[PATCH] tests: add test for symlink regression fixed by ca5cc2976574

[Mads Kiilerich]

Markus Zapke-Gründemann wrote, On 07/04/2012 08:19 PM:
> Markus Zapke-Gründemann schrieb:
>> # HG changeset patch
>> # User Markus Zapke-Gründemann <markus at keimlink.de>
>> # Date 1341189237 -7200
>> # Branch stable
>> # Node ID d5c33f148540cf8e06211ce22fe2ceb78e507548
>> # Parent  2440822446ce248175dd597f0ec8df24d8d10217
>> tests: add test for symlink regression fixed by ca5cc2976574
>>
>> This is a test for a regression introduced by fcf66193b186 and fixed by
>> ca5cc2976574 for 2.1.2. The fix had a regression test only for merging.
>>
>> diff --git a/tests/test-symlinks.t b/tests/test-symlinks.t
>> --- a/tests/test-symlinks.t
>> +++ b/tests/test-symlinks.t
>> @@ -253,3 +253,25 @@ Issue995: hg copy -A incorrectly handles
>>     $ mv dirlink newdir/dirlink
>>     $ hg mv -A dirlink newdir/dirlink
>>   
>> +
>> +move file and create symlink to it at the original location, then update
>> +regression was introduced by fcf66193b186 and fixed by ca5cc2976574 for 2.1.2
>> +
>> +  $ mkdir d1 d2
>> +  $ echo c1 > d1/f1
>> +  $ hg commit -Am "add real file f1"
>> +  adding d1/f1
>> +  $ mv d1/f1 d2
>> +  $ cd d1
>> +  $ ln -s ../d2/f1
>> +  $ cd ..
>> +  $ hg commit -Am "move f1 and create symlink to original"
>> +  adding d2/f1
>> +  $ "$TESTDIR/readlink.py" d1/f1
>> +  d1/f1 -> ../d2/f1
>> +  $ hg up 3
>> +  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
>> +  $ hg up
>> +  2 files updated, 0 files merged, 0 files removed, 0 files unresolved
>> +  $ "$TESTDIR/readlink.py" d1/f1
>> +  d1/f1 -> ../d2/f1
> Any thoughts on the patch? ca5cc2976574 added also a test but I think it's
> testing merging. Therefore I submitted this patch to test the scenario
> explicitly where a user encountered the error after upgrading from 1.8.1 to 2.1.

The first thought is that an update really _is_ a merge. That is most 
obvious in daily use when you have uncommitted changes and they are 
'merged' with the revision you are updating to. A test for the merge 
case is thus sufficient and adding more tests will just make the test 
suite slightly slower without improving the test coverage.

ca5cc2976574 did however already introduce a few lines of test for 
updating from symlink to file that (unintentionally, I presume) wasn't 
run. That is now fixed in crew with 
http://hg.intevation.org/mercurial/crew/rev/722fd97f6645 .

Adding a small additional change to the test do however give a result 
that indicates that there still is an issue in this area:

--- a/tests/test-merge-types.t
+++ b/tests/test-merge-types.t
@@ -80,6 +80,28 @@

  Update to link with local change should cause a merge prompt (issue3200):

+  $ ln -sf someotherlink a
+  $ HGMERGE= hg up -y --debug 0
+  resolving manifests
+   overwrite: False, partial: False
+   ancestor: 521a1e40188f, local: 521a1e40188f+, remote: c334dc3be0da
+   a: versions differ -> m
+  preserving a for resolve of a
+  updating: a 1/1 files (100.00%)
+  couldn't find merge tool hgmerge
+  picked tool 'internal:prompt' for a (binary False symlink True)
+   no tool found to merge a
+  keep (l)ocal or take (o)ther? l
+  0 files updated, 1 files merged, 0 files removed, 0 files unresolved
+  $ hg diff --git
+  diff --git a/a b/a
+  --- a/a
+  +++ b/a
+  @@ -1,1 +1,1 @@
+  -a
+  +someotherlink
+  \ No newline at end of file
+
    $ hg up -C 0
    1 files updated, 0 files merged, 0 files removed, 0 files unresolved
    $ echo data > a

Taking 'local' should not give a file but a symlink.

/Mads