further refine Handler validation #695

pjenvey · 2016-10-12T22:03:19Z

Improvements we can do here:

o how we pass validation results to handlers. @threaded_validate currently attaches valid_input to the handler obj yet continues calling the handler method w/ unvalidated kwargs (which we always ignore and are thus confusing/dangerous to even keep around).

We should probably pass **valid_input as the kwargs instead

o We can also pass thru richer objects (like a fully constructed WebPushNotification object for e.g. in web/message.py). Either as fields.Raw or in a @post_load (not clear to me yet what works better)?

jrconlin · 2016-10-12T22:05:28Z

I'd prefer if we used fields.Raw. passing back as part of @post_load feels too "magical" for my tastes. We declare the fields at the top of the class and should stick to that so we don't lose data.

- params -> router_data, kill unused vapid_info/body - kill bogus/unused timing vars - app_id isn't optional issue #695

- try to isolate it to the deferred definition and avoid passing args back and forth when not necessary - also put all_channels call back in a thread issue #695

- params -> router_data, kill unused vapid_info/body - kill bogus/unused timing vars - app_id isn't optional issue #695

- try to isolate it to the deferred definition and avoid passing args back and forth when not necessary - also put all_channels call back in a thread issue #695

- params -> router_data, kill unused vapid_info/body - kill bogus/unused timing vars - app_id isn't optional issue #695

- try to isolate it to the deferred definition and avoid passing args back and forth when not necessary - also put all_channels call back in a thread issue #695

- instead of valid_input call handler methods w/ the resulting dict as args/kwargs (similarly to how cyclone does) - bundle router info together in registration to ease passing it around - kill the now unneeded MessageSchema args/kwargs shuffle and its associated test_delete_topic_success2 - fix LogCheckSchema typo closes #695

more accurately reflects what's happening now and cyclone's path_args/kwargs are exactly the same as its calling args/kwargs anyway issue #695

- instead of valid_input call handler methods w/ the resulting dict as args/kwargs (similarly to how cyclone does) - bundle router info together in registration to ease passing it around - kill the now unneeded MessageSchema args/kwargs shuffle and its associated test_delete_topic_success2 - fix LogCheckSchema typo closes #695

more accurately reflects what's happening now and cyclone's path_args/kwargs are exactly the same as its calling args/kwargs anyway issue #695

pjenvey · 2017-04-13T21:04:22Z

While shaking this out it seemed like @post_load is maybe more appropriate (webpush.py already uses it). fields.Raw would require the @post_loadish code to live inside of the @validates_schema methods most of the time (and we already have some larger ones).

Almost tempted to document what the schema produces to the handler in schema class docstring to make it very clear

- instead of valid_input call handler methods w/ the resulting dict as args/kwargs (similarly to how cyclone does) - bundle router info together in registration to ease passing it around - kill the now unneeded MessageSchema args/kwargs shuffle and its associated test_delete_topic_success2 - fix LogCheckSchema typo closes #695

more accurately reflects what's happening now and cyclone's path_args/kwargs are exactly the same as its calling args/kwargs anyway issue #695

- instead of valid_input call handler methods w/ the resulting dict as args/kwargs (similarly to how cyclone does) - bundle router info together in registration to ease passing it around - kill the now unneeded MessageSchema args/kwargs shuffle and its associated test_delete_topic_success2 - fix LogCheckSchema typo closes #695

pjenvey added this to the PUSHSVC-0: quality milestone Oct 12, 2016

pjenvey self-assigned this Oct 12, 2016

bbangert added ready p2 labels Oct 20, 2016

bbangert added p1 and removed p2 labels Nov 14, 2016

pjenvey added p3 and removed p1 labels Feb 9, 2017

pjenvey added in progress and removed ready labels Apr 3, 2017

pjenvey added a commit that referenced this issue Apr 4, 2017

refactor: various cleanup

ce32a6c

- params -> router_data, kill unused vapid_info/body - kill bogus/unused timing vars - app_id isn't optional issue #695

