#193 RFC4180Parser is broken when a multi-line quoted field contains line ends with quote

[Edward Xia]

When a multi-line quoted field contains a line (but not the last line) which ends with dquote escaped dquote, it terminates the field unexpectedly.

Fisrt, this works as expected:

a,"b""x
c",d

It get you 1 record with 3 fields, 1st field is a, 3rd field is d, and 2nd field (multi-line) is:

b"x
c

However, with a little change to the csv (removing that extra x), it will be broken:

a,"b""
c",d

It becomes 2 record, each with 2 fields:

First record is a, b". Second record is c",b. In this case it's broken.

[Scott Conway]

At first I was going to say we fixed this but looking through the history it was a different issue (bug 165). I was able to recreate this with a simple unit test and will look at fixing this for the next release.

[Scott Conway]

Okay - I have found a solution that causes the test to pass, and more importantly not cause the existing tests to fail.

I have merged it in trunk and it will go out in the next release.

[Jonathan Kulzick]

Thanks for the fix Scott. I just ran into this issue as well. Do you know roughly when the next release is planned or is there a patch or something I can apply in the meantime?

[Scott Conway]

Sorry no patches. Short term you can clone the project and build the jar file yourself.

I was going to do a release last weekend but I caught a cold. I am hoping for this weekend if life works out.

Scott :)

[Jonathan Kulzick]

Haha. Thanks Scott!

[Scott Conway]

Fix has been released please download version 4.5 and test it out.

[Jonathan Kulzick]

Thanks again! I've verified the fix works for me.

[Scott Conway]

Thanks for the quick verification.