Maxime Coste
3c265acd6c
Remove posB from information given by the diff algorithm
...
posB is always the sum of previous add len and previous keep len,
so very easy to keep track of.
2020-03-20 20:27:50 +11:00
Maxime Coste
401ef84a4b
Remove uses of reverse_iterator in diff implementation
2020-03-20 20:24:42 +11:00
Maxime Coste
60782b3dc9
Make diff implementation able to work on different iterator types
2020-02-18 19:26:01 +11:00
Frank LENORMAND
63092d12c0
src: Fix the build on Mac
...
When compiled on Mac with `clang`, the following error occurs at
compile-time:
```
./diff.hh:56:28: error: no member named 'min' in namespace 'std'
const int max_D = std::min((M + N + 1) / 2 + 1, cost_limit);
~~~~~^
```
2019-12-01 21:55:55 +01:00
Maxime Coste
4fdbf21ff8
Refactor diff to make allocating a diff vector optional
...
The diff interface now goes through a for_each_diff function that
uses a callback for each found diff.
2019-11-30 11:29:36 +11:00
Maxime Coste
bbf62d1779
diff: try to improve code readability
2018-02-09 21:31:10 +11:00
Maxime Coste
fdf7cf3944
Small code simplifications
2017-09-07 19:38:34 +08:00
Maxime Coste
d1c005dd8c
Limit diff algorithm complexity
...
Return a non-optimal, but valid, diff when we detect too many
iterations
Fixes #1509
2017-07-23 21:33:12 +02:00
Maxime Coste
d43e2ac843
More cleanups in diff code
2017-07-20 17:47:50 +02:00
Maxime Coste
7b96d56996
Use the provided equal functor for prefix/suffix detection in diff
...
We were wrongly using the `==` operator.
2017-07-18 17:53:30 +02:00
Maxime Coste
9197dd393c
More refactoring of the diff code in order to make it cleaner
2017-07-18 16:11:24 +02:00
Maxime Coste
50fec86749
Change diff Implementation to use end indices instead of length
...
Having absolute begin and relative lenght was a bit strange to
work with. Rename middle_snake to snake.
2017-07-18 12:14:52 +02:00
Maxime Coste
793c2ed9cf
Slight style change
2017-07-17 19:29:57 +02:00
Maxime Coste
d90cd6de77
Refactor find_diff_rec and detect kept prefix/suffixes early
...
Certain cases, like diffing an empty buffer with a big buffer, were
very slow, now this should get better as we will directly detect the
matching eol at the end of both buffers, end then immediatly detect
we need to add the rest of the big buffer.
We still are too slow on some general diff when there is a lot of
differences.
2017-07-17 13:30:11 +02:00
Maxime Coste
388ada8142
Remove MirroredArray for diff implementation
...
We can index native arrays negatively, so just setup V1 and V2 to
point in the middle of the work arrays and remove the need for
creating MirroredArray.
2017-07-17 19:28:52 +09:00
Maxime Coste
5eae7aacc7
Small code cleanup in diff implementation
2017-07-15 17:17:27 +09:00
Maxime Coste
2b6fc6beb7
Remove unused Diff::posA field
2017-01-29 12:49:52 +00:00
Michael Vetter
71f6ee3234
Remove trailing whitespaces
2015-11-19 16:31:00 +01:00
Maxime Coste
c1b9ea4d1e
Formatting tweak
2015-07-13 13:57:43 +01:00
Maxime Coste
9f46d75b27
Add headers guard to diff.hh along with a comment about the algorithm
2015-05-19 19:46:24 +01:00
Maxime Coste
38bbecef62
Fix bug in diff implementations (missing snake after d=1 change) and refactor
2015-05-18 22:59:59 +01:00
Maxime Coste
0a6ad4dcf4
Only initialize element 1 in mirrored arrays.
2015-05-17 20:13:11 +01:00
Maxime Coste
73ddf18dc3
Another bug fix in diff implementation
2015-05-15 13:55:39 +01:00
Maxime Coste
e9af3a4217
always_inline a few methods
2015-05-14 19:05:41 +01:00
Maxime Coste
cc97d4ba41
Fix bugs in diff implementation
2015-05-14 13:57:03 +01:00
Maxime Coste
7a8c2d7f56
Fix diff implementation and change the Diff struct format
2015-05-13 23:22:29 +01:00
Maxime Coste
49def73e4e
Add initial diff implementation based Eugene W. Myers' algorithm
2015-05-12 23:41:35 +01:00
