Harro Verton

Great. Not. The problem is that in the QUERY_STRING, the first "=" is also URL encoded. I don't know if this is a bug in PHP or something that has...

Attempt two. I used this URL for testing: ```` http://19develop.dev/welcome/index/a%20space/a+plus/?Auth=DBUW1Dc%2bKwPn7eE%2bZuxzg2fnlww%3d&date=2013-10-17T22%3A09%3A24%2B00%3A00 ```` resulting in thse URI segments: ```` array (size=4) 0 => string 'welcome' (length=7) 1 => string 'index' (length=5) 2...

It shouldn't. I went through the commit history, appearently the decoding/encoding was added to address some nginx quirk. But the examples given are handled buy this code as well. I'm...

It is indeed to use the same soap object for multiple requests, instead of forging a new one for each request. If you pass "trace" as one of the options...

It would, yes. But that isn't really a regex for a float, that would also match ",.,.,.,.". A better one would be something like: ````^[-+]?[0-9]*[\.,]?[0-9]+([eE][-+]?[0-9]+)?$```` ?

The fact that a developer is using a framework doesn't mean the develop shouldn't know how PHP works, i.e. what the effect is of using the ENT_HTML5 flag, it isn't...

I think you do. I've checked all drivers, and the apc and redis drivers contain the same code. My initial reaction, without testing, is that for these three the prefixing...

Correction, I think the cache_id shouldnt be there, so L149 should be ```` // determine the section index name $section = empty($section) ? '' : '.'.$section; ```` Are you able...

It was never in the design that a module route could overload a global route. For that reason, all module routes have to be prefixed with the module name for...

Ah, yes, an empty cache id would hide the bug... :smirk: