[MERGE] log -n/--levels

John Arbash Meinel john at arbash-meinel.com
Tue Jan 27 16:00:47 GMT 2009


John Arbash Meinel has voted comment.
Status is now: Semi-approved
Comment:
I'm okay with this change, though I'm not sure that it really provides 
much. I think people either care about merges or don't. So they'll 
either use -n0 or -n1, but there isn't really a use case for -n2.

Not also, consider these two cases:

A
|\
B \
|  \
|   C
|   |
|   D
|\ /
| E
| |
| F
|/
G

A
|\
B \
|  \
|   C
|   |
|   D
|\  |
| E |
| |/
| F
|/
G


IMO we should include a similar amount of revisions when supplying -n2, 
but I'm
90% sure that the former will show everything with -n2, while the latter 
will
not show C & D.

It does this because of how merge_sorted actually numbers the merge 
depth,
which I don't think is entirely accurate versus how a human might do it.

As such, I don't think we need to accentuate the differences, and 
instead just
a flag that decides whether only the mainline is shown, or whether 
mainline +
all merges are shown is reasonable.


That said, if someone can provide a concrete use case for "-n2" I don't 
have a
problem supporting it.


For details, see: 
http://bundlebuggy.aaronbentley.com/project/bzr/request/%3C49798D72.1060107%40internode.on.net%3E
Project: Bazaar



More information about the bazaar mailing list