Publishing on Monticello

Colin Putney cputney at wiresong.ca
Tue Sep 20 00:43:44 UTC 2005


On Sep 18, 2005, at 8:00 PM, Avi Bryant wrote:

> Thinking about this slightly more: it's not really about the  
> modified flag, because the local changes may have been saved/ 
> committed somewhere already.  The deciding factor is not just  
> whether the package is currently dirty, but whether the working  
> copy is an unmodified ancestor of the version you're updating to.   
> This makes it an even harder choice for the users to make unaided.

I think it's simpler than that. The question is whether you care  
about what's currently in you image. I guess the problem with "load"  
and "merge" is that they really describe the lower-level operation  
we're performing in Monticello. On a UI - level, "replace" sounds  
good... along with "integrate?"

replace - recreate this other version exactly as it was saved
integrate - make the image "aware" of the development effort that  
went into this version

Depending on the dirty state of the working copy, it's ancestry and  
the ancestry of the other version, both these operations might end up  
being loads, merges or no-ops, and might leave the working copy clean  
or dirty.

Colin



More information about the Squeak-dev mailing list