We try to only save on pagehide, but this event doesn't seem to be sent on fast quit (like Cmd-Q of Safari or Ctrl-C when run from command line). Just save on every modification, since it's probably cheap in most realistic scenarios. If it turns out to be expensive for some reason, we could coalesce updates using a timer.
<rdar://problem/23653243>
Created attachment 266122 [details] Proposed Fix
*** Bug 129715 has been marked as a duplicate of this bug. ***
Comment on attachment 266122 [details] Proposed Fix View in context: https://bugs.webkit.org/attachment.cgi?id=266122&action=review > Source/WebInspectorUI/UserInterface/Controllers/DebuggerManager.js:877 > + let savedBreakpoints = this._breakpoints > + .filter((breakpoint) => !!breakpoint.url) > + .map((breakpoint) => breakpoint.info); Not a fan of hanging periods.
Comment on attachment 266122 [details] Proposed Fix View in context: https://bugs.webkit.org/attachment.cgi?id=266122&action=review >> Source/WebInspectorUI/UserInterface/Controllers/DebuggerManager.js:877 >> + .map((breakpoint) => breakpoint.info); > > Not a fan of hanging periods. Oh, I know.. I always write it this way then change it back before submitting the patch. It's just so much easier for me to think "oh, a bunch of leading periods, this must be a chain, I'll look up and to the right" than to see the line mysteriously indented (is it a completion handler? a chain? I don't know). But if nobody else likes it (as it seems to be the case) I'll switch it back.
Committed r192771: <http://trac.webkit.org/changeset/192771>