I think we also miss "Performance regressions". Those also should get high priority. We should at least make sure they don't stick too long in the tree. The reason for this is that it could obscure other regressions or in the worst case scenario be present on merge date. Making it hard to track and have to big change it only gets fixed on trunk and not backported.
Normally people get some small time to respond (2-3 days). During merge/multiple regressions one can respond more actively by just backing out. So since this list is about SpiderMonkey priorities it might be good to add those too. I don't know what the best way would be. The process of fixing them can be very organic, so only listing bugs could be hard. We currently have three active ones: - 5% Octane/kraken regression on mac OSX only on GGC. (bug 987047) - 4% Octane regression: due to GGC fix (Bug 984101 or Bug 992535). Terrence knows this, but I'm not pushing him too much, since there is higher priority work with GGC currently. - 2.5% Sunspider regression: due to removal of sin/cos cache (Bug 967709). Fix being tracked in bug 994993 grtz Hannes On Fri, Apr 11, 2014 at 1:53 PM, Till Schneidereit < [email protected]> wrote: > On Fri, Apr 11, 2014 at 11:37 AM, Nicolas B. Pierron < > [email protected]> wrote: > > > On 04/11/2014 12:52 AM, Chris Peterson wrote: > > > >> Following up on our brainstorming sessions at the JS work week, I'd like > >> to > >> get your feedback on SpiderMonkey's priorities and document them in an > >> ordered list that we can use as our "moral compass". :) > >> > >> Below I also propose a lightweight planning process to keep our focus > >> trained on those project priorities. > >> > >> EXAMPLE LIST OF PROJECT PRIORITIES: > >> > >> 1. Chemspill bugs > >> 2. sec-crit bugs > >> 3. Release tracking bugs: Beta > >> 4. Release tracking bugs: Aurora > >> 5. Release tracking bugs: Nightly > >> 6. topcrash bugs > >> 7. Crashes and assertion failures > >> 8. Misc regressions > >> 9a. Firefox OS performance > >> 9b. Games Initiative > >> 9c. Dev Tools > >> 9d. Firefox features and performance > >> 9e. ES6 features > >> 10a. Shumway > >> 10b. Technical debt? > >> 11a. Infrastructure debt like intermittent orange tests? > >> 11b. sec-high bugs > >> 11c. fuzzblockers? > >> 11d. tsan bugs > >> 11e. ES7 features? > >> > > > > Thanks for doing this priority list, at least this make things clear. :) > > I would suggest to amend it as follow: > > > > 6a. *reproducible* topcrash bugs > > 6b. fuzzblockers > > [...] > > 9f. *non-reproducible* topcrash bugs > > > > - top-crash are not always coming with a clear STR, in such case we > would > > mostly be wasting time at trying to reproduce something. The > reproducible > > one should indeed be important as we are able to investigate them. On the > > other hand, the non-reproducible one should not block our work, while we > > are waiting for somebody to report an STR. > > > > - fuzzblocker are often related to a crash / assertion which is too > > noisy. The fuzzblocker flag was made such as we know which crash should > be > > fixed first, as the assertion is masking all other crash and breaks the > > quality of reports of our Security Team. > > > > > > In Scrum terminology, the P1 bucket is like a Sprint Backlog ("fix now") > >> and > >> the P2 bucket is like a Product Backlog ("fix soon"). > >> > > > > I can understand that we might not want to set a time-frame to "fix > soon", > > but after how many "skip" iterations of P2 should we lower its priority? > > > > https://wiki.mozilla.org/User:Cpeterson > >> > > > > Nice. > > > > I see two issues: > > - Javascript:JIT component seems to be missing. > > > > Also the other sub-components. > > > > - Security bugs are not listed (which make sense on a public wiki) > > > > Nothing we can do about that, I guess. > > > I just went through and triaged the lists in a somewhat ad-hoc manner. Lots > of outdated stuff, so I closed about a dozen and updated a few more. > _______________________________________________ > dev-tech-js-engine-internals mailing list > [email protected] > https://lists.mozilla.org/listinfo/dev-tech-js-engine-internals > _______________________________________________ dev-tech-js-engine-internals mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-js-engine-internals
