TutorialConflict

Tutorial - Merging conflicting changes

(This page is part 8 of 9 of the [:Tutorial] series. Previous part is [:TutorialMerge], next part is [:TutorialConclusion])

We learned how to deal with simple [:Merge:merges] in TutorialMerge.

Mercurial handles more complex merge cases, too. It is not all that uncommon for two people to edit the exact same lines of a file, and then have to figure out what to do. These cases are called [:Conflict:conflicts]; figuring out what to do about a conflict is called [:ResolveConflict:resolving] it.

Let's first create an artificial conflict situation. As we did previously, let's start by making a [:Clone:clone] of my-hello:

$ cd ..
$ hg clone my-hello my-hello-not-cvs
updating working directory
2 files updated, 0 files merged, 0 files removed, 0 files unresolved

Now let's add a new line of output to hello.c:

$ cd my-hello-not-cvs
$ vi hello.c

We change main to read like this:

int main(int argc, char **argv)
{
        printf("hello, world!\n");
        printf("sure am glad I'm not using CVS!\n");
        return 0;
}

And we [:Commit:commit] the change:

$ hg commit -m "Give thanks for dodging bullet"

Recall that in TutorialFirstChange, we created a [:ChangeSet:changeset] in my-hello-new-output that also added a second line of output. What happens when we try to [:Pull:pull] that change in here?

$ hg pull ../my-hello-new-output
pulling from ../my-hello-new-output
searching for changes
adding changesets
adding manifests
adding file changes
added 1 changesets with 1 changes to 1 files (+1 heads)
(run 'hg heads' to see heads, 'hg merge' to merge)

So far, so good. Let's try an [:Update:update].

$ hg update
abort: crosses branches (use 'hg merge' or 'hg update -C')

As in [:TutorialMerge], we have to run hg merge. It will not be able to merge automatically, because the same line of the same source file has been modified in a different way by each changeset (the one we just [:Commit:commited], and the one we just pulled).

$ hg merge

At this point, what happens depends on how Mercurial is configured (see [:MergeToolConfiguration]). Per default, Mercurial drops into the editor marking the conflicts in the file for manual resolution:

/*
 * hello.c
 *
 * Placed in the public domain by Bryan O'Sullivan
 *
 * This program is not covered by patents in the United States or other
 * countries.
 */

#include <stdio.h>

int main(int argc, char **argv)
{
        printf("hello, world!\n");
<<<<<<< /home/adi/tmp/tutorial/my-hello-not-cvs/hello.c
        printf("sure am glad I'm not using CVS!\n");
=======
        printf("sure am glad I'm using Mercurial!\n");
>>>>>>> /tmp/hello.c~other.2xAVqv
        return 0;
}

In the editor, we delete the conflict markers and keep the "sure am glad I'm using Mercurial!\n" line, deleting the line about CVS. When we leave the editor, Mercurial asks:

Was the merge successful? [y/n]

Here, we answer "y". Mercurial now reports the summary of the manual merge process:

0 files updated, 1 files merged, 0 files removed, 0 files unresolved
(branch merge, don't forget to commit)

As before, be sure to commit this change to the repository once the merge is complete:

$ hg commit -m "Merged changes from my-hello-new-output"

What we have seen here is the default behaviour of Mercurial. However, Mercurial can be configured to call external three-way merge tools. Information about configuring three-way merge tools can be found at [:MergeToolConfiguration].

Now let's continue and finish on to [:TutorialConclusion].

CategoryTutorial