https://bugzilla.wikimedia.org/show_bug.cgi?id=50518
--- Comment #4 from Alex Monk <kren...@wikimedia.org> --- We discussed how to do this in #mediawiki-visualeditor a bit. Some quotes: <Krinkle> when a certain amount of time has passed (potentially taking the document complexity in account as well, avoid measuring network by accident), set a boolean flag, which in getEditnotices() will make it include/exclude that notice accordingly <Krenair> I'm wondering what operation we should measure and what timing shows a reasonable speed <Krinkle> Krenair: one or more of the actions in the setup of the surface (parsing of html into data DOM, conversion to linenar data model, building of CE etc.) <Krenair> but those operations will all take different times depending on document length/complexity :/ <Krinkle> Krenair: Yes, but avoiding I mean, taking into account <Krinkle> e.g. not `timespend > maxtime`, but something like `timespent/doc-size > maxratio` <Krinkle> RoanKattouw might have some math ideas that are more useful <Krinkle> However, we should probably do some research and see how slow the 'too slow' browsers really are. <Krinkle> We might be able to measure it with a very simple operation at a small scale (e.g. create an iframe and attach it, just a random unused iframe), that that one operation is measurable enough, we can do it that way. <Krenair> Hm. This is getting increasingly out of the 'MediaWiki integration' part <Krinkle> The function calls are all in viewpagetarget in setupSurface <Krinkle> though ideally this would probably be in ve-core, seems a useful statistis Roan, could you comment on this when you have time please? -- You are receiving this mail because: You are on the CC list for the bug. _______________________________________________ Wikibugs-l mailing list Wikibugs-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikibugs-l
