Undefined offset in Squiz.Strings.ConcatenationSpacing during live coding

[thiemowmde]

This can happen when PHPCS runs on a file that is currently being worked on, but not yet completed. The file might end with a dot. We can not assume there are always 2 more tokens after a dot.

[jrfnl]

Hi @thiemowmde I see what you are trying to do here. I'd like to suggest a change though, and that is to bow out completely when $tokens[($stackPtr + 1)] isn't set as doing any checking for the space after the concatenation operator if there isn't a next token is futile.

What do you think ?

[thiemowmde]

Nice catch. Indeed, my original proposal would create awkward reports that don't make much sense if the line of code is not even complete. I changed it to don't report anything in such a situation.

[jrfnl]

I changed it to don't report anything in such a situation.

That's not what I intended. The "spacing before" could still be safely reported/fixed in that case. It's only the spacing after where the problem lies.

[thiemowmde]

There is no separate message for "spacing before". There are only messages that talk about "Concat operator must [not] be surrounded by …". All these messages are confusing when the line of code that ends with something like $s = '…' . is not even complete. Sure, we could introduce separate messages that only talk about the "spacing before" in such a situation. But that would be quite some overkill.

[jrfnl]

@thiemowmde You're right. Sorry, I was getting confused by the metrics which are recorded separately
Bit surprising nonetheless as most of these types of sniffs would have separate error codes, but as this one doesn't, I agree, no need to change that at this time.

[gsherwood]

Thanks a lot for this PR