Bugzilla Problems, Issues, Challenges & Ideas
![/!\](/moin_static1911/modern/img/alert.png "/!\"){kind=small}
this page is quite stale: last edited 2016-07-21
List is unordered, and is not indicative of importance or priority. Note that is also intentionally light on solutions
- No way to formally declare/track things as upstream other than URL
- Encourage upstreaming, root-cause resolution, less maintenance downstream
- Importance = Priority / Severity unused, not well defined
- Can't maximise/ensure efficient use of limited people/time/resources
Security issue responsibility ambiguity between Secteams & Maintainers/Committers
Dedicated Security components in Base / Ports & Packages products
- Clearer single path/location for security reports. Dependent issues created by secteams (eg: vuxml + port update, SA + comms)
- Codification of informal hints that are applied now (post-hoc annotation like keyword: security, cc: secteam)
- Would need to extend auto-assigner to these components (for parsing category/portname, maintainer assignment, etc)
- Is a specific example of a broader issue of higher priority issue classes
- Change (patch/attachment) level things are issue specific (mfc, merge-quarterly)
- Change them to attachment flags
- Feedback is often needed from non maintainers, but only maintainer-feedback flag exists
Rename to needs-feedback ? <value>
- Do we lose desired specificity of a specific maintainer-feedback flag? Not sure.
- QA confirmation/results are currently informal, adhoc, done in attachments or comments
- Codify it: Attachment flags? testport: +, portlint: +
- With more flags, we need to make them more obvious and easy to edit
- Resolutions: Some overlap, are unclear, ambiguous, could be better or are not covered
Rejected -> Denied
- Overcome by Events, what?
- Not Enough Information subset of Unable to Reproduce?
Not a Bug <> Works As Intended?
- Timeouts (Maintainer, Feedback) are handled manually, and inconsistently applied (committers, non-committers)
- Note: consider that un-assigning could infact delay progression in some/many circumstances
Base System Product components is incomplete, but we cant have a component per kernel component (eg: per driver/chipset)