-
Improve log viewer - it might not work very well with a lot of requests in a day.
-
Consider backing up logs to s3 as soon as they are uploaded - right now termination of the EC2 means all of the recent logs get lost.
-
CrashReportSection#additionalInfo displaying
-
There seems to be a lot of inefficiencies.
- Elements being rendered multiple times
- mappings being requested too many times
- Mapping happening too many times: mappingViaProviderChain is calling stuff multiple times for the same mappable
-
Provide a way to reset rate limiting on a specific IP gated behind auth
-
Try to migrate to vite again
-
Crashalytics?
-
'Copy Link' button
- Link expiry date
-
Mobile support:
- Share button
-
Light mode support
-
Authentication solution
- Allow crash log expiry
-
Improve 'Mods' panel
- Click each mod to reach their website
- Button to create new issue if github issues is available
- Improve visibility of 'causing mod' to increase the chance an issue will be made
-
Crash discussion
- Make sure crashes are easily searchable in Google
- Built-in crash search system (so you can search for an exception and see other people getting it)
- find some solution for secret info getting released, maybe make crashes private by default and have a button to make them public
- Add a system for a 'solution' to be posted on the crash (e.g. update to version X, remove mod Y).
-
'Click to copy'
- Crash context
- Crash title
- Entire stack trace (copying will give you the raw trace)
- Trace line
- Trace title
- Mod
- Entire section
- Section detail
-
Configurable UI
-
Explore crashes
-
Add 'upload to crashy' to forge init crash screen
-
Expose and document API
-
'New Crash' button (maybe home and 'new crash' will be seperated in the future)
-
Parse vanilla crashes
-
When verifying fails, add a button 'report failure' to automatically Log failures in crash report parsing, so we can identify ourselves more crashes that are not being parsed correctly.
-
Stricter verifying to find cases where verifying succeeds but crash log looks completely wrong, possible by checking strings exist when skipping them. Because that is slower maybe add a 'strict' flag that is only used on the server.
-
Nicer Mods screen:
- Group mods by parent (i.e. 30 fabric api mods shouldn't obscure the list)
- Make suspected mods more prominent, and explain why they were blamed (in the future it could be mixins)
- Add something in the main screen that makes suspected mods more discoverable
-
Log failures in crash report parsing for decently large reports, so we can identify ourselves more crashes that are not being parsed correctly.
-
When a failure happens, we always log the exception itself, with optionally logging the entire crash report.
-
Stricter verifying to find cases where verifying succeeds but crash log looks completely wrong, possible by checking strings exist when skipping them. Because that is slower maybe add a 'strict' flag that is only used on the server. (enable strict in server)
-
Suggest fixes. For example, when a mod is blamed, suggest updating it (maybe even somehow detect it has an update)
-
Pay a flat amount to save a crash report forever.