33053 2009-12-30 06:58:15 -0800 JSON.stringify and JSON.parse implementations needlessly process properties in the prototype chain 2010-04-24 20:29:20 -0700 1 1 1 Unclassified WebKit JavaScriptCore 528+ (Nightly build) PC Linux RESOLVED FIXED ES5 P2 Normal --- 1 kent.hansen kent.hansen commit-queue eric erights oliver pmuellr webkit.review.bot oldest_to_newest 175910 0 kent.hansen 2009-12-30 06:58:15 -0800 ES5 section 15.12.3, abstract Operation JO(value), step 6: "Let K be an internal List of Strings consisting of the names of all the own properties of value [...]". The JavaScriptCore implementation is calling JSObject::getPropertyNames(), whereas it could be calling JSObject::getOwnPropertyNames() instead. Making this change does not affect the final output (the prototype-inherited properties are ignored anyways, since getOwnPropertySlot() subsequently returns false for them), but could make the implementation faster and use less memory. 175914 1 45662 kent.hansen 2009-12-30 07:05:02 -0800 Created attachment 45662 Proposed patch 175915 2 webkit.review.bot 2009-12-30 07:06:34 -0800 style-queue ran check-webkit-style on attachment 45662 without any errors. 175918 3 kent.hansen 2009-12-30 07:24:29 -0800 Hmm, there should probably be a check to see if the OverridesGetPropertyNames flag is on, and in that case call getPropertyNames() as before. It scares me a bit that there are now two virtual functions for getting property names. Can't we just make getPropertyNames() non-virtual now that there is getOwnPropertyNames()? Then rename the OverridesGetPropertyNames flag to OverridesGetOwnPropertyNames and update the bindings accordingly. I can do the refactoring if that sounds sensible. 175929 4 pmuellr 2009-12-30 08:47:47 -0800 If you're doing some refactoring in this area, there may be some issues relating to bug 32242 - which needs to obtain the non-enumerable properties of an object. I'm very unfamiliar with the JSC code, but attempted to look into this support anyway, and found I was going to have to hit a lot of code to go the first route I attempted (see the bug for details). May well be it's a n00b issue for me, but if you're in the middle of property-retrieval refactorings, something to think about it. 176125 5 45662 oliver 2009-12-30 22:56:50 -0800 Comment on attachment 45662 Proposed patch Can you determine what correct spec behaviour is given: javascript:a={__proto__:{foo:"bar"}, get b() { this.foo="PASS"; }}; alert(JSON.stringify(a)) 176603 6 kent.hansen 2010-01-04 01:43:48 -0800 (In reply to comment #5) > (From update of attachment 45662 [details]) > Can you determine what correct spec behaviour is given: > javascript:a={__proto__:{foo:"bar"}, get b() { this.foo="PASS"; }}; > alert(JSON.stringify(a)) Thanks, that's a good test case to include. Fairly sure it should be "{}", as per 15.12.3.JO.6. The prototype should not be consulted. SpiderMonkey produces "{}" with both the built-in JSON.stringify and http://www.json.org/json2.js. JSC produces "{"foo":"PASS"}" without the patch, "{}" with the patch. JSC produces "{"foo":"PASS"}" even with json2.js. That code uses for..in in combination with hasOwnProperty() to only process "own" properties of the object. This means it _will_ see the property name "b" before hasOwnProperty() rejects it, but that shouldn't cause the _getter_ itself to be invoked. Uh-oh, maybe it's because the JSObject::hasOwnProperty() implementation calls getOwnPropertySlot()? It should probably be calling getOwnPropertyDescriptor(). I'll investigate and spin off a separate bug report. 176615 7 kent.hansen 2010-01-04 02:10:50 -0800 (In reply to comment #6) Sorry, I misread the test input. b of course _is_ an "own" property. But yeah, what's happening is, "foo" will be the last item in the PropertyNameArray after the call to getPropertyNames(), even though the property is not an "own" property; after the call to the "b" getter, however, "foo" will be an own property, so it won't be rejected by the next call to hasOwnProperty(), hence the side-effect-generated "foo" property ends up in the output. Again, coming back to the spec, the propertylist is initialized before the property values are stringified, so I believe properties added during the stringification should not be part of the output. 176619 8 kent.hansen 2010-01-04 02:33:56 -0800 (In reply to comment #6) > JSC produces "{"foo":"PASS"}" even with json2.js. That code uses for..in in > combination with hasOwnProperty() to only process "own" properties of the > object. Looks like a bug in the for..in implementation. SpiderMonkey: js> a={__proto__:{foo:"bar"}, get b() { this.foo="PASS"; }}; ({get b () {this.foo = "PASS";}}) js> for (var p in a) print(p, a[p], a.hasOwnProperty(p)) b undefined true JSC: > a={__proto__:{foo:"bar"}, get b() { this.foo="PASS"; }} [object Object] > for (var p in a) print(p, a[p], a.hasOwnProperty(p)) b undefined true foo PASS true ES5 section 12.6.4 states that "if new properties are added to the object being enumerated during enumeration, the newly added properties are guaranteed not to be visited in the active enumeration." So, the "foo" added by the b getter should not be included in the enumeration. The reason the "foo" in the prototype should not be included is because of the "no-shadowing" rule, also given in section 12.6.4. I.e. it looks like even though newly added properties are not included in the enumeration, they do affect shadowing (at least if you take SpiderMonkey's implementation to be "correct"). 