pjenvey added a commit that referenced this issue Apr 4, 2017

refactor: cleanup some deferred arg bookkeeping

07f8178

- try to isolate it to the deferred definition and avoid passing args back and forth when not necessary - also put all_channels call back in a thread issue #695

pjenvey added a commit that referenced this issue Apr 4, 2017

refactor: various cleanup

f6add5e

- params -> router_data, kill unused vapid_info/body - kill bogus/unused timing vars - app_id isn't optional issue #695

pjenvey added a commit that referenced this issue Apr 4, 2017

refactor: cleanup some deferred arg bookkeeping

ce318e2

- try to isolate it to the deferred definition and avoid passing args back and forth when not necessary - also put all_channels call back in a thread issue #695

pjenvey added a commit that referenced this issue Apr 4, 2017

refactor: cleanup some deferred arg bookkeeping

024b9d5

- try to isolate it to the deferred definition and avoid passing args back and forth when not necessary - also put all_channels call back in a thread issue #695

pjenvey added a commit that referenced this issue Apr 4, 2017

refactor: cleanup some deferred arg bookkeeping

2175d7d

- try to isolate it to the deferred definition and avoid passing args back and forth when not necessary - also put all_channels call back in a thread issue #695

pjenvey mentioned this issue Apr 4, 2017

refactor: cleanup some deferred arg bookkeeping #859

Merged

pjenvey added a commit that referenced this issue Apr 4, 2017

refactor: cleanup some deferred arg bookkeeping

087a382

- try to isolate it to the deferred definition and avoid passing args back and forth when not necessary - also put all_channels call back in a thread issue #695

pjenvey added a commit that referenced this issue Apr 5, 2017

refactor: cleanup some deferred arg bookkeeping

4c22d71

- try to isolate it to the deferred definition and avoid passing args back and forth when not necessary - also put all_channels call back in a thread issue #695

pjenvey added a commit that referenced this issue Apr 6, 2017

refactor: various cleanup

013db1c

- params -> router_data, kill unused vapid_info/body - kill bogus/unused timing vars - app_id isn't optional issue #695

pjenvey added a commit that referenced this issue Apr 6, 2017

refactor: cleanup some deferred arg bookkeeping

55e91c5

- try to isolate it to the deferred definition and avoid passing args back and forth when not necessary - also put all_channels call back in a thread issue #695

pjenvey added a commit that referenced this issue Apr 13, 2017

refactor: path_args/kwargs -> args/kwargs

fa7ac22

more accurately reflects what's happening now and cyclone's path_args/kwargs are exactly the same as its calling args/kwargs anyway issue #695

pjenvey added a commit that referenced this issue Apr 13, 2017

refactor: path_args/kwargs -> args/kwargs

8276f2c

more accurately reflects what's happening now and cyclone's path_args/kwargs are exactly the same as its calling args/kwargs anyway issue #695

pjenvey added p1 and removed p3 labels Apr 13, 2017

pjenvey modified the milestones: 2017Q2-2.2 Four code refactor issues completed and obsolete legacy code removed..., PUSHSVC-0: quality Apr 13, 2017

pjenvey mentioned this issue Apr 13, 2017

refactor: refine the Handler validation #870

Merged

pjenvey added a commit that referenced this issue Apr 13, 2017

refactor: path_args/kwargs -> args/kwargs

4f46687

more accurately reflects what's happening now and cyclone's path_args/kwargs are exactly the same as its calling args/kwargs anyway issue #695

bbangert closed this as completed in #870 Apr 14, 2017

bbangert removed in progress labels Apr 14, 2017

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

further refine Handler validation #695

further refine Handler validation #695

pjenvey commented Oct 12, 2016 •

edited by bbangert

Loading

jrconlin commented Oct 12, 2016

pjenvey commented Apr 13, 2017 •

edited

Loading

further refine Handler validation #695

further refine Handler validation #695

Comments

pjenvey commented Oct 12, 2016 • edited by bbangert Loading

jrconlin commented Oct 12, 2016

pjenvey commented Apr 13, 2017 • edited Loading

pjenvey commented Oct 12, 2016 •

edited by bbangert

Loading

pjenvey commented Apr 13, 2017 •

edited

Loading