41480 2010-07-01 12:50:52 -0700 [Qt][Symbian] Variable max heap size between target/emulator 2010-10-20 07:31:23 -0700 1 1 1 Unclassified WebKit Platform 528+ (Nightly build) PC All RESOLVED FIXED Qt P2 Normal --- 1 laszlo.gombos s.mathur commit-queue hausmann koshuin laszlo.gombos s.mathur oldest_to_newest 245497 0 laszlo.gombos 2010-07-01 12:50:52 -0700 Siddharth is telling me that on the simulator 96 MB for max heapsize it to high. The limit in WebKit.pri should be different for WINSCW than for target. Patches welcome. 245510 1 s.mathur 2010-07-01 13:10:43 -0700 (In reply to comment #0) > Siddharth is telling me that on the simulator 96 MB for max heapsize it to high. The limit in WebKit.pri should be different for WINSCW than for target. Patches welcome. How about I kill many birds with one stone, and make a QMake function that _all_ Symbian EXEs can use to set their heap sizes? Also turn on code paging and data paging automagically? 245522 2 s.mathur 2010-07-01 13:34:57 -0700 I rephrased the title to better reflect that I am doing. 245543 3 laszlo.gombos 2010-07-01 14:20:31 -0700 Siddharth, we're already consistent I think as the HEAPSIZE is set currently for all .exe's that are built from the WebKit tree (WebKit.pri), including tests and DumpRenderTree exe's. Also note that I do not know what are the Symbian defaults for code paging and data paging. 245570 4 60293 s.mathur 2010-07-01 15:02:20 -0700 Created attachment 60293 Make max heap variable for ARM/emulator. Also use paging Make max heap variable for ARM, emulator so that developers don't see jammed emulator. Also use paging. OSes that don't support the bit flag in the EXE will silently ignore it. Verified using QTestBrowser and other test EXEs that directives show up in MMP and ARMV5 UREL shows correct values using elf2e32 245713 5 60293 laszlo.gombos 2010-07-01 19:55:26 -0700 Comment on attachment 60293 Make max heap variable for ARM/emulator. Also use paging Looks good - 2 minor issues: 1./ Can you make a reference to the bugzilla bug in the ChangeLog (prepare-ChangeLog --bug <bugid>will do that for you)? 2./ Patch does not seems to apply on the bot and has tab character in it 245786 6 koshuin 2010-07-02 00:17:23 -0700 Hi, PAGED keyword is deprecated and should not be used with SBSv2. Instead you should be using codepaging and datapaging keywords. These keywords also come from Qt and are added to all Qt DLLs so you don't need to add them separately. 