Confirmed users
125
edits
Platform/2015-04-14: Difference between revisions
→MemShrink (njn): add blog post link
(Created page with "<!-- Maybe don't screw with these links unless you've read this blog post: http://blog.johnath.com/2011/01/20/automatic-date-links-in-mediawiki/ Just copy them to new pages an...") |
(→MemShrink (njn): add blog post link) |
||
| Line 40: | Line 40: | ||
Brian Hackett took advantage of the removal of object parents to reduce the size of base shapes ({{bug|1143256}}). The JS engine creates many of these to track the characteristics of JS objects. | Brian Hackett took advantage of the removal of object parents to reduce the size of base shapes ({{bug|1143256}}). The JS engine creates many of these to track the characteristics of JS objects. | ||
Nick Fitzgerald greatly reduced the amount of memory used for saving JS exception stacks ({{bug|1038238}}). This reduced the memory of a test case compiled from Dart to JS from over 1GB to around 170MB ({{bug|1125259}}). | Nick Fitzgerald greatly reduced the amount of memory used for saving JS exception stacks ({{bug|1038238}}). This reduced the memory of a test case compiled from Dart to JS from over 1GB to around 170MB ({{bug|1125259}}). Blog post: http://fitzgeraldnick.com/weblog/61/ | ||
Seth Fowler fixed a bad leak regression ({{bug|1150127}}) where closing a page while it was in the process of loading could leak the page and cause long browser pauses. Seth also fixed a pair of issues that caused very large memory usage (more than 1GB) when scrolling in the downloads list ({{bug|1148682}} and {{bug|1148684}}). | Seth Fowler fixed a bad leak regression ({{bug|1150127}}) where closing a page while it was in the process of loading could leak the page and cause long browser pauses. Seth also fixed a pair of issues that caused very large memory usage (more than 1GB) when scrolling in the downloads list ({{bug|1148682}} and {{bug|1148684}}). | ||
Bill McCloskey fixed an issue where a docshell was being held alive for too long on e10s ({{bug|1137933}}). This was one of the remaining blockers to enabling e10s Mochitest browser-chrome tests, which will give us more test coverage for leaks. | Bill McCloskey fixed an issue where a docshell was being held alive for too long on e10s ({{bug|1137933}}). This was one of the remaining blockers to enabling e10s Mochitest browser-chrome tests, which will give us more test coverage for leaks. | ||
===Media=== | ===Media=== | ||