176654 9 kent.hansen 2010-01-04 03:29:19 -0800 (In reply to comment #8) > Looks like a bug in the for..in implementation. > > ES5 section 12.6.4 states that "if new properties are added to the object being > enumerated during enumeration, the newly added properties are guaranteed not to > be visited in the active enumeration." Ugh, I was looking at a previous draft of the spec. In the final draft, the order of "guaranteed" and "not" has been swapped again, back to how it was in ES3 (makes a big difference, huh!): http://www.mail-archive.com/es-discuss@mozilla.org/msg03118.html So arguably json2.js is buggy because it relies on non-standard behavior. Ideally building the property list and stringifying the values should be decoupled, like it is in the ES5 algorithm. It'd be nice to have a test that sanctions this for-in behavior in JSC, though. 176757 10 kent.hansen 2010-01-04 10:58:20 -0800 (In reply to comment #3) > Hmm, there should probably be a check to see if the OverridesGetPropertyNames > flag is on, and in that case call getPropertyNames() as before. > > It scares me a bit that there are now two virtual functions for getting > property names. Can't we just make getPropertyNames() non-virtual now that > there is getOwnPropertyNames()? Then rename the OverridesGetPropertyNames flag > to OverridesGetOwnPropertyNames and update the bindings accordingly. I can do > the refactoring if that sounds sensible. I created https://bugs.webkit.org/show_bug.cgi?id=33159 to get some feedback on this issue. 176762 11 kent.hansen 2010-01-04 11:09:27 -0800 (In reply to comment #4) > If you're doing some refactoring in this area, there may be some issues > relating to bug 32242 - which needs to obtain the non-enumerable properties of > an object. I'm very unfamiliar with the JSC code, but attempted to look into > this support anyway, and found I was going to have to hit a lot of code to go > the first route I attempted (see the bug for details). May well be it's a n00b > issue for me, but if you're in the middle of property-retrieval refactorings, > something to think about it. Yeah, see Maciej's latest comment. I'm pursuing the "add extra flag argument to get(Own)PropertyNames" approach. It should be orthogonal to this change, though. 178629 12 46130 kent.hansen 2010-01-08 03:58:33 -0800 Created attachment 46130 Revised patch (adds tests) I haven't added the checks for overridesGetPropertyNames() as I mentioned in comment #3, because that would still mean the behavior would be broken for classes that do override it (and there are quite a few). If a class reimplements getPropertyNames(), it should reimplement getOwnPropertyNames() too (i.e. the OverridesGetPropertyNames should imply that both of the property name functions are reimplemented). I can't think of a use case where it's meaningful to only reimplement getPropertyNames(), at least. The difference between the result created by json2.js (subject to the for-in issue) and the native implementation has been marked as FAIL. I'm not sure how to best proceed: 1) Acknowledge that the for-in behavior is a bug in JSC, and make it a blocker for this bug; or 2) Try to convince Crockford that json2.js should be updated to work around this behavior (i.e. detect for-in behavior for injected properties and fall back to a slower codepath); or 3) Accept that the native JSON and json2.js results will differ in this case. 178632 13 eric 2010-01-08 04:09:42 -0800 Due to the two recent rollouts, I must as: Did you run run-javascriptcore-tests with this patch? I've filed bug 33376 about the commit-queue running run-javascriptcore-tests in the future. (note, you're not the only one who has found this split confusing. bug 5253 is one bug on the subject.) 178647 14 eric 2010-01-08 05:01:29 -0800 the commit-queue is now running run-javascriptcore-tests before committing, so we won't have rollouts again like we had this evening. :) 178650 15 kent.hansen 2010-01-08 05:34:53 -0800 (In reply to comment #14) > the commit-queue is now running run-javascriptcore-tests before committing, so > we won't have rollouts again like we had this evening. :) Much appreciated! Regarding the patch for this bug, I've run the javascriptcore tests after applying it, and there is no change. JSON is a feature introduced in ES5 anyway, and JavaScriptCore/tests/mozilla doesn't yet contain the ecma_5/ directory that's present in mozilla trunk. Like we discussed online, I want to investigate bringing all tests under JavaScriptCore/tests/mozilla up-to-date. 179229 16 oliver 2010-01-10 19:45:16 -0800 Sorry for delayed response, i've been on holiday, and am now back again (In reply to comment #8) > (In reply to comment #6) > > JSC produces "{"foo":"PASS"}" even with json2.js. That code uses for..in in > > combination with hasOwnProperty() to only process "own" properties of the > > object. > > Looks like a bug in the for..in implementation. > > SpiderMonkey: > js> a={__proto__:{foo:"bar"}, get b() { this.foo="PASS"; }}; > ({get b () {this.foo = "PASS";}}) > js> for (var p in a) print(p, a[p], a.hasOwnProperty(p)) > b undefined true > > JSC: > > a={__proto__:{foo:"bar"}, get b() { this.foo="PASS"; }} > [object Object] > > for (var p in a) print(p, a[p], a.hasOwnProperty(p)) > b undefined true > foo PASS true > > ES5 section 12.6.4 states that "if new properties are added to the object being > enumerated during enumeration, the newly added properties are guaranteed not to > be visited in the active enumeration." So, the "foo" added by the b getter > should not be included in the enumeration. actually the foo added by the b getter should be included. Evaluation of for-in is for (i in o) if (o.hasOwnProperty(i)) log(o[i]) names = getPropertyNames(o) -> ["b" /* from o */, "foo" /* from o.