In the current stable release of Chrome filters seem to simply be dropping parts of the set when multiple filters are applied and reset. After much digging in my current project I found the same error appearing consistently in the canonical ASA data set example.
To reproduce simply set all the filters and reset them. The result is as as follows – the number of selected flights varies, but never matches the total:
37,407 of 231,083 flights selected.
If you continue to filter and reset the groupAll sum changes, sometimes rising, sometimes falling.
Firefox, Safari and Opera still has completely sane behaviour.
Yikes. This looks tricky to debug.
I'd offer my time, but the internals of Crossfilter seem daunting. Scary, even. Substituting a firm grasp of reality with intuition I'd go with a change in the typed arrays. I'll check the dev and canary builds to see if help is on the way upstream.
Both 24.0.1284.2 dev and 24.0.1285.0 canary display the same behaviour.
I also found this issue when working on DataZooka. If you play around with it, you can see weird things going on. I noticed that sometimes it would throw a TypeError here in the updateMany routine.
It could be this issue that I spotted in beta a while ago. It was only fixed in V8 recently, so I don't know how long it'll take to filter down. I'll see if I can get the fix backported.
Update: it looks like it's definitely the same issue, and it hasn't been merged into the release branches of V8 yet. I'll keep you updated.
Excellent, and well spotted. I'll be watching that thread.
I forked a quick example here with the code split between node.js and index.html:
This code is live for a short time here:
Another browser issue is data-size. Chrome kills a page when memory use goes over 100mb. I manage to avoid the problem by running the filter on a server instead.
Moved the flights example to Heroku: http://flights-ajax.herokuapp.com/
And here is another example close to home: http://icequake.herokuapp.com/
I also ran in to this issue. It seems to be fixed in beta channel now (using 23.0.1271.26)
Yes, the fix was merged into M23 a few days ago, but I think it’s yet to be backported to M21 (stable).
hopefully work around square/#40
M23 is now stable, so this issue should be resolved.
Yes, as far as I can tell this has been resolved.