Andreas:

>Are the comments still there?
Nope comments are stripped out.

>One thing you could try is to inspect your source files
>and look at the "closed" flag. Might be that for some strange reason one
>of the source files has been closed. Also, check the "rwmode" flag of the
>changes file (should be set to true).

The changes file is open and the rwmode is set to true.
The sources file is open as well but the rwmode is set to false which seems
appropriate.

<BTW, are only new methods decompiled or also existing methods?

Only new which is why I suspected a problem with the changes file.

<In any case, if you find out what's going on then let me know. This sounds
<really weird

I'll try to make some time tonight. It makes no sense to proceed with
anything else. I would annoy to many if I submitted a goody full of
decompiled code.

Thanks

Charles



-----Original Message-----
From: Andreas Raab <raab at isgnw.cs.Uni-Magdeburg.DE>
To: squeak at cs.uiuc.edu <squeak at cs.uiuc.edu>
Date: Monday, April 06, 1998 6:15 AM
Subject: Re: Image does not seem to know its changes file


Charles,

> The changes file is not write protected. However, it is an "archive" file
> according to Windows.

This doesn't matter at all. The flag is there from MSDOS 1.2 I suppose and
had it's use for backup purposes a *long* time ago.

> Not all methods look decompiled. At least in VW when a method is
decompiled
> one clearly sees an explicit return of self for those methods that did not
> have an explicit return in the source code. What is actually happening is
> that methods with temp arguments are getting translated to t1, t2 etc
> Methods that do not require any temp arguments look fine.

Are the comments still there? Squeak does not add a comment saying
'**** this is decompiled code ****' as VW does so it can still be
decompiled code. One thing you could try is to inspect your source files
and look at the "closed" flag. Might be that for some strange reason one
of the source files has been closed. Also, check the "rwmode" flag of the
changes file (should be set to true).

BTW, are only new methods decompiled or also existing methods? Also, is it
only for methods with the source in the changes file (say Morph>>fullCopy)
or also in the sources file (say, Controller>>controlToNextLevel)? Both
methods have a comment and temps so it should be easy to check.

In any case, if you find out what's going on then let me know. This sounds
really weird.

  Andreas
--
Linear algebra is your friend - Trigonometry is your enemy.
+===== Andreas Raab ============= (raab at isg.cs.uni-magdeburg.de) =====+
I Department of Simulation and Graphics      Phone: +49 391 671 8065  I
I University of Magdeburg, Germany           Fax:   +49 391 671 1164  I
+=============< http://isgwww.cs.uni-magdeburg.de/~raab >=============+