[Bug 8061] Fix meta handling for $suppl_attrib->{rule_hits}
https://bz.apache.org/SpamAssassin/show_bug.cgi?id=8061 Kevin A. McGrail changed: What|Removed |Added CC||kmcgr...@apache.org --- Comment #5 from Kevin A. McGrail --- +1 from me -- You are receiving this mail because: You are the assignee for the bug.
[Bug 8061] Fix meta handling for $suppl_attrib->{rule_hits}
https://bz.apache.org/SpamAssassin/show_bug.cgi?id=8061 spamassas...@arcsin.de changed: What|Removed |Added CC||spamassas...@arcsin.de --- Comment #4 from spamassas...@arcsin.de --- Many thanks, V2 works for us. -- You are receiving this mail because: You are the assignee for the bug.
[Bug 8061] Fix meta handling for $suppl_attrib->{rule_hits}
https://bz.apache.org/SpamAssassin/show_bug.cgi?id=8061 Henrik Krohns changed: What|Removed |Added Attachment #5842|0 |1 is obsolete|| --- Comment #3 from Henrik Krohns --- Created attachment 5843 --> https://bz.apache.org/SpamAssassin/attachment.cgi?id=5843=edit Fix rule_hits, meta declaration V2 This V2 version should fix undefined warnings that result from () returning undef. (Note, currently parser accepts all sorts of funky things like "meta FOOBAR &&", probably should be a bit more strict, but that's for another day and bug) -- You are receiving this mail because: You are the assignee for the bug.
[Bug 8061] Fix meta handling for $suppl_attrib->{rule_hits}
https://bz.apache.org/SpamAssassin/show_bug.cgi?id=8061 --- Comment #2 from Henrik Krohns --- Note, the above patch does not prevent using "meta __TRUNCATE 0" in the case of rule_hits API, it will work too. -- You are receiving this mail because: You are the assignee for the bug.
[Bug 8061] Fix meta handling for $suppl_attrib->{rule_hits}
https://bz.apache.org/SpamAssassin/show_bug.cgi?id=8061 --- Comment #1 from Henrik Krohns --- Created attachment 5842 --> https://bz.apache.org/SpamAssassin/attachment.cgi?id=5842=edit Fix rule_hits, meta declaration So the problem here is any rule that is not pre-declared in ruleset. If such rule is used in a meta as dependency, it will be considered unrun and the whole meta will not evaluate - unless of course got_hit() is called by rule_hits API. A good real-world example is __TRUNCATE rule that amavisd may dynamically hit, if a message is truncated. meta FOOBAR OTHER_RULE && !__TRUNCATE In that case FOOBAR will never be run if __TRUNCATE does not hit. In the case of rule_hits API, pre-declaring with "meta __TRUNCATE 0" could work as rule_hits is processed right at the start of Check.pm, but that does not work for other types of plugins that may dynamically hit some rule at later stage. So I propose a following format to be future-proof: meta __TRUNCATE () That would be considered a declaration, which will not be evaluated immediately (which other metas without dependencies currently do). Attached patch for your consideration. May add tests and doc later. -- You are receiving this mail because: You are the assignee for the bug.
[Bug 8061] Fix meta handling for $suppl_attrib->{rule_hits}
https://bz.apache.org/SpamAssassin/show_bug.cgi?id=8061 Henrik Krohns changed: What|Removed |Added Depends on||7735 Referenced Bugs: https://bz.apache.org/SpamAssassin/show_bug.cgi?id=7735 [Bug 7735] Meta rules need to handle missing/unrun dependencies -- You are receiving this mail because: You are the assignee for the bug.
[Bug 8061] Fix meta handling for $suppl_attrib->{rule_hits}
https://bz.apache.org/SpamAssassin/show_bug.cgi?id=8061 Henrik Krohns changed: What|Removed |Added Target Milestone|Undefined |4.0.0 CC||apa...@hege.li -- You are receiving this mail because: You are the assignee for the bug.