16MB for emulator? please, that will not run anything. If you are having troubles getting more for emulator you need to tweak your environment. max heapsize defaults to 20MB in Qt which is way too low for webkit apps. I've used 48MB in my test apps as that is the roughly the limit of my test phone and with that amount never had issues of emulator not starting but have had issues of running out of max app heap. I'd set it to 64MB and learn how to get more out of your environment. My machine has 3GB of RAM so you might have issues if you have less than that. Closing other programs esp. webbrowser before running emulator helps on releasing commited memory. 245793 7 koshuin 2010-07-02 00:39:22 -0700 This actually raises a question. Have you modified epoc.ini ? I have set MegabytesOfFreeMemory 256 and test apps as 48MB... You do realise if you set a separate limit to emulator and target you will get a very different behaviour which leads to hundreds of 'bugs' in the emulator env which can't be reproduced on HW. Or worse crashes in the emulator might hide the real bugs that only appear once we have enough memory and that is in the HW. Users will not understand this and we end up answering a lot of forum questions and spending lot of time on debugging bogus reports. Ofc. this is issue of the emulator which we can't do much about except drop the support for good. From my perspective the app not starting at all is much easier to explain than myriad crashes. 245918 8 60369 s.mathur 2010-07-02 08:32:30 -0700 Created attachment 60369 increased emulator limit, and explicitly used pageddata and pagedcode directives 251912 9 s.mathur 2010-07-15 08:49:44 -0700 (In reply to comment #8) > Created an attachment (id=60369) [details] > increased emulator limit, and explicitly used pageddata and pagedcode directives Janne, Did I address your concerns the other day over IRC? I upped the emulator limit to the higher value we discussed. Hopefully, "normal" Qt users as well as QtProd users will get more out-of-the-box usable values with this patch. Also, the engineers trying to get DRT to work on N8 will also see fewer crashes because the DRT binary now gets data paging by default. 254765 10 hausmann 2010-07-22 07:01:35 -0700 (In reply to comment #6) > PAGED keyword is deprecated and should not be used with SBSv2. Instead you should be using codepaging and datapaging keywords. These keywords also come from Qt and are added to all Qt DLLs so you don't need to add them separately. Mahesh, what about Janne's comment here? 254769 11 s.mathur 2010-07-22 07:11:17 -0700 (In reply to comment #10) > (In reply to comment #6) > > PAGED keyword is deprecated and should not be used with SBSv2. Instead you should be using codepaging and datapaging keywords. These keywords also come from Qt and are added to all Qt DLLs so you don't need to add them separately. > > Mahesh, what about Janne's comment here? My patch uses pageddata and pagedcode keywords, which are the more fine grained way of achieving the same effect. Janne was OK with this in my IRC chat with him, but he is on vacation it seems. Note that if your compiling QtWebkit/Symbian on an older toolchain say ABLD on 3.2 or 5.0, the tools will ignore the pageddata keyword with a warning printed about "unidentified keyword" . That's OK. Hopefully, with this patch DRT can run better on Symbian^3 due to use of data paging. 259086 12 s.mathur 2010-08-02 13:51:05 -0700 Removing paging part from bug description, as we will let Qt-Symbian's configuration handle that part. 259116 13 63260 s.mathur 2010-08-02 14:22:41 -0700 Created attachment 63260 updated patch with paging config removed 259137 14 63264 s.mathur 2010-08-02 15:01:14 -0700 Created attachment 63264 Correct screw up in previous patch where older TARGET.EPOCHEAPSIZE wasn't deleted Verified that the following shows up OK in \WebKitTools\QtTestBrowser\QtTestBrowser_0xA000E543.mmp and WebKitTools\DumpRenderTree\qt\DumpRenderTree_0xE5192995.mmp (for example). #ifdef WINSCW EPOCHEAPSIZE 0x40000 0x2000000 // Min 256kB, Max 32MB #else EPOCHEAPSIZE 0x40000 0x6000000 // Min 256kB, Max 96MB #endif 259239 15 63264 laszlo.gombos 2010-08-02 18:11:37 -0700 Comment on attachment 63264 Correct screw up in previous patch where older TARGET.EPOCHEAPSIZE wasn't deleted Looks good to me now that the explicit paging rules are removed. r+. Thanks. 260106 16 63264 commit-queue 2010-08-04 11:09:16 -0700 Comment on attachment 63264 Correct screw up in previous patch where older TARGET.EPOCHEAPSIZE wasn't deleted Clearing flags on attachment: 63264 Committed r64657: <http://trac.webkit.org/changeset/64657> 260107 17 commit-queue 2010-08-04 11:09:21 -0700 All reviewed patches have been landed. Closing bug. 266677 18 laszlo.gombos 2010-08-19 05:48:34 -0700 This patch introduced a new build warning we should look into getting rid of it.. Warning: Restricted statements detected in MMP_RULES: (EPOCHEAPSIZE) Use corresponding qmake variable(s) instead. 