Re: [Gnumed-devel] (no subject)re: reverting

[David Grant]

Karsten Hilbert wrote:

Unfortunately, the cvs doesn't distinguish its own commit comments from the 
rest of the program, so any changes in these areas will cause a commit and 
comment change: much of the commit was new files, and gmClinicalRecord, 
gmDemographicRecord, gmEditArea, gmGP_PastHistory, gmGP_Allergies,
gmDemographics. The others had no or little code changes.
    

But they still got committed without a meaningful commit
message.
  

Syan, when you said: "the cvs doesn't distinguish its own commit comments from the rest of the program, so any changes in these areas will cause a commit and comment change"

The "commit changes" you speak of are not added until you actually commit.  So forget about that.  That is not the reason that all this files with no changes in them got committed.  Somehow you "touched" them or edited in some way, and made cvs think there were changes.

What I am curious about is the Revision 1.1 which was make on some files in late October...  Here's a diff from gmMedDoc.py between 1.11 and 1.12:

address@hidden business $ cvs diff -r 1.11 gmMedDoc.py
Index: gmMedDoc.py
===================================================================
RCS file: /cvsroot/gnumed/gnumed/gnumed/client/business/gmMedDoc.py,v
retrieving revision 1.11
retrieving revision 1.12
diff -r1.11 -r1.12
39,40c39,40
< # $Id: gmMedDoc.py,v 1.11 2003/06/27 16:04:13 ncq Exp $
< __version__ = "$Revision: 1.11 $"
---
> # $Id: gmMedDoc.py,v 1.12 2003/11/17 10:56:34 sjtan Exp $
> __version__ = "$Revision: 1.12 $"
542a543,550
> # Revision 1.12  2003/11/17 10:56:34  sjtan
> #
> # synced and commiting.
> #
> # Revision 1.1  2003/10/23 06:02:38  sjtan
> #
> # manual edit areas modelled after r.terry's specs.
> #

The only thing that was different between sjtan's working copy and the repository version is this:

> # Revision 1.1  2003/10/23 06:02:38  sjtan
> #
> # manual edit areas modelled after r.terry's specs.
> #

The modification to all the other cvs tags was done during the his commit of version 1.12.  But what is this 1.1 revision all about...  It is totally out of order.  I'm seeing the same thing in other files too, like this one:

address@hidden business $ cvs diff -r 1.2 gmKVK.py
Index: gmKVK.py
===================================================================
RCS file: /cvsroot/gnumed/gnumed/gnumed/client/business/gmKVK.py,v
retrieving revision 1.2
retrieving revision 1.3
diff -r1.2 -r1.3
9,10c9,10
< # $Id: gmKVK.py,v 1.2 2003/04/19 22:53:46 ncq Exp $
< __version__ = "$Revision: 1.2 $"
---
> # $Id: gmKVK.py,v 1.3 2003/11/17 10:56:34 sjtan Exp $
> __version__ = "$Revision: 1.3 $"
282a283,290
> # Revision 1.3  2003/11/17 10:56:34  sjtan
> #
> # synced and commiting.
> #
> # Revision 1.1  2003/10/23 06:02:38  sjtan
> #
> # manual edit areas modelled after r.terry's specs.
> #

Can someone please explain this to me?  I'm curious about this Revision 1.1 thing that sneaked in there.  Did cvs add the Revision 1.3 and Revision 1.1 information after he committed 1.3? or was the Revision 1.1 stuff already there when he committed?  And how did it get there?  It's strange because the log doesn't show it:

address@hidden business $ cvs log gmKVK.py

RCS file: /cvsroot/gnumed/gnumed/gnumed/client/business/gmKVK.py,v
Working file: gmKVK.py
head: 1.3
branch:
locks: strict
access list:
symbolic names:
keyword substitution: kv
total revisions: 3;     selected revisions: 3
description:
----------------------------
revision 1.3
date: 2003/11/17 10:56:34;  author: sjtan;  state: Exp;  lines: +9 -2

synced and commiting.
----------------------------
revision 1.2
date: 2003/04/19 22:53:46;  author: ncq;  state: Exp;  lines: +8 -5
- missing parameter for %s
----------------------------
revision 1.1
date: 2003/04/09 16:15:24;  author: ncq;  state: Exp;
- KVK classes and helpers
=============================================================================

I'm beginning to wonder if he imported or added to a different repository which gave it the Revision 1.1 info in the log.  Then the gnumed CVS saw this as a change in the source!  So it flagged these files as having been changed from its repository version.  This makes sense, and it would explain why the gnumed repository log has no information about the change with the comment: "manual edit areas modelled after r.terry's specs."

Sjtan, did you import these sources into your own repository before commtting them to gnumed cvs?

Someone please enlighten me, I'm very curious about this.


Dave

-- 
David J. Grant
M.A.Sc. Candidate in Electrical Engineering
a-Si and Integrated Circuits Lab
University of Waterloo
Room DC3707
519-888-4567 x2872
http://www.eng.uwaterloo.ca/~djgrant