__proto__ */] for (_i = 0; _i < names.length; _i++) { var i = names[_i]; if (o.hasOwnProperty(i)) log(o[i]) } The first iteration of this will be i == "b", o[i] then results in the o.b getter being called which then creates o.foo (no prototype involved in assignment), in the next loop foo will now be directly on the object so will be visited. I am unsure why it is not working in firefox 179735 17 kent.hansen 2010-01-12 01:46:31 -0800 (In reply to comment #16) > actually the foo added by the b getter should be included. Evaluation of > for-in is > > for (i in o) if (o.hasOwnProperty(i)) log(o[i]) > > names = getPropertyNames(o) -> ["b" /* from o */, "foo" /* from o.__proto__ */] > for (_i = 0; _i < names.length; _i++) { > var i = names[_i]; > if (o.hasOwnProperty(i)) > log(o[i]) > } > > The first iteration of this will be i == "b", o[i] then results in the o.b > getter being called which then creates o.foo (no prototype involved in > assignment), in the next loop foo will now be directly on the object so will be > visited. I am unsure why it is not working in firefox It's because of the no-shadowing rule that I mentioned. From ES5 section 12.6.4: "Enumerating the properties of an object includes enumerating properties of its prototype, and the prototype of the prototype, and so on, recursively; but a property of a prototype is not enumerated if it is “shadowed” because some previous object in the prototype chain has a property with the same name." Pseudo-code execution: 1. Get own properties of o -> ["b"] 2. Enumerate "b" (one pass through the for-in body); as a side-effect, this will create "foo" on o 3. Recurse: Get own properties of o.__proto__ -> ["foo"]. 4. Check if some previous object in the prototype chain has the property "foo" -> yes, o has a property called "foo", so "foo" is not visited. (The fact that "foo" has not actually been enumerated before does not matter.) 179876 18 oliver 2010-01-12 10:14:34 -0800 (In reply to comment #17) > (In reply to comment #16) > > actually the foo added by the b getter should be included. Evaluation of > > for-in is > > > > for (i in o) if (o.hasOwnProperty(i)) log(o[i]) > > > > names = getPropertyNames(o) -> ["b" /* from o */, "foo" /* from o.__proto__ */] > > for (_i = 0; _i < names.length; _i++) { > > var i = names[_i]; > > if (o.hasOwnProperty(i)) > > log(o[i]) > > } > > > > The first iteration of this will be i == "b", o[i] then results in the o.b > > getter being called which then creates o.foo (no prototype involved in > > assignment), in the next loop foo will now be directly on the object so will be > > visited. I am unsure why it is not working in firefox > > It's because of the no-shadowing rule that I mentioned. > From ES5 section 12.6.4: "Enumerating the properties of an object includes > enumerating properties of its prototype, and the prototype of > the prototype, and so on, recursively; but a property of a prototype is not > enumerated if it is “shadowed” > because some previous object in the prototype chain has a property with the > same name." > > Pseudo-code execution: > 1. Get own properties of o -> ["b"] > 2. Enumerate "b" (one pass through the for-in body); as a side-effect, this > will create "foo" on o > 3. Recurse: Get own properties of o.__proto__ -> ["foo"]. > 4. Check if some previous object in the prototype chain has the property "foo" > -> yes, o has a property called "foo", so "foo" is not visited. (The fact that > "foo" has not actually been enumerated before does not matter.) No, you're misunderstanding the purpose of that rule -- it prevents you from visiting the same property name multiple times, the algorithm is 1. names = getPropertyNames(o); 2. for each name in names i. if !o.hasProperty(name) goto step 2 ii. do loop body getPropertyNames(o) 1. names = []; 2. while (o) i. temp = getOwnPropertyNames(o) ii. for each name in temp a. if names.indexOf(name) != -1 goto 2.ii b. names.push(name) 3. return names Eg. the list of property names for an object is a uniqued ordered list 180123 19 kent.hansen 2010-01-13 01:25:32 -0800 (In reply to comment #18) > No, you're misunderstanding the purpose of that rule -- it prevents you from > visiting the same property name multiple times, the algorithm is > 1. names = getPropertyNames(o); > 2. for each name in names > i. if !o.hasProperty(name) goto step 2 > ii. do loop body > > getPropertyNames(o) > 1. names = []; > 2. while (o) > i. temp = getOwnPropertyNames(o) > ii. for each name in temp > a. if names.indexOf(name) != -1 goto 2.ii > b. names.push(name) > 3. return names > > Eg. the list of property names for an object is a uniqued ordered list Where does that algorithm come from? I'm just following the ECMA-262 wording. Some more interesting cases: Case 1: try swapping out the __proto__ property while o is being enumerated. SpiderMonkey: js> o = { a : 1, __proto__: { b : 2 } }; ({a:1}) js> for (var p in o) { if (p == "a") o.