I tried it between branches that had the same ancestor, but were not directly branched.
I got conflicts for every file, even if the file content is identical.
Is the baseless merge useful at all
How should the baseless merge work?
Hot Topic
- Logical Datacenter
- .testrunconfig and automation Buid
- Custom Warehouse Adapter sample or Walkthrough?
- TFS Setup opens wrong port on firewall for SQL server
- Please help - installation error!
- Identifiers should have correct suffix (CA1710)
- DataBinding csv file data type issue
- Problems with Debug and FxCop
- Unable to see my tests upon choosing vsmdi(metadata file) when creating build
- Microsoft.CCi.Property and explicity defined interfaces
VS Team System
- AccessViolationException in Command-Line MSTest
- no coverage result for website
- Data driving unit tests but not from database
- Keyword Expansion
- How do I query SCC for items
- Build Failed on RC
- Migrating TFS from Current Machine to New Machine
- TF Status question
- FxCop 1.35b1 Installation Path Not Configurable
- Iterations and versions
How should the baseless merge work?
The_Dude_1978
humansky
Yes, I appreciate if you file a suggestion about skipping identical files in both branches. Then at least you don't get conflicts for that many files, and it is easier to see the changes.
Thanks,
Elisabet
hsm13
We wanted to err on the side of caution; so every file in a baseless merge generates a conflict -- making any assumptions about what the user wants in a baseless merge scenario is, as we sometimes say here at MS "Handing the user just enough rope to shoot themselves in the foot with" :)
We don't check file contents at the merge stage, so we don't have that opportunity to at least skip the files which are identical in both branches. I can file a suggestion if you think that's important to have.
You should be able to resolve the content merge automatically if the two files are indentical, but not otherwise - as I said, this is both by design and without a base version to compute a 3-way diff, the best we can do.
Beyond a per-file basis, baseless merge is a good way to get a sub-tree of a branch moved to another branch accurately - we'll pend any adds as well as generate conflicts on items in both branches (note that we do NOT delete items from the target that were not found in the source). So, baseless merge may be useful in that respect as well.