+---+
| Bugzilla Bug ID |
| +-+
| | Status: UNC=Unconfirmed NEW=New ASS=Assigned
In a nutshell the long term goal has always been to get the c parts of apreq
incorporated into httpd distributions so the perl parts can ship with modperl.
This is still along those lines. In order to continue to expose the cool cgi
code that Issac added to libapreq we need to ensure there is
+1
If/before we do this, maybe we can at least agree on
skiplist_insert/add() vs skiplist_insert/addne().
I known it depends on whether or not names will finally be rollbacked
in APR, but we'd better not copy skiplist's code from there before
this is decided.
Just to avoid having different
My 2c is that I don't like the idea that add and insert
means 2 different things with 2 different behaviors. I
know we were kinda forced into that due to me, mistakenly,
not realizing that dups were the *compliant* impl.
The real rub is what do we call place into skiplist
unless it would
I'm +1 on that.
rant
I'll be honest, I think that in several ways the lag
between httpd and APR is putting httpd at risk. Every
possible change to APR is being held-up by, imo, irrational
concepts of what breaks the API. Now this wouldn't be
so bad if we saw releases of APR more often than every
Looking back, I think that apr_skiplist wasn't ready for general use (both
doc and code) when it was put in APR and released, and at the same time it
was unfortunate that it placed a prereq on a new APR release in order to
use Event, introducing another speedbump to using httpd's latest and