__proto__ = { c : 3 }; print(p); } a c JSC: > o = { a : 1, __proto__: { b : 2 } }; [object Object] > for (var p in o) { if (p == "a") o.__proto__ = { c : 3 }; print(p); } a Case 2: add a property to o.__proto__ while o is being enumerated. SpiderMonkey: js> o = { a : 1, __proto__: {} }; ({a:1}) js> for (var p in o) { if (p == "a") { o.__proto__.b = 2; }; print(p); } a b JSC: > o = { a : 1, __proto__: { } }; [object Object] > for (var p in o) { if (p == "a") { o.__proto__.b = 2; }; print(p); } a Case 3: delete a property on o and add a property with the same name on o.__proto__ before o.__proto__ is being enumerated. SpiderMonkey: js> o = { a : 1, __proto__: { b : 2 } }; ({a:1}) js> for (var p in o) { if ((p == "a") && o.hasOwnProperty("a")) { delete o.a; o.__proto__.a = 3; }; print(p); } a b a JSC: > o = { a : 1, __proto__: { b : 2 } }; [object Object] > for (var p in o) { if ((p == "a") && o.hasOwnProperty("a")) { delete o.a; o.__proto__.a = 3; }; print(p); } a b Again, I find the SpiderMonkey behavior fits with the ECMA-262 wording in all three cases. So maybe I'm misunderstanding the _purpose_ of the rule, but the question is whether it should be implemented as it's currently worded. I'll probably fire off a mail to es-discuss@ about this, once I've been able to check the behavior on IE8. (Attempt on Opera 10.10 was futile because it doesn't seem to support the __proto__ extension.) 180544 20 46130 oliver 2010-01-13 22:18:33 -0800 Comment on attachment 46130 Revised patch (adds tests) Okay, i think the for..in enumeration issue can be considered separate from this. So r=me 180547 21 46130 commit-queue 2010-01-13 22:37:20 -0800 Comment on attachment 46130 Revised patch (adds tests) Clearing flags on attachment: 46130 Committed r53239: <http://trac.webkit.org/changeset/53239> 180548 22 commit-queue 2010-01-13 22:37:27 -0800 All reviewed patches have been landed. Closing bug. 45662 2009-12-30 07:05:02 -0800 2010-01-08 03:58:33 -0800 Proposed patch JSONstringify.diff text/plain 1855 kent.hansen ZGlmZiAtLWdpdCBhL0phdmFTY3JpcHRDb3JlL0NoYW5nZUxvZyBiL0phdmFTY3JpcHRDb3JlL0No YW5nZUxvZwppbmRleCAwYjVmZTc4Li5kOTFkZjgyIDEwMDY0NAotLS0gYS9KYXZhU2NyaXB0Q29y ZS9DaGFuZ2VMb2cKKysrIGIvSmF2YVNjcmlwdENvcmUvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTQg QEAKKzIwMDktMTItMzAgIEtlbnQgSGFuc2VuICA8a2VudC5oYW5zZW5Abm9raWEuY29tPgorCisg ICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgIEpTT04uc3RyaW5n aWZ5IGFuZCBKU09OLnBhcnNlIG5lZWRsZXNzbHkgcHJvY2VzcyBwcm9wZXJ0aWVzIGluIHRoZSBw cm90b3R5cGUgY2hhaW4KKyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcu Y2dpP2lkPTMzMDUzCisKKyAgICAgICAgKiBydW50aW1lL0pTT05PYmplY3QuY3BwOgorICAgICAg ICAoSlNDOjpTdHJpbmdpZmllcjo6SG9sZGVyOjphcHBlbmROZXh0UHJvcGVydHkpOgorICAgICAg ICAoSlNDOjpXYWxrZXI6OndhbGspOgorCiAyMDA5LTEyLTI4ICBTaHUgQ2hhbmcgIDxDaGFuZy5T aHVAbm9raWEuY29tPgogCiAgICAgICAgIFJldmlld2VkIGJ5IExhc3psbyBHb21ib3MuCmRpZmYg LS1naXQgYS9KYXZhU2NyaXB0Q29yZS9ydW50aW1lL0pTT05PYmplY3QuY3BwIGIvSmF2YVNjcmlw dENvcmUvcnVudGltZS9KU09OT2JqZWN0LmNwcAppbmRleCBmMjhlNzBlLi5jZTBkY2ZmIDEwMDY0 NAotLS0gYS9KYXZhU2NyaXB0Q29yZS9ydW50aW1lL0pTT05PYmplY3QuY3BwCisrKyBiL0phdmFT Y3JpcHRDb3JlL3J1bnRpbWUvSlNPTk9iamVjdC5jcHAKQEAgLTUwMSw3ICs1MDEsNyBAQCBib29s IFN0cmluZ2lmaWVyOjpIb2xkZXI6OmFwcGVuZE5leHRQcm9wZXJ0eShTdHJpbmdpZmllciYgc3Ry aW5naWZpZXIsIFN0cmluZ0J1aQogICAgICAgICAgICAgICAgIG1fcHJvcGVydHlOYW1lcyA9IHN0 cmluZ2lmaWVyLm1fYXJyYXlSZXBsYWNlclByb3BlcnR5TmFtZXMuZGF0YSgpOwogICAgICAgICAg ICAgZWxzZSB7CiAgICAgICAgICAgICAgICAgUHJvcGVydHlOYW1lQXJyYXkgb2JqZWN0UHJvcGVy dHlOYW1lcyhleGVjKTsKLSAgICAgICAgICAgICAgICBtX29iamVjdC0+Z2V0UHJvcGVydHlOYW1l cyhleGVjLCBvYmplY3RQcm9wZXJ0eU5hbWVzKTsKKyAgICAgICAgICAgICAgICBtX29iamVjdC0+ Z2V0T3duUHJvcGVydHlOYW1lcyhleGVjLCBvYmplY3RQcm9wZXJ0eU5hbWVzKTsKICAgICAgICAg ICAgICAgICBtX3Byb3BlcnR5TmFtZXMgPSBvYmplY3RQcm9wZXJ0eU5hbWVzLnJlbGVhc2VEYXRh KCk7CiAgICAgICAgICAgICB9CiAgICAgICAgICAgICBtX3NpemUgPSBtX3Byb3BlcnR5TmFtZXMt PnByb3BlcnR5TmFtZVZlY3RvcigpLnNpemUoKTsKQEAgLTc0Niw3ICs3NDYsNyBAQCBORVZFUl9J TkxJTkUgSlNWYWx1ZSBXYWxrZXI6OndhbGsoSlNWYWx1ZSB1bmZpbHRlcmVkKQogICAgICAgICAg ICAgICAgIG9iamVjdFN0YWNrLmFwcGVuZChvYmplY3QpOwogICAgICAgICAgICAgICAgIGluZGV4 U3RhY2suYXBwZW5kKDApOwogICAgICAgICAgICAgICAgIHByb3BlcnR5U3RhY2suYXBwZW5kKFBy b3BlcnR5TmFtZUFycmF5KG1fZXhlYykpOwotICAgICAgICAgICAgICAgIG9iamVjdC0+Z2V0UHJv cGVydHlOYW1lcyhtX2V4ZWMsIHByb3BlcnR5U3RhY2subGFzdCgpKTsKKyAgICAgICAgICAgICAg ICBvYmplY3QtPmdldE93blByb3BlcnR5TmFtZXMobV9leGVjLCBwcm9wZXJ0eVN0YWNrLmxhc3Qo KSk7CiAgICAgICAgICAgICAgICAgLy8gZmFsbHRocm91Z2gKICAgICAgICAgICAgIH0KICAgICAg ICAgICAgIG9iamVjdFN0YXJ0VmlzaXRNZW1iZXI6Cg== 46130 2010-01-08 03:58:33 -0800 2010-01-13 22:37:20 -0800 Revised patch (adds tests) JSONstringify2.diff text/plain 15597 kent.hansen ZGlmZiAtLWdpdCBhL0phdmFTY3JpcHRDb3JlL0NoYW5nZUxvZyBiL0phdmFTY3JpcHRDb3JlL0No YW5nZUxvZwppbmRleCAwMTdjMGRhLi5lYjUwZDdlIDEwMDY0NAotLS0gYS9KYXZhU2NyaXB0Q29y ZS9DaGFuZ2VMb2cKKysrIGIvSmF2YVNjcmlwdENvcmUvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTQg QEAKKzIwMTAtMDEtMDcgIEtlbnQgSGFuc2VuICA8a2VudC5oYW5zZW5Abm9raWEuY29tPgorCisg ICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgIEpTT04uc3RyaW5n aWZ5IGFuZCBKU09OLnBhcnNlIG5lZWRsZXNzbHkgcHJvY2VzcyBwcm9wZXJ0aWVzIGluIHRoZSBw cm90b3R5cGUgY2hhaW4KKyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcu Y2dpP2lkPTMzMDUzCisKKyAgICAgICAgKiBydW50aW1lL0pTT05PYmplY3QuY3BwOgorICAgICAg ICAoSlNDOjpTdHJpbmdpZmllcjo6SG9sZGVyOjphcHBlbmROZXh0UHJvcGVydHkpOgorICAgICAg ICAoSlNDOjpXYWxrZXI6OndhbGspOgorCiAyMDEwLTAxLTA2ICBHYXZpbiBCYXJyYWNsb3VnaCAg PGJhcnJhY2xvdWdoQGFwcGxlLmNvbT4KIAogICAgICAgICBRVCBidWlsZCBmaXguCmRpZmYgLS1n aXQgYS9KYXZhU2NyaXB0Q29yZS9ydW50aW1lL0pTT05PYmplY3QuY3BwIGIvSmF2YVNjcmlwdENv cmUvcnVudGltZS9KU09OT2JqZWN0LmNwcAppbmRleCBmMjhlNzBlLi5jZTBkY2ZmIDEwMDY0NAot LS0gYS9KYXZhU2NyaXB0Q29yZS9ydW50aW1lL0pTT05PYmplY3QuY3BwCisrKyBiL0phdmFTY3Jp cHRDb3JlL3J1bnRpbWUvSlNPTk9iamVjdC5jcHAKQEAgLTUwMSw3ICs1MDEsNyBAQCBib29sIFN0 cmluZ2lmaWVyOjpIb2xkZXI6OmFwcGVuZE5leHRQcm9wZXJ0eShTdHJpbmdpZmllciYgc3RyaW5n aWZpZXIsIFN0cmluZ0J1aQogICAgICAgICAgICAgICAgIG1fcHJvcGVydHlOYW1lcyA9IHN0cmlu Z2lmaWVyLm1fYXJyYXlSZXBsYWNlclByb3BlcnR5TmFtZXMuZGF0YSgpOwogICAgICAgICAgICAg ZWxzZSB7CiAgICAgICAgICAgICAgICAgUHJvcGVydHlOYW1lQXJyYXkgb2JqZWN0UHJvcGVydHlO YW1lcyhleGVjKTsKLSAgICAgICAgICAgICAgICBtX29iamVjdC0+Z2V0UHJvcGVydHlOYW1lcyhl eGVjLCBvYmplY3RQcm9wZXJ0eU5hbWVzKTsKKyAgICAgICAgICAgICAgICBtX29iamVjdC0+Z2V0 T3duUHJvcGVydHlOYW1lcyhleGVjLCBvYmplY3RQcm9wZXJ0eU5hbWVzKTsKICAgICAgICAgICAg ICAgICBtX3Byb3BlcnR5TmFtZXMgPSBvYmplY3RQcm9wZXJ0eU5hbWVzLnJlbGVhc2VEYXRhKCk7 CiAgICAgICAgICAgICB9CiAgICAgICAgICAgICBtX3NpemUgPSBtX3Byb3BlcnR5TmFtZXMtPnBy b3BlcnR5TmFtZVZlY3RvcigpLnNpemUoKTsKQEAgLTc0Niw3ICs3NDYsNyBAQCBORVZFUl9JTkxJ TkUgSlNWYWx1ZSBXYWxrZXI6OndhbGsoSlNWYWx1ZSB1bmZpbHRlcmVkKQogICAgICAgICAgICAg ICAgIG9iamVjdFN0YWNrLmFwcGVuZChvYmplY3QpOwogICAgICAgICAgICAgICAgIGluZGV4U3Rh Y2suYXBwZW5kKDApOwogICAgICAgICAgICAgICAgIHByb3BlcnR5U3RhY2suYXBwZW5kKFByb3Bl cnR5TmFtZUFycmF5KG1fZXhlYykpOwotICAgICAgICAgICAgICAgIG9iamVjdC0+Z2V0UHJvcGVy dHlOYW1lcyhtX2V4ZWMsIHByb3BlcnR5U3RhY2subGFzdCgpKTsKKyAgICAgICAgICAgICAgICBv YmplY3QtPmdldE93blByb3BlcnR5TmFtZXMobV9leGVjLCBwcm9wZXJ0eVN0YWNrLmxhc3QoKSk7 CiAgICAgICAgICAgICAgICAgLy8gZmFsbHRocm91Z2gKICAgICAgICAgICAgIH0KICAgICAgICAg ICAgIG9iamVjdFN0YXJ0VmlzaXRNZW1iZXI6CmRpZmYgLS1naXQgYS9MYXlvdXRUZXN0cy9DaGFu Z2VMb2cgYi9MYXlvdXRUZXN0cy9DaGFuZ2VMb2cKaW5kZXggYzliZTQ4Ny4uODllOWQ1OCAxMDA2 NDQKLS0tIGEvTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCisrKyBiL0xheW91dFRlc3RzL0NoYW5nZUxv ZwpAQCAtMSwzICsxLDE3IEBACisyMDEwLTAxLTA3ICBLZW50IEhhbnNlbiAgPGtlbnQuaGFuc2Vu QG5va2lhLmNvbT4KKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAg ICAgICBKU09OLnN0cmluZ2lmeSBhbmQgSlNPTi5wYXJzZSBuZWVkbGVzc2x5IHByb2Nlc3MgcHJv cGVydGllcyBpbiB0aGUgcHJvdG90eXBlIGNoYWluCisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJr aXQub3JnL3Nob3dfYnVnLmNnaT9pZD0zMzA1MworCisgICAgICAgIEFkZCB0ZXN0cyB0aGF0IHZl cmlmeSB0aGF0IHByb3BlcnRpZXMgaW4gdGhlIHByb3RvdHlwZSBjaGFpbiBhcmUgaWdub3JlZC4K KworICAgICAgICAqIGZhc3QvanMvSlNPTi1wYXJzZS1leHBlY3RlZC50eHQ6CisgICAgICAgICog ZmFzdC9qcy9KU09OLXN0cmluZ2lmeS1leHBlY3RlZC50eHQ6CisgICAgICAgICogZmFzdC9qcy9y ZXNvdXJjZXMvSlNPTi1wYXJzZS5qczoKKyAgICAgICAgKiBmYXN0L2pzL3Jlc291cmNlcy9KU09O LXN0cmluZ2lmeS5qczoKKwogMjAxMC0wMS0wNiAgTU9SSVRBIEhhamltZSAgPG1vcnJpdGFAZ21h aWwuY29tPgogCiAgICAgICAgIFJldmlld2VkIGJ5IE5pa29sYXMgWmltbWVybWFubi4KZGlmZiAt LWdpdCBhL0xheW91dFRlc3RzL2Zhc3QvanMvSlNPTi1wYXJzZS1leHBlY3RlZC50eHQgYi9MYXlv dXRUZXN0cy9mYXN0L2pzL0pTT04tcGFyc2UtZXhwZWN0ZWQudHh0CmluZGV4IDg3ZDBkMmEuLjkz NjI5NTQgMTAwNjQ0Ci0tLSBhL0xheW91dFRlc3RzL2Zhc3QvanMvSlNPTi1wYXJzZS1leHBlY3Rl ZC50eHQKKysrIGIvTGF5b3V0VGVzdHMvZmFzdC9qcy9KU09OLXBhcnNlLWV4cGVjdGVkLnR4dApA QCAtNjcsNiArNjcsMTAgQEAgZnVuY3Rpb24gKGpzb25PYmplY3QpIHsKICAgICB9CiBQQVNTIEpT T04uc3RyaW5naWZ5KHRlc3RzW2ldKG5hdGl2ZUpTT04pKSBpcyBKU09OLnN0cmluZ2lmeSh0ZXN0 c1tpXShKU09OKSkKIGZ1bmN0aW9uIChqc29uT2JqZWN0KSB7CisgICAgICAgIHJldHVybiBqc29u T2JqZWN0LnBhcnNlKCd7Il9fcHJvdG9fXyI6NX0nKTsKKyAgICB9CitQQVNTIEpTT04uc3RyaW5n aWZ5KHRlc3RzW2ldKG5hdGl2ZUpTT04pKSBpcyBKU09OLnN0cmluZ2lmeSh0ZXN0c1tpXShKU09O KSkKK2Z1bmN0aW9uIChqc29uT2JqZWN0KSB7CiAgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnBh cnNlKCd7ImEiOjUsfScpOwogICAgIH0KIFBBU1MgdGVzdHNbaV0obmF0aXZlSlNPTikgdGhyZXcg ZXhjZXB0aW9uIFN5bnRheEVycm9yOiBVbmFibGUgdG8gcGFyc2UgSlNPTiBzdHJpbmcuCkBAIC00 MzYsNiArNDQwLDEwIEBAIGZ1bmN0aW9uIChqc29uT2JqZWN0KSB7CiAgICAgfQogUEFTUyBKU09O LnN0cmluZ2lmeSh0ZXN0c1tpXShuYXRpdmVKU09OKSkgaXMgSlNPTi5zdHJpbmdpZnkodGVzdHNb aV0oSlNPTikpCiBmdW5jdGlvbiAoanNvbk9iamVjdCkgeworICAgICAgICByZXR1cm4ganNvbk9i amVjdC5wYXJzZSgneyJfX3Byb3RvX18iOnsiYSI6NX19JywgbG9nKTsKKyAgICB9CitQQVNTIEpT T04uc3RyaW5naWZ5KHRlc3RzW2ldKG5hdGl2ZUpTT04pKSBpcyBKU09OLnN0cmluZ2lmeSh0ZXN0 c1tpXShKU09OKSkKK2Z1bmN0aW9uIChqc29uT2JqZWN0KSB7CiAgICAgICAgIGxvZ09yZGVyU3Ry aW5nID0gIiI7CiAgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnBhcnNlKCJ0cnVlIiwgbG9nT3Jk ZXIpOwogICAgIH0KZGlmZiAtLWdpdCBhL0xheW91dFRlc3RzL2Zhc3QvanMvSlNPTi1zdHJpbmdp ZnktZXhwZWN0ZWQudHh0IGIvTGF5b3V0VGVzdHMvZmFzdC9qcy9KU09OLXN0cmluZ2lmeS1leHBl Y3RlZC50eHQKaW5kZXggMzJhYWQ4Ny4uOTE0NmZiNSAxMDA2NDQKLS0tIGEvTGF5b3V0VGVzdHMv ZmFzdC9qcy9KU09OLXN0cmluZ2lmeS1leHBlY3RlZC50eHQKKysrIGIvTGF5b3V0VGVzdHMvZmFz dC9qcy9KU09OLXN0cmluZ2lmeS1leHBlY3RlZC50eHQKQEAgLTE2NCw2ICsxNjQsMTIgQEAgZnVu Y3Rpb24gKGpzb25PYmplY3QpIHsKIFBBU1MgdGVzdHNbaV0obmF0aXZlSlNPTikgaXMgdGVzdHNb