266712 19 s.mathur 2010-08-19 07:14:27 -0700 (In reply to comment #18) > This patch introduced a new build warning we should look into getting rid of it.. > > Warning: Restricted statements detected in MMP_RULES: > (EPOCHEAPSIZE) > Use corresponding qmake variable(s) instead. That warning is qmake's way of telling us that a normal user shouldn't have a need to use anything except TARGET.EPOCHEAPSIZE directive. Since we can't achieve conditional heap sizes with simply TARGET.EPOCHEAPSIZE, this warning is OK to incur. If you can verify that your generated webcore_<UID3>.mmp has the correct syntax, then things are OK. 296705 20 hausmann 2010-10-20 02:17:43 -0700 Revision r64657 cherry-picked into qtwebkit-2.1 with commit 80daa72 <http://gitorious.org/webkit/qtwebkit/commit/80daa72> 296707 21 hausmann 2010-10-20 02:18:31 -0700 re-opening as it was before the cherry-pick 296787 22 s.mathur 2010-10-20 07:31:23 -0700 I don't think any action is required on this. Hence resolving. 60293 2010-07-01 15:02:20 -0700 2010-07-02 08:32:30 -0700 Make max heap variable for ARM/emulator. Also use paging try1.patch text/plain 1761 s.mathur ZGlmZiAtLWdpdCBhL0NoYW5nZUxvZyBiL0NoYW5nZUxvZwppbmRleCA2MTFlZmY3Li5lYTA2NTdk IDEwMDY0NAotLS0gYS9DaGFuZ2VMb2cKKysrIGIvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTEgQEAK KzIwMTAtMDctMDEgIFNpZGRoYXJ0aCBNYXRodXIgIDxzaWRkaGFydGgubWF0aHVyQG5va2lhLmNv bT4KKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBbUXRd W1N5bWJpYW5dIFZhcmlhYmxlIG1heCBoZWFwIHNpemUgYmV0d2VlbiB0YXJnZXQvZW11bGF0b3Is IGFuZCBwYWdpbmcgYXMgZGVmYXVsdAorCisgICAgICAgICogV2ViS2l0LnByaToKKwogMjAxMC0w Ny0wMSAgU2ltb24gSGF1c21hbm4gIDxzaW1vbi5oYXVzbWFubkBub2tpYS5jb20+CiAKICAgICAg ICAgUnViYmVyLXN0YW1wZWQgYnkgTGFzemxvIEdvbWJvcy4KZGlmZiAtLWdpdCBhL1dlYktpdC5w cmkgYi9XZWJLaXQucHJpCmluZGV4IDA3YmMxODMuLjA1ZDA3YjQgMTAwNjQ0Ci0tLSBhL1dlYktp dC5wcmkKKysrIGIvV2ViS2l0LnByaQpAQCAtMzAsNyArMzAsMjAgQEAgYnVpbGRpbmctbGlicyB7 CiAgICAgICAgICAgICAgICAgTElCUyArPSAtbFF0V2ViS2l0CiAgICAgICAgICAgICAgICAgc3lt YmlhbiB7CiAgICAgICAgICAgICAgICAgICAgIFRBUkdFVC5FUE9DU1RBQ0tTSVpFID0gMHgxNDAw MCAvLyA4MCBrQgotICAgICAgICAgICAgICAgICAgICBUQVJHRVQuRVBPQ0hFQVBTSVpFID0gMHgy MDAwMCAweDYwMDAwMDAgLy8gTWluIDEyOGtCLCBNYXggMzJNQgorICAgICAgICAgICAgICAgICAg ICAjIEhlYXAgY29uZmlndXJhdGlvbiBmb3IgRVhFIHRhcmdldHM6IG1ha2UgbGFyZ2UgZW5vdWdo IHRvIGJlIHVzYWJsZQorICAgICAgICAgICAgICAgICAgICBUQVJHRVQuRVBPQ0hFQVBTSVpFID0g CisgICAgICAgICAgICAgICAgICAgIGhlYXBTaXplUnVsZSA9IFwKKyAgICAgICAgICAgICAgICAg ICAgIiQke0xJVEVSQUxfSEFTSH1pZmRlZiBXSU5TQ1ciIFwKKyAgICAgICAgICAgICAgICAgICAg ICAgICJFUE9DSEVBUFNJWkUgIDB4NDAwMDAgMHgxMDAwMDAwIC8vIE1pbiAyNTZrQiwgTWF4IDE2 TUIiIFwKKyAgICAgICAgICAgICAgICAgICAgIiQke0xJVEVSQUxfSEFTSH1lbHNlIiBcCisgICAg ICAgICAgICAgICAgICAgICAgICAiRVBPQ0hFQVBTSVpFICAweDQwMDAwIDB4NjAwMDAwMCAvLyBN