Failed to parse lines, starting in 0!

[LionelMarbot]

Hi,

Our diffs are generated in PHP with https://github.com/sebastianbergmann/diff. Here is an example diff:

"--- Original\n+++ New\n@@ @@\n-This is test.\n-A multiline test.\n-With some minor differencs.\n+This is a test.\n+A multi-line test.\n+With some minor differences.\n","translator_comment":"This is a test.\r\nA multi-line test.\r\nWith some minor differences."

While Diff2HtmlUI does parse the diff and shows HTML, it always throws error Failed to parse lines, starting in 0!.

Longer diffs then look like this:

As you can see in the image, the numbering is wrong. Also, further lines without diffs are still being shown. In above example, after line 2, there are no more diffs. Still, the complete file (200 more lines) is being shown.

Any idea?

Best,
Lionel

[rtfpessoa]

diff2html is prepared to parse unified diffs, as defined in http://www.gnu.org/software/diffutils/manual/html_node/Detailed-Unified.html#Detailed-Unified. Over this it can also track the extra information added by git diff. In you case you have some different version of the unified diff without base lines.

diff2html does some effort to recover from failed line identification and starts from 0. In this case it seems to not be enough. I can take a look, but since you are not providing the line info it will never have the right lines.

[rtfpessoa]

@LionelMarbot can you provide a full diff example for me to test?

[LionelMarbot]

@rtfpessoa Thank you for your quick reply.

I have set up an example here: https://testing.enuvo.ch/diff/index.html

The diff is generated from those two contents:

Original: https://testing.enuvo.ch/diff/a.txt
New: https://testing.enuvo.ch/diff/b.txt

Hope that helps!

[rtfpessoa]

@LionelMarbot, after reading the diff you gave me I cannot find any problem on my side.

diff2html is parsing the diff as best as it can.

The line number problem is related to the fact that sebastianbergmann/diff is not generating a proper unified diff.

It should return something like @@ 20 21 @@ being 20 and 21 the starting number from the old and the new file.

Regarding the whole file without changes in the end, diff2html only parses raw diff t html, and in this case the diff contains more lines that the regular diff. In the future I want to be able to collapse those lines but now it is still not possible.

Not sure if I addressed all your questions, let me know if you need any other clarification.

[LionelMarbot]

@rtfpessoa, thank you. I will try to figure out a way to add the line numbers to the diff.

[bryannielsen]

For anyone else ending up here like I did, the solution is to use the StrictUnifiedDiffOutputBuilder