aV0oSlNPTikKIGZ1bmN0aW9uIChqc29uT2JqZWN0KSB7CiAgICAgICAgIHZhciBhbGxTdHJpbmcg PSB0cnVlOworICAgICAgICB2YXIgYXJyYXkgPSBbe3RvU3RyaW5nOmZ1bmN0aW9uKCl7YXJyYXlb MF09J2EnOyBhcnJheVsxXT0nYyc7IGFycmF5WzJdPSdiJzsgcmV0dXJuICdhJ319XTsKKyAgICAg ICAgcmV0dXJuIGpzb25PYmplY3Quc3RyaW5naWZ5KHNpbXBsZU9iamVjdFdpdGhQcm90bywgYXJy YXkpOworICAgIH0KK1BBU1MgdGVzdHNbaV0obmF0aXZlSlNPTikgaXMgdGVzdHNbaV0oSlNPTikK K2Z1bmN0aW9uIChqc29uT2JqZWN0KSB7CisgICAgICAgIHZhciBhbGxTdHJpbmcgPSB0cnVlOwog ICAgICAgICB2YXIgYXJyYXkgPSBbMSwgbmV3IE51bWJlcigyKSwgTmFOLCBJbmZpbml0eSwgLUlu ZmluaXR5LCBuZXcgU3RyaW5nKCJzdHIiKV07CiAgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0 cmluZ2lmeSh7IjEiOiIxIiwiMiI6IjIiLCJOYU4iOiJOYU4iLCJJbmZpbml0eSI6IkluZmluaXR5 IiwiLUluZmluaXR5IjoiLUluZmluaXR5Iiwic3RyIjoic3RyIn0sIGFycmF5KTsKICAgICB9CkBA IC0yNjEsNiArMjY3LDEwOCBAQCBmdW5jdGlvbiAoanNvbk9iamVjdCkgewogICAgIH0KIFBBU1Mg dGVzdHNbaV0obmF0aXZlSlNPTikgaXMgdGVzdHNbaV0oSlNPTikKIGZ1bmN0aW9uIChqc29uT2Jq ZWN0KSB7CisgICAgICAgIHZhciBhbGxTdHJpbmcgPSB0cnVlOworICAgICAgICB2YXIgYXJyYXkg PSBbIjEiLCIyIiwiMyJdOworICAgICAgICByZXR1cm4ganNvbk9iamVjdC5zdHJpbmdpZnkoc2lt cGxlQXJyYXlXaXRoUHJvdG8sIGFycmF5KTsKKyAgICB9CitQQVNTIHRlc3RzW2ldKG5hdGl2ZUpT T04pIGlzIHRlc3RzW2ldKEpTT04pCitmdW5jdGlvbiAoanNvbk9iamVjdCkgeworICAgICAgICBy ZXR1cm4ganNvbk9iamVjdC5zdHJpbmdpZnkoc2ltcGxlQXJyYXlXaXRoUHJvdG8sIG51bGwsICIg ICIpOworICAgIH0KK1BBU1MgdGVzdHNbaV0obmF0aXZlSlNPTikgaXMgdGVzdHNbaV0oSlNPTikK K2Z1bmN0aW9uIChqc29uT2JqZWN0KSB7CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmlu Z2lmeShzaW1wbGVBcnJheVdpdGhQcm90bywgbnVsbCwgNCk7CisgICAgfQorUEFTUyB0ZXN0c1tp XShuYXRpdmVKU09OKSBpcyB0ZXN0c1tpXShKU09OKQorZnVuY3Rpb24gKGpzb25PYmplY3QpIHsK KyAgICAgICAgcmV0dXJuIGpzb25PYmplY3Quc3RyaW5naWZ5KHNpbXBsZUFycmF5V2l0aFByb3Rv LCBudWxsLCAiYWIiKTsKKyAgICB9CitQQVNTIHRlc3RzW2ldKG5hdGl2ZUpTT04pIGlzIHRlc3Rz W2ldKEpTT04pCitmdW5jdGlvbiAoanNvbk9iamVjdCkgeworICAgICAgICByZXR1cm4ganNvbk9i amVjdC5zdHJpbmdpZnkoc2ltcGxlQXJyYXlXaXRoUHJvdG8sIG51bGwsIDQpOworICAgIH0KK1BB U1MgdGVzdHNbaV0obmF0aXZlSlNPTikgaXMgdGVzdHNbaV0oSlNPTikKK2Z1bmN0aW9uIChqc29u T2JqZWN0KSB7CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmluZ2lmeShzaW1wbGVPYmpl Y3RXaXRoUHJvdG8sIG51bGwsICIgICIpOworICAgIH0KK1BBU1MgdGVzdHNbaV0obmF0aXZlSlNP TikgaXMgdGVzdHNbaV0oSlNPTikKK2Z1bmN0aW9uIChqc29uT2JqZWN0KSB7CisgICAgICAgIHJl dHVybiBqc29uT2JqZWN0LnN0cmluZ2lmeShzaW1wbGVPYmplY3RXaXRoUHJvdG8sIG51bGwsIDQp OworICAgIH0KK1BBU1MgdGVzdHNbaV0obmF0aXZlSlNPTikgaXMgdGVzdHNbaV0oSlNPTikKK2Z1 bmN0aW9uIChqc29uT2JqZWN0KSB7CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmluZ2lm eShzaW1wbGVPYmplY3RXaXRoUHJvdG8sIG51bGwsICJhYiIpOworICAgIH0KK1BBU1MgdGVzdHNb aV0obmF0aXZlSlNPTikgaXMgdGVzdHNbaV0oSlNPTikKK2Z1bmN0aW9uIChqc29uT2JqZWN0KSB7 CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmluZ2lmeShzaW1wbGVPYmplY3RXaXRoUHJv dG8sIG51bGwsIDQpOworICAgIH0KK1BBU1MgdGVzdHNbaV0obmF0aXZlSlNPTikgaXMgdGVzdHNb aV0oSlNPTikKK2Z1bmN0aW9uIChqc29uT2JqZWN0KSB7CisgICAgICAgIHJldHVybiBqc29uT2Jq ZWN0LnN0cmluZ2lmeShzaW1wbGVPYmplY3RXaXRoUHJvdG8sIG51bGwsIDEwKTsKKyAgICB9CitQ QVNTIHRlc3RzW2ldKG5hdGl2ZUpTT04pIGlzIHRlc3RzW2ldKEpTT04pCitmdW5jdGlvbiAoanNv bk9iamVjdCkgeworICAgICAgICByZXR1cm4ganNvbk9iamVjdC5zdHJpbmdpZnkoc2ltcGxlT2Jq ZWN0V2l0aFByb3RvLCBudWxsLCAxMSk7CisgICAgfQorUEFTUyB0ZXN0c1tpXShuYXRpdmVKU09O KSBpcyB0ZXN0c1tpXS5leHBlY3RlZAorZnVuY3Rpb24gKGpzb25PYmplY3QpIHsKKyAgICAgICAg cmV0dXJuIGpzb25PYmplY3Quc3RyaW5naWZ5KHNpbXBsZU9iamVjdFdpdGhQcm90bywgbnVsbCwg IiAgICAgICAgICAiKTsKKyAgICB9CitQQVNTIHRlc3RzW2ldKG5hdGl2ZUpTT04pIGlzIHRlc3Rz W2ldLmV4cGVjdGVkCitmdW5jdGlvbiAoanNvbk9iamVjdCkgeworICAgICAgICByZXR1cm4ganNv bk9iamVjdC5zdHJpbmdpZnkoc2ltcGxlT2JqZWN0V2l0aFByb3RvLCBudWxsLCAiICAgICAgICAg ICAiKTsKKyAgICB9CitQQVNTIHRlc3RzW2ldKG5hdGl2ZUpTT04pIGlzIHRlc3RzW2ldLmV4cGVj dGVkCitmdW5jdGlvbiAoanNvbk9iamVjdCkgeworICAgICAgICByZXR1cm4ganNvbk9iamVjdC5z dHJpbmdpZnkoY29tcGxleEFycmF5V2l0aFByb3RvLCBudWxsLCAiICAiKTsKKyAgICB9CitQQVNT IHRlc3RzW2ldKG5hdGl2ZUpTT04pIGlzIHRlc3RzW2ldKEpTT04pCitmdW5jdGlvbiAoanNvbk9i amVjdCkgeworICAgICAgICByZXR1cm4ganNvbk9iamVjdC5zdHJpbmdpZnkoY29tcGxleEFycmF5 V2l0aFByb3RvLCBudWxsLCA0KTsKKyAgICB9CitQQVNTIHRlc3RzW2ldKG5hdGl2ZUpTT04pIGlz IHRlc3RzW2ldKEpTT04pCitmdW5jdGlvbiAoanNvbk9iamVjdCkgeworICAgICAgICByZXR1cm4g anNvbk9iamVjdC5zdHJpbmdpZnkoY29tcGxleEFycmF5V2l0aFByb3RvLCBudWxsLCAiYWIiKTsK KyAgICB9CitQQVNTIHRlc3RzW2ldKG5hdGl2ZUpTT04pIGlzIHRlc3RzW2ldKEpTT04pCitmdW5j dGlvbiAoanNvbk9iamVjdCkgeworICAgICAgICByZXR1cm4ganNvbk9iamVjdC5zdHJpbmdpZnko Y29tcGxleEFycmF5V2l0aFByb3RvLCBudWxsLCA0KTsKKyAgICB9CitQQVNTIHRlc3RzW2ldKG5h dGl2ZUpTT04pIGlzIHRlc3RzW2ldKEpTT04pCitmdW5jdGlvbiAoanNvbk9iamVjdCkgeworICAg ICAgICByZXR1cm4ganNvbk9iamVjdC5zdHJpbmdpZnkoY29tcGxleE9iamVjdFdpdGhQcm90bywg bnVsbCwgIiAgIik7CisgICAgfQorUEFTUyB0ZXN0c1tpXShuYXRpdmVKU09OKSBpcyB0ZXN0c1tp XShKU09OKQorZnVuY3Rpb24gKGpzb25PYmplY3QpIHsKKyAgICAgICAgcmV0dXJuIGpzb25PYmpl Y3Quc3RyaW5naWZ5KGNvbXBsZXhPYmplY3RXaXRoUHJvdG8sIG51bGwsIDQpOworICAgIH0KK1BB U1MgdGVzdHNbaV0obmF0aXZlSlNPTikgaXMgdGVzdHNbaV0oSlNPTikKK2Z1bmN0aW9uIChqc29u T2JqZWN0KSB7CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmluZ2lmeShjb21wbGV4T2Jq ZWN0V2l0aFByb3RvLCBudWxsLCAiYWIiKTsKKyAgICB9CitQQVNTIHRlc3RzW2ldKG5hdGl2ZUpT T04pIGlzIHRlc3RzW2ldKEpTT04pCitmdW5jdGlvbiAoanNvbk9iamVjdCkgeworICAgICAgICBy ZXR1cm4ganNvbk9iamVjdC5zdHJpbmdpZnkoY29tcGxleE9iamVjdFdpdGhQcm90bywgbnVsbCwg