aW4gMjU2a0IsIE1heCA5Nk1CIiBcCisgICAgICAgIAkgICAgIiQke0xJVEVSQUxfSEFTSH1lbmRp ZiIKKyAgICAgICAgCSAgICBNTVBfUlVMRVMgKz0gaGVhcFNpemVSdWxlCisKKyAgICAgICAgICAg ICAgICAgICAgIyBQYWdpbmcgZm9yIGJvdGggRExMIGFuZCBFWEUgdGFyZ2V0czogZW5hYmxlIGNv ZGUgcGFnaW5nIChPRFApIGFuZCBkYXRhIHBhZ2luZyAoV0RQKQorICAgICAgICAgICAgICAgICAg ICAjIGh0dHA6Ly9kZXZlbG9wZXIuc3ltYmlhbi5vcmcvbWFpbi9kb2N1bWVudGF0aW9uL3JlZmVy ZW5jZS9zMy9wZGsvR1VJRC02RUY3NjJCOC1ERTkzLTUxQzAtOEE1RC04OUZDNTI4OUI3RDAuaHRt bAorICAgICAgICAgICAgICAgICAgICAjIEltcG9ydGFudDogeW91IG5lZWQgYSBuZXdpc2ggdG9v bGNoYWluIEFORCBhIFN5bWJpYW5eMyBvciBsYXRlciBkZXZpY2UgZm9yIHlvdXIgYXBwIHRvIGRh dGEgcGFnZSBzdWNjZXNzZnVsbHkKKyAgICAgICAgICAgICAgICAgICAgTU1QX1JVTEVTICo9IHBh Z2VkCQkKICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICB9CiAgICAgICAgIH0K 60369 2010-07-02 08:32:30 -0700 2010-08-02 14:22:41 -0700 increased emulator limit, and explicitly used pageddata and pagedcode directives try3.patch text/plain 1887 s.mathur ZGlmZiAtLWdpdCBhL0NoYW5nZUxvZyBiL0NoYW5nZUxvZwppbmRleCA2MTFlZmY3Li5kNzlkMDQw IDEwMDY0NAotLS0gYS9DaGFuZ2VMb2cKKysrIGIvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTIgQEAK KzIwMTAtMDctMDIgIFNpZGRoYXJ0aCBNYXRodXIgIDxzaWRkaGFydGgubWF0aHVyQG5va2lhLmNv bT4KKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBWYXJp YWJsZSBtYXggaGVhcCBzaXplIGJldHdlZW4gdGFyZ2V0L2VtdWxhdG9yLCBhbmQgcGFnaW5nIGFz IGRlZmF1bHQKKyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lk PTQxNDgwCisKKyAgICAgICAgKiBXZWJLaXQucHJpOgorCiAyMDEwLTA3LTAxICBTaW1vbiBIYXVz bWFubiAgPHNpbW9uLmhhdXNtYW5uQG5va2lhLmNvbT4KIAogICAgICAgICBSdWJiZXItc3RhbXBl ZCBieSBMYXN6bG8gR29tYm9zLgpkaWZmIC0tZ2l0IGEvV2ViS2l0LnByaSBiL1dlYktpdC5wcmkK aW5kZXggMDdiYzE4My4uZGQwZGFhOCAxMDA2NDQKLS0tIGEvV2ViS2l0LnByaQorKysgYi9XZWJL aXQucHJpCkBAIC0zMCw3ICszMCwyMyBAQCBidWlsZGluZy1saWJzIHsKICAgICAgICAgICAgICAg ICBMSUJTICs9IC1sUXRXZWJLaXQKICAgICAgICAgICAgICAgICBzeW1iaWFuIHsKICAgICAgICAg ICAgICAgICAgICAgVEFSR0VULkVQT0NTVEFDS1NJWkUgPSAweDE0MDAwIC8vIDgwIGtCCi0gICAg ICAgICAgICAgICAgICAgIFRBUkdFVC5FUE9DSEVBUFNJWkUgPSAweDIwMDAwIDB4NjAwMDAwMCAv LyBNaW4gMTI4a0IsIE1heCAzMk1CCisgICAgICAgICAgICAgICAgICAgICMgRm9yIEVYRXMgb25s eTogc2V0IGhlYXAgdG8gdXNhYmxlIHZhbHVlCisgICAgICAgICAgICAgICAgICAgIFRBUkdFVC5F UE9DSEVBUFNJWkUgPSAKKyAgICAgICAgICAgICAgICAgICAgaGVhcFNpemVSdWxlID0gXAorICAg ICAgICAgICAgICAgICAgICAiJCR7TElURVJBTF9IQVNIfWlmZGVmIFdJTlNDVyIgXAorICAgICAg ICAgICAgICAgICAgICAgICAgIkVQT0NIRUFQU0laRSAgMHg0MDAwMCAweDIwMDAwMDAgLy8gTWlu IDI1NmtCLCBNYXggMzJNQiIgXAorICAgICAgICAgICAgICAgICAgICAiJCR7TElURVJBTF9IQVNI