Closed Bug 72208 Opened 25 years ago Closed 24 years ago

Starting app with sidebar closed creates panels, should wait until first open

Categories

(SeaMonkey :: Sidebar, defect)

x86
Windows 2000
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED FIXED
mozilla0.9.6

People

(Reporter: mikepinkerton, Assigned: bugzilla)

Details

(Keywords: perf, Whiteboard: [nav+perf])

Attachments

(2 files)

I've noticed (because of an assert in a treewidget frame) that if you have the sidebar enabled but collapsed at startup, it goes ahead and creates the panel anyway even though you can't see it. This has an impact on both startup and new window performance. What we should do is have it not load the panel until you expand the sidebar for the first time, and then it can keep it alive from that point onward to retain state, etc when collapsed.
Keywords: perf
nav triage team: not going to happen for beta, marking nsbeta1-
Keywords: nsbeta1-
Probably not enough to change anyone's mind, but I went and measured this on the test machine for win98 - PII 266 / 128MB, and here's what I have: Average time for 10 consecutive starts (very little scatter), with sidebar ... Collapsed Open (to 'News') Hidden 7.74 7.71 7.50 So, that's about 2/10ths of a second, but only about 3% overall.
jrgm, what if that sidebar panel we're creating is AIM? that would unnecessarily drag in all the aim libraries for no good reason. which panel did you use to do your tests?
Thanks for reminding me: I was going to come back to this again to mention that perhaps 'on startup' was the wrong way to think about this 0.2 second cost. It may be the case that we pay this cost on every new window, which would be much more significant in proportion. But to answer your questions: I was using the bookmarks panel for my test, with the default bookmark list that ships with Netscape. As for AIM, I believe it is the case that we pull in the AIM dll's on startup no matter what (although maybe it is the case that it is the presence of the sidebar panel that triggers this loading). (There's another problem with AIM for which I will file a bug in bugscape, but I'll mention it here: if the AIM sidebar panel is the selected panel, but the sidebar is collapsed and I have not done the initial signup dance, then every time I create a new navigator window, an AIM dialog window appears out of nowhere, asking me to signin, etc.)
--> me
Assignee: matt → blake
Whiteboard: [nav+perf]
Status: NEW → ASSIGNED
Target Milestone: --- → mozilla0.9.3
Target Milestone: mozilla0.9.3 → mozilla0.9.4
Target Milestone: mozilla0.9.4 → mozilla0.9.5
Mass moving lower-priority bugs to 0.9.6 (with Blake's pre-consent) to make room for remaining 0.9.4/eMojo bugs and MachV planning, performance and feature work. If anyone disagrees with the new target, please let me know.
Target Milestone: mozilla0.9.5 → mozilla0.9.6
Attached patch patchSplinter Review
Samir, review?
Samir: when the sidebar is totally hidden (i.e. closed), we don't even create the frames for it or anything, so as far as I know we already don't load the sidebar if it's closed. An alert in search-panel.xul's onload (which I was using to test) seems to confirm that.
Good point. Although we don't need to do the RDF initialization stuff if the sidebar is hidden so the modified patch doesn't hurt. (sidebar_overlay_init() itself gets called even when hidden with your patch, not when collapsed though, but arguably does little.) This is a nit I don't fel strongly about so r=sgehani on either patch.
I didn't see sidebar_overlay_init() getting called when the sidebar was completely hidden so I checked in the patch as-is.
Status: ASSIGNED → RESOLVED
Closed: 24 years ago
Resolution: --- → FIXED
Product: Browser → Seamonkey
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: