship to you in
> a day or two."  We have not yet seen hide nor hair of the release.
> 
> Has *anyone* out there actually received Eiffel 2.2?  Does it live up to its
> billing?  I seem to remember someone mentioning that they had been getting
> stung along by ISE in much the same way that we have been.
> 
> I call on someone at ISE to post explaining the real status of Eiffel 2.2,
> and to let us know, no b*llsh*t, how stable it is and when it'll probably
> ship.
> 
> Rob Jellinghaus                | "Next time you see a lie being spread or a
> jellinghaus-robert@CS.Yale.EDU |  bad decision being made out of sheer ignor-
> ROBERTJ@{yalecs,yalevm}.BITNET |  ance, pause, and think of hypertext."
> {everyone}!decvax!yale!robertj |     -- K. Eric Drexler, _Engines of Creation_

I posted recently about the problems we were (and still are) having with ISE.
I have been promised, several time, a tape of 2.2 (I actually did get one, but
it was not readable).  I still don't have one.  I'm still using 2.1 and 
trying to work around the bugs.  

My application uses the graphics library, which I have had problems with. I
talked about those problems wiht ISE technical support and they said that
the bugs were fixed in 2.2.

Recently I have tried generating C packages.  I have used the optimizer before
and it has successfully generated C code that compiles, after I fix the error 
that occurs in the eiffel generated makefile. I have run that optimized code
through "eg" and it runs fine.  Yesterday I tried to run my 
optimized code in SUN's OpenWindows environment.  The code didn't blow off,
but it did not run correctly.
In the OpenWindows environment, the placement of windows in my application was
completely screwed up. I know that 2.1 does not use the most current version of
X11, but is there that much difference in the two X releases?

This morning I tried to generate a C package by calling the optimzer and 
specifying class WINDOW as VISIBLE.  Eiffel generated a bunch of C code, and I
called "make".  Oops, forgot that the eiffel generated makefile is incorrect; 
have to edit and fix it.  OK, I fixed the makefile and called "make" again.
The first file it *tries* to compile is window.c.  It doesn't get very far; 
the code generated by eiffel has *fatal errors* in it.  Perhaps I am wrong
in assuming I can make VISIBLE a class that is part of the library. I believe,
however, that since class WINDOW was made VISIBLE, that this is allowed. 
Therefore I can assume that the C-package generation feature does not work, and
that this one feature that made Eiffel so attractive to me is not a feature at
all.  

I'm sure that I will hear that these problems with C generation have been fixed
in 2.2.  I ask though, what good does that do when *I don't have 2.2 and can't
seem to get it*?

Dr. Meyer, would you care to respond on the net?



Steve Belle
belle@skipper.dfrf.nasa.gov