fWVsc2UiIFwKKyAgICAgICAgICAgICAgICAgICAgICAgICJFUE9DSEVBUFNJWkUgIDB4NDAwMDAg MHg2MDAwMDAwIC8vIE1pbiAyNTZrQiwgTWF4IDk2TUIiIFwKKyAgICAgICAgICAgICAgICAgICAg IiQke0xJVEVSQUxfSEFTSH1lbmRpZiIKKyAgICAgICAgICAgICAgICAgICAgTU1QX1JVTEVTICs9 IGhlYXBTaXplUnVsZQorCisgICAgICAgICAgICAgICAgICAgICMgRm9yIEVYRXMgb25seTogYWxs b3cgZGF0YSBwYWdpbmcgb2YgdGhlIHByb2Nlc3MgaGVhcCAoV0RQKSAKKyAgICAgICAgICAgICAg ICAgICAgIyBodHRwOi8vZGV2ZWxvcGVyLnN5bWJpYW4ub3JnL21haW4vZG9jdW1lbnRhdGlvbi9y ZWZlcmVuY2UvczMvcGRrL0dVSUQtNkVGNzYyQjgtREU5My01MUMwLThBNUQtODlGQzUyODlCN0Qw Lmh0bWwKKyAgICAgICAgICAgICAgICAgICAgIyBJbXBvcnRhbnQ6IHlvdSBuZWVkIGEgbmV3aXNo IHRvb2xjaGFpbiBBTkQgYSBTeW1iaWFuXjMgb3IgbGF0ZXIgZGV2aWNlIGZvciB5b3VyIGFwcCB0 byBkYXRhIHBhZ2Ugc3VjY2Vzc2Z1bGx5CisgICAgICAgICAgICAgICAgICAgIE1NUF9SVUxFUyAq PSBwYWdlZGRhdGEJCQorCisgICAgICAgICAgICAgICAgICAgICMgRm9yIGJvdGggRExMcyBhbmQg RVhFczogYWxsb3cgY29kZSBwYWdpbmcgKE9EUCkKKyAgICAgICAgICAgICAgICAgICAgTU1QX1JV TEVTICo9IHBhZ2VkY29kZQkKICAgICAgICAgICAgICAgICB9CiAgICAgICAgICAgICB9CiAgICAg ICAgIH0K 63260 2010-08-02 14:22:41 -0700 2010-08-02 15:01:14 -0700 updated patch with paging config removed 41480_v4.patch text/plain 1393 s.mathur ZGlmZiAtLWdpdCBhL0NoYW5nZUxvZyBiL0NoYW5nZUxvZwppbmRleCAwYmY0ZWQ2Li4yM2FjY2Jk IDEwMDY0NAotLS0gYS9DaGFuZ2VMb2cKKysrIGIvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTIgQEAK KzIwMTAtMDgtMDIgIFNpZGRoYXJ0aCBNYXRodXIgIDxzaWRkaGFydGgubWF0aHVyQG5va2lhLmNv bT4KKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBbUXRd W1N5bWJpYW5dIFZhcmlhYmxlIG1heCBoZWFwIHNpemUgYmV0d2VlbiB0YXJnZXQvZW11bGF0b3IK KyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTQxNDgwCisK KyAgICAgICAgKiBXZWJLaXQucHJpOiBTeW1iaWFuLW9ubHkgY29kZSBibG9jayBmb3IgRVBPQ0hF QVBTSVpFIGNvbmZpZ3VyYXRpb24KKwogMjAxMC0wOC0wMiAgQ2hyaXMgRmxlaXphY2ggIDxjZmxl aXphY2hAYXBwbGUuY29tPgogCiAgICAgICAgIFJldmlld2VkIGJ5IERhdmlkIEtpbHplci4KZGlm ZiAtLWdpdCBhL1dlYktpdC5wcmkgYi9XZWJLaXQucHJpCmluZGV4IGMwNWMyYTIuLmM4ZWRhYWQg MTAwNjQ0Ci0tLSBhL1dlYktpdC5wcmkKKysrIGIvV2ViS2l0LnByaQpAQCAtMzAsNiArMzAsMTUg QEAgYnVpbGRpbmctbGlicyB7CiAgICAgICAgICAgICAgICAgTElCUyArPSAtbFF0V2ViS2l0CiAg ICAgICAgICAgICAgICAgc3ltYmlhbiB7CiAgICAgICAgICAgICAgICAgICAgIFRBUkdFVC5FUE9D U1RBQ0tTSVpFID0gMHgxNDAwMCAvLyA4MCBrQgorICAgICAgICAgICAgICAgICAgICAjIEZvciBF WEVzIG9ubHk6IHNldCBoZWFwIHRvIHVzYWJsZSB2YWx1ZQorICAgICAgICAgICAgICAgICAgICBU QVJHRVQuRVBPQ0hFQVBTSVpFID0gCisgICAgICAgICAgICAgICAgICAgIGhlYXBTaXplUnVsZSA9 IFwKKyAgICAgICAgICAgICAgICAgICAgIiQke0xJVEVSQUxfSEFTSH1pZmRlZiBXSU5TQ1ciIFwK KyAgICAgICAgICAgICAgICAgICAgICAgICJFUE9DSEVBUFNJWkUgIDB4NDAwMDAgMHgyMDAwMDAw