da3f0325a2
Grouping: expose the recent data-modelling work in the UI
2024-04-08 16:17:45 +02:00
e98dddf2c4
Use a denormalized field (event_count) in the admin that's been available for a while
2024-04-08 15:34:28 +02:00
cc3a119998
Remove TODO; I stopped doubting (for now)
...
issue.title is based on the first event, but if this ever becomes a problem
I should probably deal with it by making issue.title better (generic over the
events or even editable) and not by making an issue-page which changes its
display based on which event is selected
2024-04-08 15:33:07 +02:00
652823f8c3
Store calculated type and value on issue and event and use these values in the templates
2024-04-08 15:30:41 +02:00
729a4c7ea1
Make <no transaction> explicit;
...
and more moving-around-of-code in preparation for our next step
2024-04-08 15:03:02 +02:00
cb75d318af
Remove the event_type_name from the grouper
...
it adds very little, especially now that we've put this information in the type
('Log Message') and removing this allows us to refactor more freely
2024-04-08 14:52:09 +02:00
8d17e7b128
For log messages, demote the msg into the value
...
and make the type 'Log Message'.
This may just be a matter of personal taste, but I've always found these
messages-as-type super-confusing. I'd rather have some made up (but clear,
and thanks to the whitespace unlikely-to-otherwise-exist) type "Log Message".
This is also inline in what I've already done in the UI.
2024-04-08 14:48:39 +02:00
a70ac7e1cb
Factor out the idea (exception -> type_, value) what we have
2024-04-08 14:39:14 +02:00
77209bde5d
truncatechars: just use the Django version
2024-04-08 14:36:34 +02:00
48307daa0f
Introduce 'Grouping' data-modeling
2024-04-08 11:41:15 +02:00
226415ce43
Further simplications in ErrorEvent.get_title()
2024-04-05 16:03:10 +02:00
7caf6bbac4
docstring
2024-04-05 15:50:06 +02:00
9dfae5a829
Use the diamond-separator when generating fingerprint-based groupers too
2024-04-05 15:41:12 +02:00
54e8009e73
Reorganize/cleanup title-generating code for clarification
2024-04-05 15:37:24 +02:00
301c6416ea
Remove class-hierarchy (in preparation for further cleanup)
2024-04-05 15:05:00 +02:00
80919d01b8
Remove the concept of 'culprit' from the grouping code
...
This has been deprecated in Sentry in 2016; no reason for us to support it.
It is noted that despite the deprecation (in favor of `transaction`) an
explicitly provided culprit actually took a higher precedence than an
explicitly provided transaction in the now-removed code.
2024-04-05 15:02:56 +02:00
eaa5ee6838
Tests for existing grouping functionality
2024-04-05 12:10:51 +02:00
2011006e74
Remove 'TODO' that was apparently already acted on
2024-04-05 11:13:08 +02:00
83c55e5f12
Remove unused base template
2024-04-05 10:13:56 +02:00
0729523cde
Guess the pygments Lexer (using the filename)
2024-04-05 09:59:11 +02:00
aca522ef8f
Document thoughts on the design of breadcrumbs
2024-04-05 09:14:06 +02:00
e263bcb96f
'No info available' for breadcrumbs
2024-04-05 09:10:33 +02:00
dfa92a5725
Comment-out trace info from detail view
2024-04-04 22:00:46 +02:00
82251814e9
minor textual
2024-04-04 21:58:36 +02:00
098bc3dbc9
Detail page: margins
2024-04-04 21:55:37 +02:00
0747a808b9
(custom) tags on the detail page
2024-04-04 21:53:12 +02:00
97c1df4f83
Document thoughts about log messages
2024-04-04 21:35:14 +02:00
be2a2157e5
'no detail info'/'no stacktrace' implemented
2024-04-04 21:19:36 +02:00
b0760c8f1e
Fix non-displayed descenders of 'g'
...
as per https://stackoverflow.com/questions/11161216/why-does-overflow-xhidden-clip-my-descenders
there does not actually seem to be a good solution for this; I'm resorting to 'just add padding'
2024-04-04 21:05:31 +02:00
9870acf6d2
show raise/show in-app jump-to-location
...
TODO: when multiple buttons are pressed in quick succession, the transitions may create some
race-condition
2024-04-04 16:18:44 +02:00
5510002a47
show text 'line' only when there is a lineno
2024-04-04 16:00:46 +02:00
e7f67e9016
Clarify that 'no code or variables' is about a given frame
...
textually, but also by just left-aligning (for some reason that signals 'this frame' to me)
2024-04-04 15:57:59 +02:00
d94bfa8aa6
Log Messages: my first take
...
they should somehow show up in the title; in the interface it should be clear that we're
dealing with log messages (rather than exceptions)
2024-04-04 15:40:31 +02:00
7064287395
Details page: less harsh borders
2024-04-04 13:43:00 +02:00
b77f16e2db
Margin fix (target: the details page)
2024-04-04 13:36:29 +02:00
b17cfd8c43
Anchor tags in the details page
2024-04-04 13:34:54 +02:00
b9e13ef52b
details page: show relevant info, in a non-fugly way
2024-04-04 13:30:19 +02:00
5c2686a10e
Event admin: search
2024-04-04 12:34:52 +02:00
fbd94976bd
Add default ordering (-timestamp) on Events in admin
2024-04-04 12:33:34 +02:00
33f530c2ea
Slightly more margin below vars
2024-04-04 12:31:12 +02:00
09a716a8d7
Stacktrace ordering: stack-of-plates as an option
...
implemented in-python; doing this using flex quickly became a mess
2024-04-04 10:40:32 +02:00
2854d6f3ee
Labels for points-in-time in the stack
2024-04-04 09:26:38 +02:00
d4d0601bbe
margins/padding of code/variables
...
designed given 4 kinds of frame: with/without both code & vars
2024-04-03 11:48:15 +02:00
f37d7b7c0e
guard against issue-without-event-creation
2024-04-03 11:29:16 +02:00
9acff514ea
Fix --fresh-id param on send_json command
...
probably never worked, the right attr is simply 'event_id'
2024-04-03 11:21:37 +02:00
10785c1838
When no context_line is available, don't display code sample (again, fixed)
2024-04-03 10:54:33 +02:00
d3afb83dec
pygments css: remove stuff related to background colors
...
pygments can do the foreground, bu let me (tailwind) take care of the backgrounds please
2024-04-03 10:53:18 +02:00
90197fccc1
When no context_line is available, don't display code sample
2024-04-03 10:31:11 +02:00
eba7f1c628
When no line numbers are available, don't display them
...
(Whether this is needed in practice is unsure; I implemented this
before I noticed that in such cases frame.context_line is also
always missing in my data-set; but I havent' checked the spec and
leaving it in defensively really won't hurt)
2024-04-03 10:29:40 +02:00
f40ee395ee
Make issues/templatetags more robust
...
(the need for which was exposed by the recently added tests)
2024-04-03 10:15:13 +02:00
Klaas van Schelven