mozilla :: #activity-stream

11 Jan 2017
01:17as-github-botjkerim: Hey! Someone just assigned you a PR for review: https://github.com/mozilla/activity-stream/pull/1943
16:03as-github-botursula: Hey! Someone just assigned you a PR for review: https://github.com/mozilla/activity-stream/pull/1944
16:24emtwotspurway: nanj k88hudson jkerim ursula hello friends. would anybody like to have ramen for lunch today?
16:24jkerim
16:26k88hudsonyes!
16:26k88hudsonemtwo: ^
16:27emtwo\o/
16:31nanjursula: ^. yeah, i'm looking at you :P
16:31jkerimursula <<
16:31ursulahaha yaa i&#39;m in
16:32jkerimcat | ursula
16:32ursula
16:37emtwowow ursula you just said yes to ramen
16:37jkerimpeer pressure works wonders
16:38emtwousually it&#39;s harder to reject peer pressure when it&#39;s verbal rather than written
16:38emtwobut ursula seems to have it the other way around XD
16:39ursulacan we go at like 12 tho? i have to go to talk to the rrsp person in finch
16:39emtwoi am ok with that.
16:45ursulaok i&#39;m done we can go whenever
18:21ursuladmose: ping
19:18dmoseursula: pong
19:19ursulahey, so i was looking at your pr for enzyme, i noticed that highlightcontext was also stateless, is it possible to also convert that one?
19:19ursulain the same file as where the placeholder highlight context is
19:21ursulahttps://github.com/mozilla/activity-stream/blob/master/content-src/components/HighlightContext/HighlightContext.js#L7-L25 this guy
19:22ursulathat way we could remove the pure render mixin package altogether - since this is the last one that would be using it
19:24* dmose looks
19:25dmoseyes i can do that one
19:25dmosealthough i have some thoughts of reverting the Snippet changes
19:25dmosebecause they allocate a function on every render call which is said to slow
19:25dmosebut i may also have a way around that
19:26ursulado you want me to hold off on the review?
19:26dmosesure, i can get you something within an hour or so, i think
19:26ursulasounds good
19:26dmose. o O ( famous last words )
19:26ursulastateless components + enzyme is a cool combo - was reading up on it
19:27dmoseyeah, it&#39;s good stuff
19:27dmoseshould help our code structure become cleaner over time
19:27dmoseI also found this, which is interesting: https://medium.com/@esamatti/react-js-pure-render-performance-anti-pattern-fb88c101332f#.jlq56es1i
19:31ursulaoh wow
19:35ursulai wonder if these would be some easy wins for perf for us
19:35dmosei was woindering the same thing
19:36dmoseeslint shows that we have a bunch of violations
19:36dmose(i tried turning the rule on locally)
19:36dmoseprobably not hard to fix
19:37dmosewe really need regular perf measurement in our CI
19:37dmosewe have no way of telling when changes regress or improve our perf right now
20:45ursulananj jkerim did we do this? https://github.com/mozilla/activity-stream/issues/1400
20:48nanjursula: not yet, but we do expire the entries in the performance table
20:48ursulaso can i close this? since i don&#39;t think we&#39;re going to be compressing the pings
20:48nanjkeep 30 days of data in that table
20:48nanjyup, go ahead
20:50ursulasweet, thanks!
20:53nanjursula: np
20:54nanjMardak: i am trying your branch now. But i need to turn on the &quot;combined&quot; experiment, right?
20:54Mardaknanj: i overwrote it with `true ||`
20:55nanjk, so it&#39;s already on
20:59jkerimursula: yeah we talked about that idea but eventually threw it out
21:00jkerimthat shouldve just been closed
21:00jkerimin fact if this materialized view thing works, we can even delete stuff out of the performance table altogehter once its flattened into its materialized view
21:01jkerimdmose: your earlier question about sentry filtering
21:02jkerimi found that the python sentry client had a built in thing that scrubs all post data from error reports
21:02jkerimso i used taht
21:02jkerimhttps://github.com/mozilla/embedly-proxy/blob/master/app/proxy/app.py#L31
21:02jkerimhttps://github.com/getsentry/raven-python/blob/c69abaa1506a426802e979c7f8c6a4ab0ef9c4d7/raven/processors.py#L52
21:03jkerimso it doesnt super match the js/addon/clientside use case
21:04nanjMardak: it worked
21:05Mardaknanj: yay! working on tests right now
21:05nanji took those google sites out of my blocklist
21:05nanjthey didn&#39;t show up in my topsites any more
21:06nanjMardak: heh, let me see what&#39;s your magic sauce in your branch
21:06Mardaknanj: just does a group by to select the top page of a www.host and no-www.host
21:07Mardakinstead of adding up all the pages from both www.host and no-www.host
21:08nanjah, i see
21:09nanjyet another experiment for topsites
21:09nanjit works great for me so far
21:11as-github-botncloudioj: Hey! Someone just assigned you a PR for review: https://github.com/mozilla/activity-stream/pull/1946
21:17as-github-botMardak: Hey! Someone just assigned you a PR for review: https://github.com/mozilla/activity-stream/pull/1946
21:18Mardaknanj: thanks for tracking down the issue. i actually thought only the top 2 pages were getting frecency combined already but your example made it clear it wasn&#39;t
21:19nanjMardak: np, happy sql hustle
21:29dmosejkerim: thanks!
21:46grishaemtwo: https://dxr.mozilla.org/mozilla-central/source/mobile/android/docs/activitystreamtelemetry.rst
21:47as-github-botMardak: Hey! Someone just assigned you a PR for review: https://github.com/mozilla/activity-stream/pull/1948
22:11as-github-botursula: Hey! Someone just assigned you a PR for review: https://github.com/mozilla/activity-stream/pull/1948
12 Jan 2017
No messages
   
Last message: 5 days and 19 hours ago