Hello all,
I recently encountered a problem where the Hg automerge inserted some
duplicate code into the file.
Basically the inputs were like this:
..long stretch of identical code..
..line A of identical code except for a change in a function name..
..two lines of identical code..
..line B of identical code except for a change in a function name..
..block of identical code..
..long stretch of identical code..
and the result of automerge was that line A and B are merged
properly, but the 'block of identical code' gets duplicated.
It appears this is due to some complex revision history [partly
involving changes to the 'block of identical code'] since the problem
doesn't occur if I get rid of the history.
I can't post the code here, but it appears that a similar situation
is described at
http://stackoverflow.com/questions/5003577/how-safe-are-automatic-merges-in-mercurial
In that Stackoverflow discussion it mentions that the _default_
context is three lines wide, and I'm thinking that our problems might be
lessened with more context. (though i realize we might get more prompts
for manual merges as well)
So my question: is there some way to specify an increased number of
lines to use for context?
The only option I could find was for the diff tool, but the help
explicitly says this is only for the diff tool.
thanks for any help!
Daniel Faken
_______________________________________________
Mercurial mailing list
Mercurial@mercurial-scm.org
https://www.mercurial-scm.org/mailman/listinfo/mercurial
