Make parenthesized expressions advance the limit count on preg_split #2860
Conversation
|
@tml thanks a lot! |
|
you should add a unit test which reproduces the issue this change is fixing |
|
Sure; but the fix was so painfully obvious, I figured it'd be good to get
it out there and see if anyone argues that the existing behaviour is
preferable/intentional while I brush up on how to write a pht - it's been a
while. :)
On Oct 18, 2017 12:48 AM, "Markus Staab" <[email protected]> wrote:
you should add a unit test which reproduces the issue this change is fixing
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#2860 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AACJTCCVPhLLiRSMpqIa5KnQUrqRid7Vks5stZ88gaJpZM4P9Mx2>
.
|
… count on preg_split)
|
@krakjoe I don't know how to make the Appveyor builds run, any help there? |
|
@krakjoe it seems some AppVeyor issue with the cache. I've posted to the existing ticket a day ago http://help.appveyor.com/discussions/problems/8572-cache-is-broken . I'll do some experiments yet, that might fix it. Will check with them further, anyway. Thanks. |
|
Thanks. Can we remove the waiting on author tag? |
|
@tml I've disabled the AppVeyor cache, that's the solution for now. If you pull new from the dev branch, it'll fetch the new .appveyor.yml. Thanks. |
|
The logic here does not look right to me. E.g. if you only have NO_EMPTY but not DELIM_CAPTURE, this is still going to count empty matches, even though they won't be returned, right? |
|
This appears to have been rejected. |
|
No, it was merged in commit c19d519. :) |
Perhaps the existing behaviour is preferred; if so, the documentation should make this more clear. As I view it, I think counting the parenthesized expressions feels like the right thing to do.