NCk7CisgICAgfQorUEFTUyB0ZXN0c1tpXShuYXRpdmVKU09OKSBpcyB0ZXN0c1tpXShKU09OKQor ZnVuY3Rpb24gKGpzb25PYmplY3QpIHsKKyAgICAgICAgcmV0dXJuIGpzb25PYmplY3Quc3RyaW5n aWZ5KG9iamVjdFdpdGhTaWRlRWZmZWN0R2V0dGVyKTsKKyAgICB9CitGQUlMIHRlc3RzW2ldKG5h dGl2ZUpTT04pIHNob3VsZCBiZSB7ImZvbyI6MX0uIFdhcyB7fS4KK2Z1bmN0aW9uIChqc29uT2Jq ZWN0KSB7CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmluZ2lmeShvYmplY3RXaXRoU2lk ZUVmZmVjdEdldHRlckFuZFByb3RvKTsKKyAgICB9CitGQUlMIHRlc3RzW2ldKG5hdGl2ZUpTT04p IHNob3VsZCBiZSB7ImZvbyI6MX0uIFdhcyB7fS4KK2Z1bmN0aW9uIChqc29uT2JqZWN0KSB7Cisg ICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmluZ2lmeShhcnJheVdpdGhTaWRlRWZmZWN0R2V0 dGVyKTsKKyAgICB9CitQQVNTIHRlc3RzW2ldKG5hdGl2ZUpTT04pIGlzIHRlc3RzW2ldKEpTT04p CitmdW5jdGlvbiAoanNvbk9iamVjdCkgeworICAgICAgICByZXR1cm4ganNvbk9iamVjdC5zdHJp bmdpZnkoYXJyYXlXaXRoU2lkZUVmZmVjdEdldHRlckFuZFByb3RvKTsKKyAgICB9CitQQVNTIHRl c3RzW2ldKG5hdGl2ZUpTT04pIGlzIHRlc3RzW2ldKEpTT04pCitmdW5jdGlvbiAoanNvbk9iamVj dCkgewogICAgICAgICByZXBsYWNlVHJhY2tlciA9ICIiOwogICAgICAgICBqc29uT2JqZWN0LnN0 cmluZ2lmeShbMSwyLDMsLCwsNCw1LDZdLCByZXBsYWNlRnVuYyk7CiAgICAgICAgIHJldHVybiBy ZXBsYWNlVHJhY2tlcjsKZGlmZiAtLWdpdCBhL0xheW91dFRlc3RzL2Zhc3QvanMvcmVzb3VyY2Vz L0pTT04tcGFyc2UuanMgYi9MYXlvdXRUZXN0cy9mYXN0L2pzL3Jlc291cmNlcy9KU09OLXBhcnNl LmpzCmluZGV4IGU4M2NiNzAuLmI3OGRmMjUgMTAwNjQ0Ci0tLSBhL0xheW91dFRlc3RzL2Zhc3Qv anMvcmVzb3VyY2VzL0pTT04tcGFyc2UuanMKKysrIGIvTGF5b3V0VGVzdHMvZmFzdC9qcy9yZXNv dXJjZXMvSlNPTi1wYXJzZS5qcwpAQCAtNjUsNiArNjUsOSBAQCBmdW5jdGlvbiBjcmVhdGVUZXN0 cygpIHsKICAgICAgICAgcmV0dXJuIGpzb25PYmplY3QucGFyc2UoJ3siYSI6NX0nKTsKICAgICB9 KTsKICAgICByZXN1bHQucHVzaChmdW5jdGlvbihqc29uT2JqZWN0KXsKKyAgICAgICAgcmV0dXJu IGpzb25PYmplY3QucGFyc2UoJ3siX19wcm90b19fIjo1fScpOworICAgIH0pOworICAgIHJlc3Vs dC5wdXNoKGZ1bmN0aW9uKGpzb25PYmplY3QpewogICAgICAgICByZXR1cm4ganNvbk9iamVjdC5w YXJzZSgneyJhIjo1LH0nKTsKICAgICB9KTsKICAgICByZXN1bHRbcmVzdWx0Lmxlbmd0aCAtIDFd LnRocm93cyA9IHRydWU7CkBAIC0zOTYsNiArMzk5LDkgQEAgZnVuY3Rpb24gY3JlYXRlVGVzdHMo KSB7CiAgICAgcmVzdWx0LnB1c2goZnVuY3Rpb24oanNvbk9iamVjdCl7CiAgICAgICAgIHJldHVy biBqc29uT2JqZWN0LnBhcnNlKEpTT04uc3RyaW5naWZ5KGNvbXBsZXhPYmplY3QpLCBsb2cpOwog ICAgIH0pOworICAgIHJlc3VsdC5wdXNoKGZ1bmN0aW9uKGpzb25PYmplY3QpeworICAgICAgICBy ZXR1cm4ganNvbk9iamVjdC5wYXJzZSgneyJfX3Byb3RvX18iOnsiYSI6NX19JywgbG9nKTsKKyAg ICB9KTsKICAgICB2YXIgbG9nT3JkZXJTdHJpbmc7CiAgICAgZnVuY3Rpb24gbG9nT3JkZXIoa2V5 LCB2YWx1ZSkgewogICAgICAgICBsb2dPcmRlclN0cmluZyArPSBrZXkgKyI6IitKU09OLnN0cmlu Z2lmeSh2YWx1ZSk7CmRpZmYgLS1naXQgYS9MYXlvdXRUZXN0cy9mYXN0L2pzL3Jlc291cmNlcy9K U09OLXN0cmluZ2lmeS5qcyBiL0xheW91dFRlc3RzL2Zhc3QvanMvcmVzb3VyY2VzL0pTT04tc3Ry aW5naWZ5LmpzCmluZGV4IGUwOWY3NDguLmQ1OGI1ZmUgMTAwNjQ0Ci0tLSBhL0xheW91dFRlc3Rz L2Zhc3QvanMvcmVzb3VyY2VzL0pTT04tc3RyaW5naWZ5LmpzCisrKyBiL0xheW91dFRlc3RzL2Zh c3QvanMvcmVzb3VyY2VzL0pTT04tc3RyaW5naWZ5LmpzCkBAIC00LDYgKzQsMTkgQEAgZnVuY3Rp b24gY3JlYXRlVGVzdHMoKSB7CiAgICAgdmFyIHNpbXBsZU9iamVjdCA9IHthOiIxIiwgYjoiMiIs IGM6IjMifTsKICAgICB2YXIgY29tcGxleEFycmF5ID0gWydhJywgJ2InLCAnYycsLCxzaW1wbGVP YmplY3QsIHNpbXBsZUFycmF5LCBbc2ltcGxlT2JqZWN0LHNpbXBsZUFycmF5XV07CiAgICAgdmFy IGNvbXBsZXhPYmplY3QgPSB7YToiMSIsIGI6IjIiLCBjOiIzIiwgZDp1bmRlZmluZWQsIGU6bnVs bCwgIiI6MTIsIGdldCBmKCl7IHJldHVybiBzaW1wbGVBcnJheTsgfSwgYXJyYXk6IGNvbXBsZXhB cnJheX07CisgICAgdmFyIHNpbXBsZUFycmF5V2l0aFByb3RvID0gWydkJywgJ2UnLCAnZiddOwor ICAgIHNpbXBsZUFycmF5V2l0aFByb3RvLl9fcHJvdG9fXyA9IHNpbXBsZU9iamVjdDsKKyAgICB2 YXIgc2ltcGxlT2JqZWN0V2l0aFByb3RvID0ge2Q6IjQiLCBlOiI1IiwgZjoiNiIsIF9fcHJvdG9f XzpzaW1wbGVPYmplY3R9OworICAgIHZhciBjb21wbGV4QXJyYXlXaXRoUHJvdG8gPSBbJ2QnLCAn ZScsICdmJywsLHNpbXBsZU9iamVjdFdpdGhQcm90bywgc2ltcGxlQXJyYXlXaXRoUHJvdG8sIFtz aW1wbGVPYmplY3RXaXRoUHJvdG8sc2ltcGxlQXJyYXlXaXRoUHJvdG9dXTsKKyAgICBjb21wbGV4 QXJyYXlXaXRoUHJvdG8uX19wcm90b19fID0gc2ltcGxlT2JqZWN0V2l0aFByb3RvOworICAgIHZh ciBjb21wbGV4T2JqZWN0V2l0aFByb3RvID0ge2Q6IjQiLCBlOiI1IiwgZjoiNiIsIGc6dW5kZWZp bmVkLCBoOm51bGwsICIiOjEyLCBnZXQgaSgpeyByZXR1cm4gc2ltcGxlQXJyYXlXaXRoUHJvdG87 IH0sIGFycmF5MjogY29tcGxleEFycmF5V2l0aFByb3RvLCBfX3Byb3RvX186Y29tcGxleE9iamVj dH07CisgICAgdmFyIG9iamVjdFdpdGhTaWRlRWZmZWN0R2V0dGVyID0ge2dldCBiKCkge3RoaXMu Zm9vPTE7fX07CisgICAgdmFyIG9iamVjdFdpdGhTaWRlRWZmZWN0R2V0dGVyQW5kUHJvdG8gPSB7 X19wcm90b19fOntmb286ImJhciJ9LCBnZXQgYigpIHt0aGlzLmZvbz0xO319OworICAgIHZhciBh cnJheVdpdGhTaWRlRWZmZWN0R2V0dGVyID0gW107CisgICAgYXJyYXlXaXRoU2lkZUVmZmVjdEdl dHRlci5fX2RlZmluZUdldHRlcl9fKCJiIiwgZnVuY3Rpb24oKXt0aGlzLmZvbz0xO30pOworICAg IHZhciBhcnJheVdpdGhTaWRlRWZmZWN0R2V0dGVyQW5kUHJvdG8gPSBbXTsKKyAgICBhcnJheVdp dGhTaWRlRWZmZWN0R2V0dGVyQW5kUHJvdG8uX19kZWZpbmVHZXR0ZXJfXygiYiIsIGZ1bmN0aW9u KCl7dGhpcy5mb289MTt9KTsKKyAgICBhcnJheVdpdGhTaWRlRWZmZWN0R2V0dGVyQW5kUHJvdG8u X19wcm90b19fID0ge2ZvbzoiYmFyIn07CiAgICAgdmFyIHJlc3VsdCA9IFtdOwogICAgIHJlc3Vs dC5wdXNoKGZ1bmN0aW9uKGpzb25PYmplY3QpewogICAgICAgICByZXR1cm4ganNvbk9iamVjdC5z dHJpbmdpZnkoMSk7CkBAIC0xNDMsNiArMTU2LDExIEBAIGZ1bmN0aW9uIGNyZWF0ZVRlc3RzKCkg ewogICAgIH0pOwogICAgIHJlc3VsdC5wdXNoKGZ1bmN0aW9uKGpzb25PYmplY3QpewogICAgICAg ICB2YXIgYWxsU3RyaW5nID0gdHJ1ZTsKKyAgICAgICAgdmFyIGFycmF5ID0gW3t0b1N0cmluZzpm dW5jdGlvbigpe2FycmF5WzBdPSdhJzsgYXJyYXlbMV09J2MnOyBhcnJheVsyXT0nYic7IHJldHVy biAnYSd9fV07CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmluZ2lmeShzaW1wbGVPYmpl Y3RXaXRoUHJvdG8sIGFycmF5KTsKKyAgICB9KTsKKyAgICByZXN1bHQucHVzaChmdW5jdGlvbihq c29uT2JqZWN0KXsKKyAgICAgICAgdmFyIGFsbFN0cmluZyA9IHRydWU7CiAgICAgICAgIHZhciBh