Ruediger Pluem wrote:
Sounds like a good idea to me to work this either into the setaside code on apr-util side or into ap_save_brigade
++1 to both... ap_save_brigade_ex ? For setaside, of course, we could have a setaside_ex member, but it seems to me we can have a common struct that persists as datum on the pool, itself, consisting of setaside and setaside_limit members. setaside would accumulate the size of the setaside buckets, and setaside would fail if an allocation was required and the setaside_limit would be exceeded. Thoughts?
But I guess for backwards compatibility we need to retain the old versions of setaside / ap_save_brigade that simply call the new ones with a limit of 0 (=unlimited).
Well, adding this as a datum to the pool (assuming unlimited if it's not found) means that we could implement, in theory, in 1.2.x/0.9.x. Adding any apr _ex function would require 1.3.0, or 0.9.x which doesn't have the same versioning rules. For httpd, 1.2.3 can include this if ap_save_brigade_ex bumps the mmn. ap_save_brigade_ex would know the rules for this operation and how to manipulate the pool datum, or use the accessor calls against a pool to define setaside limits. Bill
