Changes to the ticket system
I've made a few changes to the ticket system today, in preparation for a push on cleaning up the ticket database (announcement to be made shortly). The changes are: 1. There's a new field "Type of Failure", with values None/Unknown Build failure Install failure GHC doesn't work at all Compile-time crash Runtime crash GHCi crash Incorrect result at runtime Compile-time performance bug Runtime performance bug Hopefully the values are self-explanatory. Setting this correctly will help us to focus on important tickets, and may help us automatically prioritize tickets in the future. 2. The "Severity" field has gone away. This field was intended to be used as a way for ticket submitters to communicate to us the importance of the bug to them, but in practice it wasn't clear how to use it, it hasn't been used consistently, and as a result we generally ignore it. The new "Type of Failure" is a much less ambiguous way to communicate some of the same information. 3. The "Difficulty" field has more sensible values representing a range of times, rather than just "1 hour" or "1 week": Unknown Easy (less than 1 hour) Moderate (less than a day) Difficult (2-5 days) Project (more than a week) Rocket Science Cheers, Simon
Simon Marlow wrote:
I've made a few changes to the ticket system today, in preparation for a push on cleaning up the ticket database (announcement to be made shortly). The changes are:
1. There's a new field "Type of Failure", with values
None/Unknown Build failure
meaning failed building GHC, not that GHC building something failed.
Install failure GHC doesn't work at all Compile-time crash Runtime crash GHCi crash Incorrect result at runtime Compile-time performance bug Runtime performance bug
what about... Documentation is wrong. Type system is wrong (eg. rejects something that's correct), or similarly for syntax. Something strange that doesn't deserve its own category, but isn't "unknown"
On 16/11/2009 20:52, Isaac Dupree wrote:
Simon Marlow wrote:
I've made a few changes to the ticket system today, in preparation for a push on cleaning up the ticket database (announcement to be made shortly). The changes are:
1. There's a new field "Type of Failure", with values
None/Unknown Build failure
meaning failed building GHC, not that GHC building something failed.
Right. Changed to Building GHC failed
Install failure
Changed to Installing GHC failed
GHC doesn't work at all Compile-time crash Runtime crash GHCi crash Incorrect result at runtime Compile-time performance bug Runtime performance bug
what about...
Documentation is wrong.
Added "Documentation bug"
Type system is wrong (eg. rejects something that's correct), or similarly for syntax.
Good point: added "GHC rejects valid program"
Something strange that doesn't deserve its own category, but isn't "unknown"
such as? Cheers, Simon
Hi Simon,
Something strange that doesn't deserve its own category, but isn't "unknown"
such as?
I guess the distinction is between "I don't know in which of these categories the bug falls" and "I know that it does not fall in any of the above categories". e.g., "Unknown" vs "Other" ? Gr. Matthijs
On 17/11/2009 12:45, Matthijs Kooijman wrote:
Hi Simon,
Something strange that doesn't deserve its own category, but isn't "unknown"
such as?
I guess the distinction is between "I don't know in which of these categories the bug falls" and "I know that it does not fall in any of the above categories". e.g., "Unknown" vs "Other" ?
Ok, I've added "Other". On balance I guess it's useful to know that "we've looked at this bug and decided that it doesn't fall into any of the other categories" vs. "it hasn't been assigned a category yet". Cheers, Simon
participants (3)
-
Isaac Dupree -
Matthijs Kooijman -
Simon Marlow