cnJheSA9IFsxLCBuZXcgTnVtYmVyKDIpLCBOYU4sIEluZmluaXR5LCAtSW5maW5pdHksIG5ldyBT dHJpbmcoInN0ciIpXTsKICAgICAgICAgcmV0dXJuIGpzb25PYmplY3Quc3RyaW5naWZ5KHsiMSI6 IjEiLCIyIjoiMiIsIk5hTiI6Ik5hTiIsIkluZmluaXR5IjoiSW5maW5pdHkiLCItSW5maW5pdHki OiItSW5maW5pdHkiLCJzdHIiOiJzdHIifSwgYXJyYXkpOwogICAgIH0pOwpAQCAtMjIwLDYgKzIz OCw4NiBAQCBmdW5jdGlvbiBjcmVhdGVUZXN0cygpIHsKICAgICByZXN1bHQucHVzaChmdW5jdGlv bihqc29uT2JqZWN0KXsKICAgICAgICAgcmV0dXJuIGpzb25PYmplY3Quc3RyaW5naWZ5KGNvbXBs ZXhPYmplY3QsIG51bGwsIDQpOwogICAgIH0pOworICAgIHJlc3VsdC5wdXNoKGZ1bmN0aW9uKGpz b25PYmplY3QpeworICAgICAgICB2YXIgYWxsU3RyaW5nID0gdHJ1ZTsKKyAgICAgICAgdmFyIGFy cmF5ID0gWyIxIiwiMiIsIjMiXTsKKyAgICAgICAgcmV0dXJuIGpzb25PYmplY3Quc3RyaW5naWZ5 KHNpbXBsZUFycmF5V2l0aFByb3RvLCBhcnJheSk7CisgICAgfSk7CisgICAgcmVzdWx0LnB1c2go ZnVuY3Rpb24oanNvbk9iamVjdCl7CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmluZ2lm eShzaW1wbGVBcnJheVdpdGhQcm90bywgbnVsbCwgIiAgIik7CisgICAgfSk7CisgICAgcmVzdWx0 LnB1c2goZnVuY3Rpb24oanNvbk9iamVjdCl7CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0 cmluZ2lmeShzaW1wbGVBcnJheVdpdGhQcm90bywgbnVsbCwgNCk7CisgICAgfSk7CisgICAgcmVz dWx0LnB1c2goZnVuY3Rpb24oanNvbk9iamVjdCl7CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0 LnN0cmluZ2lmeShzaW1wbGVBcnJheVdpdGhQcm90bywgbnVsbCwgImFiIik7CisgICAgfSk7Cisg ICAgcmVzdWx0LnB1c2goZnVuY3Rpb24oanNvbk9iamVjdCl7CisgICAgICAgIHJldHVybiBqc29u T2JqZWN0LnN0cmluZ2lmeShzaW1wbGVBcnJheVdpdGhQcm90bywgbnVsbCwgNCk7CisgICAgfSk7 CisgICAgcmVzdWx0LnB1c2goZnVuY3Rpb24oanNvbk9iamVjdCl7CisgICAgICAgIHJldHVybiBq c29uT2JqZWN0LnN0cmluZ2lmeShzaW1wbGVPYmplY3RXaXRoUHJvdG8sIG51bGwsICIgICIpOwor ICAgIH0pOworICAgIHJlc3VsdC5wdXNoKGZ1bmN0aW9uKGpzb25PYmplY3QpeworICAgICAgICBy ZXR1cm4ganNvbk9iamVjdC5zdHJpbmdpZnkoc2ltcGxlT2JqZWN0V2l0aFByb3RvLCBudWxsLCA0 KTsKKyAgICB9KTsKKyAgICByZXN1bHQucHVzaChmdW5jdGlvbihqc29uT2JqZWN0KXsKKyAgICAg ICAgcmV0dXJuIGpzb25PYmplY3Quc3RyaW5naWZ5KHNpbXBsZU9iamVjdFdpdGhQcm90bywgbnVs bCwgImFiIik7CisgICAgfSk7CisgICAgcmVzdWx0LnB1c2goZnVuY3Rpb24oanNvbk9iamVjdCl7 CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmluZ2lmeShzaW1wbGVPYmplY3RXaXRoUHJv dG8sIG51bGwsIDQpOworICAgIH0pOworICAgIHJlc3VsdC5wdXNoKGZ1bmN0aW9uKGpzb25PYmpl Y3QpeworICAgICAgICByZXR1cm4ganNvbk9iamVjdC5zdHJpbmdpZnkoc2ltcGxlT2JqZWN0V2l0 aFByb3RvLCBudWxsLCAxMCk7CisgICAgfSk7CisgICAgcmVzdWx0LnB1c2goZnVuY3Rpb24oanNv bk9iamVjdCl7CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmluZ2lmeShzaW1wbGVPYmpl Y3RXaXRoUHJvdG8sIG51bGwsIDExKTsKKyAgICB9KTsKKyAgICByZXN1bHRbcmVzdWx0Lmxlbmd0 aCAtIDFdLmV4cGVjdGVkID0gSlNPTi5zdHJpbmdpZnkoc2ltcGxlT2JqZWN0V2l0aFByb3RvLCBu dWxsLCAxMCk7CisgICAgcmVzdWx0LnB1c2goZnVuY3Rpb24oanNvbk9iamVjdCl7CisgICAgICAg IHJldHVybiBqc29uT2JqZWN0LnN0cmluZ2lmeShzaW1wbGVPYmplY3RXaXRoUHJvdG8sIG51bGws ICIgICAgICAgICAgIik7CisgICAgfSk7CisgICAgcmVzdWx0W3Jlc3VsdC5sZW5ndGggLSAxXS5l eHBlY3RlZCA9IEpTT04uc3RyaW5naWZ5KHNpbXBsZU9iamVjdFdpdGhQcm90bywgbnVsbCwgMTAp OworICAgIHJlc3VsdC5wdXNoKGZ1bmN0aW9uKGpzb25PYmplY3QpeworICAgICAgICByZXR1cm4g anNvbk9iamVjdC5zdHJpbmdpZnkoc2ltcGxlT2JqZWN0V2l0aFByb3RvLCBudWxsLCAiICAgICAg ICAgICAiKTsKKyAgICB9KTsKKyAgICByZXN1bHRbcmVzdWx0Lmxlbmd0aCAtIDFdLmV4cGVjdGVk ID0gSlNPTi5zdHJpbmdpZnkoc2ltcGxlT2JqZWN0V2l0aFByb3RvLCBudWxsLCAxMCk7CisgICAg cmVzdWx0LnB1c2goZnVuY3Rpb24oanNvbk9iamVjdCl7CisgICAgICAgIHJldHVybiBqc29uT2Jq ZWN0LnN0cmluZ2lmeShjb21wbGV4QXJyYXlXaXRoUHJvdG8sIG51bGwsICIgICIpOworICAgIH0p OworICAgIHJlc3VsdC5wdXNoKGZ1bmN0aW9uKGpzb25PYmplY3QpeworICAgICAgICByZXR1cm4g anNvbk9iamVjdC5zdHJpbmdpZnkoY29tcGxleEFycmF5V2l0aFByb3RvLCBudWxsLCA0KTsKKyAg ICB9KTsKKyAgICByZXN1bHQucHVzaChmdW5jdGlvbihqc29uT2JqZWN0KXsKKyAgICAgICAgcmV0 dXJuIGpzb25PYmplY3Quc3RyaW5naWZ5KGNvbXBsZXhBcnJheVdpdGhQcm90bywgbnVsbCwgImFi Iik7CisgICAgfSk7CisgICAgcmVzdWx0LnB1c2goZnVuY3Rpb24oanNvbk9iamVjdCl7CisgICAg ICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmluZ2lmeShjb21wbGV4QXJyYXlXaXRoUHJvdG8sIG51 bGwsIDQpOworICAgIH0pOworICAgIHJlc3VsdC5wdXNoKGZ1bmN0aW9uKGpzb25PYmplY3Qpewor ICAgICAgICByZXR1cm4ganNvbk9iamVjdC5zdHJpbmdpZnkoY29tcGxleE9iamVjdFdpdGhQcm90 bywgbnVsbCwgIiAgIik7CisgICAgfSk7CisgICAgcmVzdWx0LnB1c2goZnVuY3Rpb24oanNvbk9i amVjdCl7CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmluZ2lmeShjb21wbGV4T2JqZWN0 V2l0aFByb3RvLCBudWxsLCA0KTsKKyAgICB9KTsKKyAgICByZXN1bHQucHVzaChmdW5jdGlvbihq c29uT2JqZWN0KXsKKyAgICAgICAgcmV0dXJuIGpzb25PYmplY3Quc3RyaW5naWZ5KGNvbXBsZXhP YmplY3RXaXRoUHJvdG8sIG51bGwsICJhYiIpOworICAgIH0pOworICAgIHJlc3VsdC5wdXNoKGZ1 bmN0aW9uKGpzb25PYmplY3QpeworICAgICAgICByZXR1cm4ganNvbk9iamVjdC5zdHJpbmdpZnko Y29tcGxleE9iamVjdFdpdGhQcm90bywgbnVsbCwgNCk7CisgICAgfSk7CisgICAgcmVzdWx0LnB1 c2goZnVuY3Rpb24oanNvbk9iamVjdCl7CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmlu Z2lmeShvYmplY3RXaXRoU2lkZUVmZmVjdEdldHRlcik7CisgICAgfSk7CisgICAgcmVzdWx0LnB1 c2goZnVuY3Rpb24oanNvbk9iamVjdCl7CisgICAgICAgIHJldHVybiBqc29uT2JqZWN0LnN0cmlu Z2lmeShvYmplY3RXaXRoU2lkZUVmZmVjdEdldHRlckFuZFByb3RvKTsKKyAgICB9KTsKKyAgICBy ZXN1bHQucHVzaChmdW5jdGlvbihqc29uT2JqZWN0KXsKKyAgICAgICAgcmV0dXJuIGpzb25PYmpl Y3Quc3RyaW5naWZ5KGFycmF5V2l0aFNpZGVFZmZlY3RHZXR0ZXIpOworICAgIH0pOworICAgIHJl c3VsdC5wdXNoKGZ1bmN0aW9uKGpzb25PYmplY3QpeworICAgICAgICByZXR1cm4ganNvbk9iamVj dC5zdHJpbmdpZnkoYXJyYXlXaXRoU2lkZUVmZmVjdEdldHRlckFuZFByb3RvKTsKKyAgICB9KTsK ICAgICB2YXIgcmVwbGFjZVRyYWNrZXI7CiAgICAgZnVuY3Rpb24gcmVwbGFjZUZ1bmMoa2V5LCB2 YWx1ZSkgewogICAgICAgICByZXBsYWNlVHJhY2tlciArPSBrZXkgKyAiKCIrKHR5cGVvZiBrZXkp KyIpIiArIEpTT04uc3RyaW5naWZ5KHZhbHVlKSArICI7IjsK