[08:16] ricotz, sorted out the e4x thing [08:16] and now just failing 1 test [08:16] gnome-shell runs, albeit with a few minor issues [08:31] fantastic ;) [09:58] darkxst, do you have a new patch-stack based on gjs master? [09:58] ricotz, just rebased and cleaning it up a [09:58] bit [09:58] darkxst, oh, i see, take your time [10:43] ricotz, https://www.dropbox.com/s/gxhzh6qkc3sak92/gjs188-r1.patch [10:43] might still be a few misplaced hunks but mostly cleaned up. [10:47] darkxst, thanks [10:47] just a minor, there are some of those faulty alignments [10:47] - jsid id, [10:47] + JSObject **obj, [10:53] darkxst, #define JSVAL_IS_OBJECT(obj) !JSVAL_IS_PRIMITIVE(obj) [10:53] this is not the same as the original method [10:53] it will miss the NULL case [10:54] meaning NULL will be handled as not an object [10:55] #define JSVAL_IS_OBJECT(obj) ( JS_IS_NULL(obj) || !JSVAL_IS_PRIMITIVE(obj)) [11:19] hmm right, I missed that somehow ;) [11:22] darkxst, are you still seeing tests failing? [11:22] localeCompare fails [11:22] assertRaises(function() { "\ud800".localeCompare("a"); }); [11:22] is suppose to fail [11:22] so you can ignore those [11:25] but it aborts the other tests? that doesnt seem right [11:26] it runs through for me while commenting those two [11:27] yeh everything else passes, although some things are probably missing tests [11:27] yes, running g-s is the actual test ;) [11:28] g-s runs, but not 100% right [11:28] dash is broken [11:28] overview search is a bit strange [11:30] I suspect I could have some JSPropertyOP's somewhere [11:31] They only segfault if the they use 'id' [11:31] other wise they just fail silently with priv == NULL [11:31] ^^could have missed [11:34] hmm, you should avoid changing the functions interface [11:34] mozilla changed it [11:35] typedef JSBool (*JSPropertyOp)( [11:35] JSContext *cx, JSObject *obj, jsid id, jsval *vp); [11:35] you are expecting **obj, ... [11:36] typedef JSBool [11:36] (* JSPropertyOp)(JSContext *cx, JSHandleObject obj, JSHandleId id, JSMutableHandleValue vp); [11:37] ah i see [11:37] the JSHandle* things are pointers rooted objects [11:37] upstream are planning on changing most of jspai to use them directly [11:37] but currently onlt the callbacks do [11:38] ok [11:43] some way to scan for all callback functions would be nice [11:45] look for gjs_*et_prop_verbose_stub [11:45] gjs_add_prop_verbose_stub gjs_delete_prop_verbose_stub [11:46] jsapi-util.c/h [11:48] hmm [11:50] and how to use that? short of plugging it into the class definitions [11:50] was just looking to similar interfaces [11:51] i guess i will try to run it later to maybe see some failure outputs [11:58] maybe could check from within priv_from_js() [13:52] darkxst, didnt really get to look into it [13:54] darkxst, it seems to run stable despite the broken dash