Bug 1772 – (D1 only) regexp.split behavior with captures needs to be documented

I want to split columns out of a row of numbers. They may be separated by comas or by just white space. So I tried this: The splitter regexp auto re_splitter = new RegExp(r"(\s+|\s*,\s*)"); char[][] numbers = re_splitter.split(line); if input is a line like: 410.90711,352.879 The output from that is the array [410.90711,,,352.879] After a bit of debugging, it turns out the problem is the grouping in the regexp. Removing the parens fixes the problem in this case, but there are cases where you need parens for grouping and not for the capturing side effect. So I think this is a bug. Only match 0 should be considered significant for splitting, not the submatches.

fixed summary

It seems I've been duped by writefln's output. Further investigation shows that this: [410.90711,,,352.879] is actually this: ["410.90711", ",", "352.879"] and not a 4-element list with two empty strings as I thought. I discovered this because I checked what python does with captures, and it is this: """ If capturing parentheses are used in pattern, then the text of all groups in the pattern are also returned as part of the resulting list. """ So that made me think maybe D could be trying to do something similar. Apparently it is. So please just document it.

Reassigning to Dmitry.

Commits pushed to master at https://github.com/D-Programming-Language/phobos https://github.com/D-Programming-Language/phobos/commit/62b464b48d61b076c89f7585dc0ac7632f57ba49 fix Issue 1772 - regexp.split behavior with captures needs to be documented A documentation clarification, the report itself is largely outdated. https://github.com/D-Programming-Language/phobos/commit/6d782c6efd9ba6a7b7a314002a52d3455fa00d8c Merge pull request #491 from blackwhale/issue-1772 fix Issue 1772 - regexp.split behavior with captures needs to be documen...

Is this fixed/D1 only now?

I'm no expert D1 stuff, but I belive issue is still applicable for D1. Come to think of, I closed few D1 issues like this in the past, maybe we should close this one too (marked as D1 for now). D1/D2 regexp is broken in many ways and nobody is doing any work on Phobos/D1 to fix it AFIAK, Tango folks have their own regex anyway.

I guess it can be closed when D1 is discontinued at the end of the year.

It's unlikely this D1 issue will get worked on, if anyone plans to work on it feel free to reopen.