mozilla :: #jsapi

12 Jul 2017
01:19sfinkwtf? static_assert on our product gcc (4.9.4) *must* have an explanation string. That was annoying to track down.
01:35shu`sickptomato-M: i got sick and asking reviewing for and backporting your patch completely slipped my mind, sorry!
01:35shu`sickptomato-M: it has a pending r? now
01:59ptomato-Mshu`sick: sorry to hear you're sick! no worries about the patch. I'm still a bit fuzzy on who to request review from but I should probably have set the flag myself
14:59jorendorffbholley: Is there a place in the code (or wiki, anyhwhere) that has a short summary of content XBL? If not, where should I put one?
15:00jorendorffit seems like a coherent enough chunk of stuff to have a single comment
15:04jorendorff(the existing comments aren't so bad)
15:09jorendorffbholley: Separately... arity: does every content XBL scopes have exactly one content scope that it works for?
15:10jorendorffbholley: i.e. in the new order, say 1 compartment contains N windows containing marquee elements, maybe we should have 1 content XBL compartment containing N content XBL scopes.
15:29jorendorffLooks like the section of xpcpublic.h where all the XBLScope functions live would be a good place to document.
15:29jorendorff*to put a few paragraphs of documentation
15:30tcampbellDid we change the behavior of MapObject recently? perf.html seems way slower this week and 85% of runtime is in MapObject::get/set
15:47joncosfink: ping
15:53sfinkjonco: pong. You can't hear me?
16:07sfinkhttp://searchfox.org/mozilla-central/source/js/src/jsgc.cpp#6297
16:08sfinkbug 1375566
16:08firebothttps://bugzil.la/1375566 NEW, sphink@gmail.com We can blow our slice budgets due to arena finalization taking too long
16:40jimbmismith: Hi!
16:49fitzgensfink: ping
17:06mrgigglesthe mozilla-inbound tree is now closed (bustage from bug 1352559)
17:06firebothttps://bugzil.la/1352559 NEW, dan1bh@yahoo.co.uk Remove NPN_NewStream and related machinery
17:12mrgigglesthe mozilla-inbound tree is now open
17:57bholleyjorendorff: I don't know if there is one. Probably somewhere in the xpc headers around where we manage the scopes
17:58bholleyjorendorff: and yes, every content xbl scope has exactly one content scope it works for
17:58bholleyjorendorff: and ever content scope has zero or one content xbl scopes
17:59jorendorffbholley: thanks. another q - I am making mIsContentXBLScope a field of CompartmentPrivate; I can do the same with mIsAddonScope, right?
17:59bholleyjorendorff: yes, those should live together
17:59jorendorffgood
18:21zhitingzIs there a good way to visualize what jit does? I'm trying to figure out why LICM doesn't perform on the one of my selfhosted code.
18:22zhitingzIONFLAGS output only says something like: constant will be hosted if its users also. But I don't see anything afterwards. How should I know what users are?
19:00* bz has had that problem....
19:00bzIt's pretty anoying
19:00bzer, annoying
19:19evilpiezhitingz: https://github.com/sstangl/iongraph can be useful
20:29jimbSo, I reviewed a bunch of Eddy Bruel's patches in bug 1271654, but then he left Mozilla. Is it okay for me to revise his patches as I requested and just land them myself?
20:29jimbTechnically, they've been reviewed.
20:29firebothttps://bugzil.la/1271654 NEW, jimb@mozilla.com Implement a C++ interface for Debugger.Script instances.
20:29jimbBut revising and landing them myself doesn't feel right.
20:30jimbDo I need to get a second peer's review?
20:42jandemif it's a matter of "rename this variable and fix this comment" I'd land it probably
20:42jandemthat's also what often happens with contributor patches
21:27mrgigglesthe mozilla-inbound tree is now closed (crashes on Windows 8 x64 opt)
23:31sfinkfitzgen: much delayed pong
23:31fitzgensfink: uhhhhh... no idea why I pung anymore
23:31fitzgensfink: you get off easy ;)
23:31sfink\o/
23:34zhitingzIs there any reason a call to UnsafeGetResearvedSlot would block LICM?
23:40zhitingzIt's a readonly function.
23:42bzDid it get inlined?
23:44bzBecause if not... we don't flag it as movable, I bet
23:48zhitingzhow can I tell whether it's inline or not?
23:50zhitingzbz: ^
23:50zhitingzDo I need to annotate on the js side?
23:51zhitingzAll I can see is call constant254 constant238 phi237 constant323. constant254 is the function
23:51zhitingzconstant function UnsafeGetReservedSlot at 10a511e70
13 Jul 2017
No messages
   
Last message: 73 days and 5 hours ago