Bug 1536091 Comment 2 Edit History

Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.

Well for some reason these seven items load from the "disk" cache rather than from the "memory" cache:

xulcache/resource/app/chrome/browser/content/browser/browser.xul
jssubloader/global/resource/app/features/formautofill@mozilla.org/chrome/content/customElements.js
jssubloader/global/resource/app/chrome/browser/content/browser/browser-sync.js
jssubloader/global/resource/gre/chrome/toolkit/content/global/viewZoomOverlay.js
jssubloader/global/resource/app/chrome/browser/content/browser/browser-allTabsMenu.js
jssubloader/global/resource/app/chrome/browser/content/browser/browser-gestureSupport.js
jssubloader/global/resource/app/chrome/browser/content/browser/browser-thumbnails.js

"disk" and "memory" caches reflect what I had the histogram record, but they're not the most descriptive names. We're not actually reading from disk for any of these items, but we are calling inflate(...) for each of them, which seems less than ideal.

All of this seems to be due to how the "memory" cache (just a hash table (`mTable`)) is only populated when we call PutBuffer, not from the result of GetBuffer. Populating the cache in GetBuffer would incur an extra copy since the caller is supposed to take ownership of the resulting buffer.

In any case, I don't have a depth of familiarity with how all of this *ought* to work or why we're taking a different path with the quantumbar. kmag, does any of this ring any bells in your head?
Well for some reason, with quantumbar enabled these seven items load from the "disk" cache rather than from the "memory" cache:

xulcache/resource/app/chrome/browser/content/browser/browser.xul
jssubloader/global/resource/app/features/formautofill@mozilla.org/chrome/content/customElements.js
jssubloader/global/resource/app/chrome/browser/content/browser/browser-sync.js
jssubloader/global/resource/gre/chrome/toolkit/content/global/viewZoomOverlay.js
jssubloader/global/resource/app/chrome/browser/content/browser/browser-allTabsMenu.js
jssubloader/global/resource/app/chrome/browser/content/browser/browser-gestureSupport.js
jssubloader/global/resource/app/chrome/browser/content/browser/browser-thumbnails.js

"disk" and "memory" caches reflect what I had the histogram record, but they're not the most descriptive names. We're not actually reading from disk for any of these items, but we are calling inflate(...) for each of them, which seems less than ideal.

All of this seems to be due to how the "memory" cache (just a hash table (`mTable`)) is only populated when we call PutBuffer, not from the result of GetBuffer. Populating the cache in GetBuffer would incur an extra copy since the caller is supposed to take ownership of the resulting buffer.

In any case, I don't have a depth of familiarity with how all of this *ought* to work or why we're taking a different path with the quantumbar. kmag, does any of this ring any bells in your head?

Back to Bug 1536091 Comment 2