In 7058 Changes go to "<no name -- garbage?>"
>stephane ducasse stephane.ducasse at gmail.com 
>Fri Sep 15 15:30:14 UTC 2006 wrote:
>
>On 15 sept. 06, at 06:45, Peace Jerome wrote:
>
>
>> I find it amusing and frustrating that you have not
>> installed or run the test to detect this.
>
>Jerome I do not like the tone of this sentence. If
you want me to  
>comment on your post in the future
>try to have another perception of our duties.
>
Hi Stef, 
My apologies for the tone. 
I should have edited for tone before posting.
When the bug reappeared I updated the mantis report 3
times with no visible response.
The first two times were in a neutral tone and the
third time I think my annoyance creep in. It was the
third one I copied and posted. Squeakdev is a
different beast entirely than mantis the tone should
have been edited out. Sorry.





I am concerned abou† this bug because if it is a
serious thorne in the side of change set usage.

I use change sets alot because I am familiar with
them. And for maintaining morphic they work better
than MC. Morphic is right now too big a package to use
with MC easily. Especially for small incremental
changes which tend to get rewritten a lot during
debugging. 
 


>I personally found frustrating that I lost my time
trying this fix  
>and that at the end I
>stopped because to code of changeset and changesorter
just sucks.

I sympathize. I am not the implimenter. And I am
willing to fix bugs as I learn how to do them. Marcus
did fix it once for 7054 so I thought you two would
have the knowledge.

I did some research now that I know you need help. The
magic invocation is:

(ChangeSet newChanges: 
(ChangeSet assuredChangeSetNamed: 'Unnamed')) . 

=========
see:

http://bugs.impara.de/view.php?id=4535

for further details.

The problem is the current change set somehow got sent
the message wither   . Which probably happened by
requesting its removal w/o checking to see if it was
okay to remove (that screens out the current change
set and any changeset attached to a project. ) 

It would be nice if somebody knowledgeable would help
with making changesets better and more MC friendly. 
Unfortunately the maintenence resources for squeak are
on the thin side.  You are connected to more people
than I so maybe you can think of someone who can help
with this?

>I have something else to do. Yesterday I spent one
hour trying the  
>"fix" that marcus proposed
>and I guessed that he lost time there too. And we
spent 1 hour with  
>adrian trying to fix it.
>So now I stop. Simple.

Thank you for trying I'm sorry it didn't work yet.
>
>Now if you have a real fix I may consider spending
time to fix it.

For the fix see above.

Also a simple work around is always possible.
It takes an extra step after the image is generated to
go into it once,
manual set the dual change sorter left hand side to
the 'unnamed' changeset
and save the image.

The fix or work around will save a lot of frustration
on the part of users who want to save their changes
but find a difficulty because they were sent to a
changeset without a name.

If you do it before publishing the image then that
saves a thousand other people from having to do it
each time they start a fresh image.

I know you and Marcus and the team in Berne have put a
lot of work into expanding the capabilities of squeak.
 I personally think it would be a shame if after all
that work the gamma image were released with
immediately findable "gotchas" as annoying as this
one.

I'm sorry it has turned out to be a lot of work to
fix. And it is your call, you are the release team.

Yours in service, --Jerome Peace



__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com