Cannnot find the processing of bounces as `hard` and `soft`
Somewhat details regarding this can be found in this thread also.
Currently, in the mailman/src/mailman/runners/bounce.py I do not see whether the separation of bounces as hard and soft is implemented. According to the previous version of Mailman, the bounce_score is increased accordingly ( see the Bounce processing section of this page ) depending upon the type of bounce. I intend to follow that implementation. If not required I can drop that implementation and just increase the bounce_score as an integer always.
Pointers on this issue are appreciated. Am I missing something here?
On Fri, Jul 19, 2019, at 10:51 PM, Aaryan Bhagat wrote:
Somewhat details regarding this can be found in this thread also.
Currently, in the
mailman/src/mailman/runners/bounce.pyI do not see whether the separation of bounces ashardandsoftis implemented.
Looking at the code in src/runners/bounce.py I don't think temporary
failures are even stored. It seems like they are just ignored.
According to the previous version of Mailman, the
bounce_scoreis increased accordingly ( see theBounce processingsection of this page ) depending upon the type of bounce. I intend to follow that implementation. If not required I can drop that implementation and just increase thebounce_scoreas an integer always.
I don't think we should be increasing the bounce_score on soft failures, which can be retried by the MTA.
Pointers on this issue are appreciated. Am I missing something here?
Mailman-Developers mailing list -- mailman-developers@python.org To unsubscribe send an email to mailman-developers-leave@python.org https://mail.python.org/mailman3/lists/mailman-developers.python.org/ Mailman FAQ: https://wiki.list.org/x/AgA3
Security Policy: https://wiki.list.org/x/QIA9
-- thanks, Abhilash Raj (maxking)
On 7/19/19 10:50 PM, Aaryan Bhagat wrote:
Currently, in the
mailman/src/mailman/runners/bounce.pyI do not see whether the separation of bounces ashardandsoftis implemented. According to the previous version of Mailman, thebounce_scoreis increased accordingly ( see theBounce processingsection of this page ) depending upon the type of bounce. I intend to follow that implementation. If not required I can drop that implementation and just increase thebounce_scoreas an integer always.
Contrary to the documentation, Mailman has never distinguished between "hard" and "soft" bounces. The recognizer used in Mailman 3, flufl.bounce, implements essentially the same process as Mailman 2.1. It distinguishes between permanent (5xx) failures and temporary (4xx) failures, but not between various kinds of permanent failures, e.g. unknown user vs full mailbox.
Mailman 2.1 has never actually scored "soft" bounces as 0.5. It scores 5xx bounces as 1.0 and ignores 4xx bounces.
The same should be done with Mailman 3. Distinguishing between "hard" and "soft" 5xx bounces would require significant changes to flufl.bounce and is outside the scope of your project.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
I understand, thanks for the reply, Mark.
So that means I can set the bounce_score_threshold as Integer instead of Float reason being that the score will always be increased by one.
Is the above inference correct? Am I missing something here?
On 7/20/19 11:45 AM, Aaryan Bhagat wrote:
I understand, thanks for the reply, Mark. So that means I can set the
bounce_score_thresholdas Integer instead of Float reason being that the score will always be increased by one.Is the above inference correct? Am I missing something here?
It mm 2.1, it has always been a float to allow for the possible implementation of fractional scores. It is still possible that fractional scores will be implemented at some point in the future so my inclination is that all these scores should be floats.
However, there is currently a contradiction in that
model/mailinglist.py has bounce_score_threshold = Column(Integer) # XXX
but
styles/base.py has mlist.bounce_score_threshold = 5.0
Whether we choose float or integer, we should be consistent in how we define and set it.
Also, bounce_score_threshold is apparently not yet exposed in REST at all.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
participants (3)
-
Aaryan Bhagat -
Abhilash Raj -
Mark Sapiro