Rob Rix
13a90624de
Stub in a quick Arbitrary instance for Child.
2016-05-20 15:34:20 -04:00
Rob Rix
72455cdecd
Add a datatype to represent children for alignBranch property tests.
2016-05-20 15:34:11 -04:00
Rob Rix
5355143395
Add some pending properties.
2016-05-20 13:05:38 -04:00
Rob Rix
018f289fb5
Abstract PrettyDiff over a type parameter.
2016-05-20 12:39:56 -04:00
Rob Rix
a2952307f6
Produce a single element with the rest of the lines in a split-up child.
2016-05-20 11:09:15 -04:00
Rob Rix
d99e4ab39b
When we only have ranges on one side, default the other to assume intersecting.
2016-05-19 16:51:33 -04:00
Rob Rix
cb2ffd188c
Add a test case of ungreedy asymmetrical node alignment.
2016-05-19 16:16:35 -04:00
Rob Rix
f1d9cc3f2f
Skip past nonintersecting children.
...
This resolves failures in two distinct cases where changing the default
for handling asymmetrical ranges had opposite outcomes.
2016-05-19 16:16:15 -04:00
Rob Rix
2fa8060a1c
Revert "Default the intersection assumption to True."
...
This reverts commit 9bf8bb93f3ddf06027e9aafd7b7062b398c5409f.
2016-05-19 16:13:43 -04:00
Rob Rix
16b1d5b4f3
Handle the empty child case with otherwise.
2016-05-19 16:12:51 -04:00
Rob Rix
0fd78af15e
🔥 the properties, since we’ve captured them.
2016-05-19 15:12:24 -04:00
Rob Rix
2df52eec1d
Ditto for the json output.
2016-05-19 14:04:16 -04:00
Rob Rix
172a0aaa47
We no longer extend an expression statement over the end of the line, apparently.
2016-05-19 14:00:22 -04:00
Rob Rix
1d2ec4c5f0
The expected alignment of this case has changed.
2016-05-19 13:54:29 -04:00
Rob Rix
c839a434e8
🔥 distribute.
2016-05-19 13:43:02 -04:00
Rob Rix
bf656592c1
🔥 pairRangesWithLine.
2016-05-19 13:42:58 -04:00
Rob Rix
f4923c63b7
🔥 spanAndSplitFirstLines.
2016-05-19 13:42:53 -04:00
Rob Rix
73746013fb
🔥 alignChildrenInRanges.
2016-05-19 13:42:46 -04:00
Rob Rix
20627f764f
Test alignBranch instead of alignChildrenInRanges.
2016-05-19 13:40:10 -04:00
Rob Rix
572082cacd
Diff Keyed nodes using alignBranch.
2016-05-19 13:38:24 -04:00
Rob Rix
527e87fc4b
Merge branch 'master' into cofree-and-bifunctors-sitting-in-a-tree-a-l-i-g-n-edly
2016-05-19 13:36:57 -04:00
Rob Rix
64ac903f3f
Align branches within a copointed context.
2016-05-19 13:26:36 -04:00
Rob Rix
ac17a13167
Test that context-only lines are paired up before the asymmetrical halves of split-up symmetrical children.
2016-05-19 12:52:51 -04:00
Rob Rix
17d2e8fe62
Default the intersection assumption to True.
...
This fixes an issue where we dropped asymmetrical children on the
opposite side from earlier asymmetrical children when splitting up
deletions/insertions.
2016-05-19 12:39:43 -04:00
Rob Rix
d25f7ddc7f
Test that we align deletions on a line before insertions.
2016-05-19 12:22:48 -04:00
Rob Rix
1979f4e32f
Make it convenient to show the raw structures for strange failure cases.
2016-05-19 12:22:33 -04:00
Rob Rix
80f58f01e3
Do some gymnastics to allow our DSL to construct terms/diffs uniformly.
2016-05-19 12:12:11 -04:00
Rob Rix
198304d45d
Use the insert/delete constructors in the test fixtures.
2016-05-19 11:10:48 -04:00
Rob Rix
7bc9a358a5
Add PatchConstructible instances to Patch/SplitPatch.
2016-05-19 11:04:19 -04:00
Rob Rix
04ef584b6c
Add a typeclass to make patches/split patches.
2016-05-19 11:03:45 -04:00
Rob Rix
3e63b464f7
Test that symmetrical nodes are split if forced by earlier symmetrical nodes.
2016-05-19 09:01:15 -04:00
Rob Rix
4f13339838
🔥 some redundant bits of the comments.
2016-05-19 00:23:04 -04:00
Rob Rix
c50493175b
Remove the redundant tests of asymmetrical nodes preceding/following symmetrical ones.
2016-05-19 00:21:41 -04:00
Rob Rix
24f98f7db9
Test that symmetrical nodes force alignment on both sides.
2016-05-19 00:19:08 -04:00
Rob Rix
17403fc2b2
Fix up the test of symmetrical nodes and following asymmetrical ones.
2016-05-19 00:18:49 -04:00
Rob Rix
b86ce458b0
Test that symmetrical nodes force alignment of later asymmetrical nodes on the same line.
2016-05-19 00:15:03 -04:00
Rob Rix
7d591e737f
Test that symmetrical nodes force alignment of earlier asymmetrical nodes on the same line.
2016-05-19 00:13:49 -04:00
Rob Rix
5c2d9b6b01
Change a test name.
2016-05-19 00:13:26 -04:00
Rob Rix
264f1cf5bd
Merge branch 'master' into cofree-and-bifunctors-sitting-in-a-tree-a-l-i-g-n-edly
2016-05-18 18:21:40 -04:00
Rob Rix
dc8d0d4fdf
Add punctuation characters to a test case to clarify the failure.
2016-05-18 17:22:16 -04:00
Rob Rix
99d0496d72
Use first/second to advance asymmetrically before symmetrical children.
2016-05-18 17:22:02 -04:00
Rob Rix
78323c8183
Swap the order of the arguments to alignChildren.
2016-05-18 16:19:40 -04:00
Rob Rix
59ca7fd08b
Fix a missing bracket in a comment.
...
This was making bracket-matcher sad.
2016-05-18 16:15:44 -04:00
Rob Rix
a92f55a053
alignChildren returns the remaining children in the nonintersecting case.
2016-05-18 16:06:15 -04:00
Rob Rix
bb38c87015
Reformat alignChildren’s cases.
2016-05-18 16:03:37 -04:00
Rob Rix
a5960a164f
Factor the recursive call of alignChildren into its where clause.
2016-05-18 16:02:38 -04:00
Rob Rix
59ff6fcccd
Handle an oxymoronic case to silence a warning (and in hope of future refactoring).
2016-05-18 15:57:32 -04:00
Rob Rix
6ce445068c
Move the symmetrical child test up a level.
...
This avoids redundancy with the does-anything-intersect-on-both-sides
case.
2016-05-18 15:51:49 -04:00
Rob Rix
5e1d111cac
Move the asymmetrical computation break into the where clause.
2016-05-18 15:44:04 -04:00
Rob Rix
43a893e843
lineAndRemainingWhere doesn’t imply any.
2016-05-18 15:37:28 -04:00
