PATCHES ACCEPTED TO BACKPORT FROM TRUNK:
[ start all new proposals below, under PATCHES PROPOSED. ]
- * mod_proxy: Make balancers server-specific, as they should have
- been. Inheritance causes too many behind-the-scene interactions
- to be reliable in a dynamic environ.
- trunk patch: http://svn.apache.org/viewvc?view=revision&revision=1387603
- http://svn.apache.org/viewvc?view=revision&revision=1388029
- 2.4.x patch: trunk patch works (minus ap_mmn.h change)
- +1: jim, druggeri
- rjung: +1 if fixing copy-and-paste error set_persist -> set_inherit
- rjung: Please adjust compatibility note.
- Because of the config incompatibility I suggest we add
- a note either to the release announcement or CHANGES.
- When we did the same change for mod_jk years ago it took
- very long for users to understand that workers defined in the global
- server where no longer working in their vhosts (and yes, there was
- a switch there to force old behavior too).
-
* mod_auth_form: Support the expr parser in the
AuthFormLoginRequiredLocation, AuthFormLoginSuccessLocation and
AuthFormLogoutLocation directives.
+1: jim
rjung: the combined 2.4 patch contains lots of unwanted "inherit" changes.
+ * mod_proxy: Make balancers server-specific, as they should have
+ been. Inheritance causes too many behind-the-scene interactions
+ to be reliable in a dynamic environ.
+ trunk patch: http://svn.apache.org/viewvc?view=revision&revision=1387603
+ http://svn.apache.org/viewvc?view=revision&revision=1388029
+ 2.4.x patch: trunk patch works (minus ap_mmn.h change)
+ +1: jim, druggeri
+ -1: rjung
+ rjung: This is about balancers *and* workers: doesn't the change also influence
+ the inheritance of simple workers defined via ProxyPass? So I'd prefer
+ WorkersInherit as the name instead of BalancersInherit or something
+ more neutral (plus docs adjustment).
+ rjung: also needs fixing copy-and-paste error set_persist -> set_inherit
+ rjung: Please adjust compatibility note.
+ Because of the config incompatibility I suggest we add
+ a note either to the release announcement or CHANGES.
+ When we did the same change for mod_jk years ago it took
+ very long for users to understand that workers defined in the global
+ server where no longer working in their vhosts (and yes, there was
+ a switch there to force old behavior too).
+
* mod_auth_form: Make sure that get_notes_auth() sets the user as does
get_form_auth() and get_session_auth(). Makes sure that REMOTE_USER
does not vanish during mod_include driven subrequests.