Closed
Bug 962026
Opened 12 years ago
Closed 4 years ago
[Meta] OS.File is too slow to start
Categories
(Toolkit Graveyard :: OS.File, defect)
Toolkit Graveyard
OS.File
Tracking
(Not tracked)
RESOLVED
WONTFIX
People
(Reporter: Yoric, Unassigned)
References
Details
(Keywords: meta, perf, Whiteboard: [Async:meeting])
Experience indicates that OS.File is not suitable for critical startup path as loading the OS.File ChromeWorker can take 700 ms on a fast machine – before even the JS code starts being executed (Telemetry pending bug 916076).
We need to fix the issue.
| Reporter | ||
Comment 1•12 years ago
|
||
I believe that the best solution is to provide a native implementation of OS.File for use in the main thread. We can perform the implementation function by function until we have everything that is needed for a fast startup.
Comment 2•9 years ago
|
||
In bug 1280671 I'm trying to move away from using XHR+chrome:// for loading l10n resources.
Unfortunately, my initial experience with OS.File.read is matching this bug - it's around 5 times slower.
I set up an experiment to load 5 l10n resource files on my non-omni.ja'ed browser (so, the files are physically in directories on the hard drive).
Then I loaded them once using following strategies:
1) XHR+chrome:// - 12ms
2) XHR+file:// - 15ms
3) NetUtil+file:// - 16ms
4) OS.File.read + /path/to/the/file - 50ms
So, this bug indicates that loading OS.File is slow. But also using OS.File.read is slow :(
Updated•4 years ago
|
Updated•3 years ago
|
Product: Toolkit → Toolkit Graveyard
You need to log in
before you can comment on or make changes to this bug.
Description
•