IC8vIE1pbiAyNTZrQiwgTWF4IDMyTUIiIFwKKyAgICAgICAgICAgICAgICAgICAgIiQke0xJVEVS QUxfSEFTSH1lbHNlIiBcCisgICAgICAgICAgICAgICAgICAgICAgICAiRVBPQ0hFQVBTSVpFICAw eDQwMDAwIDB4NjAwMDAwMCAvLyBNaW4gMjU2a0IsIE1heCA5Nk1CIiBcCisgICAgICAgICAgICAg ICAgICAgICIkJHtMSVRFUkFMX0hBU0h9ZW5kaWYiCisgICAgICAgICAgICAgICAgICAgIE1NUF9S VUxFUyArPSBoZWFwU2l6ZVJ1bGUKICAgICAgICAgICAgICAgICAgICAgVEFSR0VULkVQT0NIRUFQ U0laRSA9IDB4MjAwMDAgMHg2MDAwMDAwIC8vIE1pbiAxMjhrQiwgTWF4IDk2TUIKICAgICAgICAg ICAgICAgICB9CiAgICAgICAgICAgICB9Cg== 63264 2010-08-02 15:01:14 -0700 2010-08-04 11:09:15 -0700 Correct screw up in previous patch where older TARGET.EPOCHEAPSIZE wasn't deleted 41480_v5.patch text/plain 1404 s.mathur ZGlmZiAtLWdpdCBhL0NoYW5nZUxvZyBiL0NoYW5nZUxvZwppbmRleCAwYmY0ZWQ2Li4yM2FjY2Jk IDEwMDY0NAotLS0gYS9DaGFuZ2VMb2cKKysrIGIvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTIgQEAK KzIwMTAtMDgtMDIgIFNpZGRoYXJ0aCBNYXRodXIgIDxzaWRkaGFydGgubWF0aHVyQG5va2lhLmNv bT4KKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBbUXRd W1N5bWJpYW5dIFZhcmlhYmxlIG1heCBoZWFwIHNpemUgYmV0d2VlbiB0YXJnZXQvZW11bGF0b3IK KyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTQxNDgwCisK KyAgICAgICAgKiBXZWJLaXQucHJpOiBTeW1iaWFuLW9ubHkgY29kZSBibG9jayBmb3IgRVBPQ0hF QVBTSVpFIGNvbmZpZ3VyYXRpb24KKwogMjAxMC0wOC0wMiAgQ2hyaXMgRmxlaXphY2ggIDxjZmxl aXphY2hAYXBwbGUuY29tPgogCiAgICAgICAgIFJldmlld2VkIGJ5IERhdmlkIEtpbHplci4KZGlm ZiAtLWdpdCBhL1dlYktpdC5wcmkgYi9XZWJLaXQucHJpCmluZGV4IGMwNWMyYTIuLjIxNDA0ZGQg MTAwNjQ0Ci0tLSBhL1dlYktpdC5wcmkKKysrIGIvV2ViS2l0LnByaQpAQCAtMzAsNyArMzAsMTUg QEAgYnVpbGRpbmctbGlicyB7CiAgICAgICAgICAgICAgICAgTElCUyArPSAtbFF0V2ViS2l0CiAg ICAgICAgICAgICAgICAgc3ltYmlhbiB7CiAgICAgICAgICAgICAgICAgICAgIFRBUkdFVC5FUE9D U1RBQ0tTSVpFID0gMHgxNDAwMCAvLyA4MCBrQgotICAgICAgICAgICAgICAgICAgICBUQVJHRVQu RVBPQ0hFQVBTSVpFID0gMHgyMDAwMCAweDYwMDAwMDAgLy8gTWluIDEyOGtCLCBNYXggOTZNQgor ICAgICAgICAgICAgICAgICAgICAjIEZvciBFWEVzIG9ubHk6IHNldCBoZWFwIHRvIHVzYWJsZSB2 YWx1ZQorICAgICAgICAgICAgICAgICAgICBUQVJHRVQuRVBPQ0hFQVBTSVpFID0gCisgICAgICAg ICAgICAgICAgICAgIGhlYXBTaXplUnVsZSA9IFwKKyAgICAgICAgICAgICAgICAgICAgIiQke0xJ VEVSQUxfSEFTSH1pZmRlZiBXSU5TQ1ciIFwKKyAgICAgICAgICAgICAgICAgICAgICAgICJFUE9D SEVBUFNJWkUgIDB4NDAwMDAgMHgyMDAwMDAwIC8vIE1pbiAyNTZrQiwgTWF4IDMyTUIiIFwKKyAg ICAgICAgICAgICAgICAgICAgIiQke0xJVEVSQUxfSEFTSH1lbHNlIiBcCisgICAgICAgICAgICAg ICAgICAgICAgICAiRVBPQ0hFQVBTSVpFICAweDQwMDAwIDB4NjAwMDAwMCAvLyBNaW4gMjU2a0Is IE1heCA5Nk1CIiBcCisgICAgICAgICAgICAgICAgICAgICIkJHtMSVRFUkFMX0hBU0h9ZW5kaWYi CisgICAgICAgICAgICAgICAgICAgIE1NUF9SVUxFUyArPSBoZWFwU2l6ZVJ1bGUKICAgICAgICAg ICAgICAgICB9CiAgICAgICAgICAgICB9CiAgICAgICAgIH0K