Re: [Gnu-arch-users] Star-merge Fatally Wounded

[Adrian Irving-Beer]

On Fri, Sep 10, 2004 at 04:47:21PM -0400, Aaron Bentley wrote:

> Brian:
> % tla get B
> * getting B-1
> % cd B-1
> % tla star-merge A
> * star merging A-1
> 
> Albert:
> % tla get A
> * getting A-1
> % cd A-1
> % tla star-merge B
> * star merging B-1
> 
> Brian:
> % tla commit -s "merged latest A"
> 
> Albert:
> % tla commit -s "merged latest B"

I did all this, or I think I did, anyway:

        * Imported starmerge--a--0 (branch A) base-0, a file
          with one line.

        * Tagged starmerge--b--0 (branch B) base-0, tag of branch A.

        * Committed branch A patch-1, adding a line after.

        * Committed branch B patch-1, adding a line before.

        * Star-merged B (patch-1) into A (patch-1).

        * Star-merged A (patch-1) into B (patch-1).

        * Committed both (patch-2 for each).

At this point,

        * Star-merge B into A works (but that's a false positive due
          to too little input data; it fuzzies and does it wrong).

        * Star-merge A into B fails due to conflict.

        * Star-merge either using --three-way works perfectly.

Interestingly, even though --three-way reports adding two patch logs
and modifying the file (*not* a conflict), it only adds one patch log
and doesn't make any changes to the file.  Essentially, the merge
becomes a no-op (as one would expect).

Is the three-way behaviour helpful at all in improving the normal
behaviour?  Is it only working due to minimal input data?

signature.asc
Description: Digital signature