73847 2011-12-05 11:10:31 -0800 NRWT creates too many workers on RAM-limited machines 2022-03-01 03:08:37 -0800 1 1 1 Unclassified WebKit Tools / Tests 528+ (Nightly build) All All RESOLVED FIXED P2 Normal --- 77419 1 aroben eric abarth dpranke eric jamesr ojan rniwa tony webkit.review.bot oldest_to_newest 515222 0 aroben 2011-12-05 11:10:31 -0800 NRWT creates the same number of workers regardless of how much RAM is installed. On RAM-limited machines (e.g., 12 cores, 3GB RAM), so many workers are created that the machine starts paging like crazy and all tests end up timing out. It would be nice if NRWT were to take RAM into account when deciding how many workers to create. 515231 1 eric 2011-12-05 11:17:44 -0800 http://support.hyperic.com/display/SIGAR/Home is a way to do this. But I don't think we can auto-install that, due to the binary files it includes. I suspect we'll have to roll our own if we want this functionality. Juse like we do for Executive.cpu_count() 515232 2 eric 2011-12-05 11:18:23 -0800 How much ram should NRWT expect a DRT instance to need? 500mb? 515236 3 tony 2011-12-05 11:23:16 -0800 (In reply to comment #2) > How much ram should NRWT expect a DRT instance to need? 500mb? 500mb sounds plenty conservative. Make sure to leave ram for the system and NRWT. 515278 4 dpranke 2011-12-05 12:20:51 -0800 It hadn't occurred to me before that we might have a lot of many-core but low-RAM machines. Makes sense to change this. 515429 5 117944 eric 2011-12-05 15:04:27 -0800 Created attachment 117944 Patch 515445 6 117944 dpranke 2011-12-05 15:22:52 -0800 Comment on attachment 117944 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=117944&action=review > Tools/Scripts/run-webkit-tests:79 > + return !isQt(); Are we sure that it was a RAM issue and not something else? Also, in the future I'd probably change the defaults in port/mac.py et al. rather than duplicating the logic here in Perl. > Tools/Scripts/webkitpy/layout_tests/port/base.py:177 > + return cpu_count I'm not sure zero should be a legal value here, so maybe you should return at least 1? If it is zero (essentially saying we need at least 300MB to run), there should probably at least be a check in run_webkit_tests: _set_up_derived_options() to handle zero cleanly and bail out. 515455 7 eric 2011-12-05 15:37:54 -0800 (In reply to comment #6) > (From update of attachment 117944 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=117944&action=review > > > Tools/Scripts/run-webkit-tests:79 > > + return !isQt(); > > Are we sure that it was a RAM issue and not something else? I don't know. But it doesn't hurt to try. > Also, in the future I'd probably change the defaults in port/mac.py et al. rather than duplicating the logic here in Perl. This is intentional. run-webkit-tests is meant to match ORWT, to be a shim. Folks who run new-run-webkit-test directly always get parallel testing, regardless of the state of the bots. > > Tools/Scripts/webkitpy/layout_tests/port/base.py:177 > > + return cpu_count > > I'm not sure zero should be a legal value here, so maybe you should return at least 1? If it is zero (essentially saying we need at least 300MB to run), there should probably at least be a check in run_webkit_tests: _set_up_derived_options() to handle zero cleanly and bail out. This is no different from setting --child-count=0 directly. I think that's out of scope of this change. 515462 8 117944 dpranke 2011-12-05 15:45:06 -0800 Comment on attachment 117944 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=117944&action=review >>> Tools/Scripts/webkitpy/layout_tests/port/base.py:177 >>> + return cpu_count >> >> I'm not sure zero should be a legal value here, so maybe you should return at least 1? If it is zero (essentially saying we need at least 300MB to run), there should probably at least be a check in run_webkit_tests: _set_up_derived_options() to handle zero cleanly and bail out. > > This is no different from setting --child-count=0 directly. I think that's out of scope of this change. I disagree, in the sense that there was no way default_child_processes() could return 0 before, and now it can, potentially breaking things down the road. That said, it'll probably never come up in practice. Up to you ... 515465 9 eric 2011-12-05 15:48:59 -0800 Committed r102056: <http://trac.webkit.org/changeset/102056> 117944 2011-12-05 15:04:27 -0800 2022-03-01 03:08:37 -0800 Patch bug-73847-20111205150426.patch text/plain 8986 eric U3VidmVyc2lvbiBSZXZpc2lvbjogMTAyMDQ2CmRpZmYgLS1naXQgYS9Ub29scy9DaGFuZ2VMb2cg Yi9Ub29scy9DaGFuZ2VMb2cKaW5kZXggOGI4ZGIyZDBkMGNiYWQ0YTU1ZWNiOTlmOGVkNjg4M2Vk MDk1NDJiYy4uNjEwMjRlNWFkNzA0MGIyZGI4NGQ1ZDdjYzg2NTg4N2JiOWE3ZmQ5YyAxMDA2NDQK LS0tIGEvVG9vbHMvQ2hhbmdlTG9nCisrKyBiL1Rvb2xzL0NoYW5nZUxvZwpAQCAtMSwzICsxLDMy IEBACisyMDExLTEyLTA1ICBFcmljIFNlaWRlbCAgPGVyaWNAd2Via2l0Lm9yZz4KKworICAgICAg ICBOUldUIGNyZWF0ZXMgdG9vIG1hbnkgd29ya2VycyBvbiBSQU0tbGltaXRlZCBtYWNoaW5lcwor ICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9NzM4NDcKKwor ICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBNYWtlIE5SV1Qg cmVzdHJpY3QgaXRzIG51bWJlciBvZiB3b3JrZXJzIHdoZW4gdGhlcmUgaXNuJ3QgZW5vdWdoCisg ICAgICAgIFJBTSB0byBzdXBwb3J0IHRoZW0uCisgICAgICAgIEFsc28gcmUtZW5hYmxlIHBhcmFs bGVsIHRlc3Rpbmcgb24gdGhlIE1hYyBib3RzIHdoaWNoIHdlcmUKKyAgICAgICAgZmFpbGluZyBk dWUgdG8gbm90IGVub3VnaCByYW0uCisKKyAgICAgICAgKiBTY3JpcHRzL3J1bi13ZWJraXQtdGVz dHM6CisgICAgICAgIChwbGF0Zm9ybUlzUmVhZHlGb3JQYXJhbGxlbFRlc3RpbmcpOgorICAgICAg ICAqIFNjcmlwdHMvd2Via2l0cHkvY29tbW9uL2hvc3QucHk6CisgICAgICAgIChIb3N0Ll9faW5p dF9fKToKKyAgICAgICAgKiBTY3JpcHRzL3dlYmtpdHB5L2NvbW1vbi9zeXN0ZW0vZXhlY3V0aXZl X21vY2sucHk6CisgICAgICAgIChNb2NrRXhlY3V0aXZlLmNwdV9jb3VudCk6CisgICAgICAgICog U2NyaXB0cy93ZWJraXRweS9jb21tb24vc3lzdGVtL3BsYXRmb3JtaW5mby5weToKKyAgICAgICAg KFBsYXRmb3JtSW5mby5fX2luaXRfXyk6CisgICAgICAgIChQbGF0Zm9ybUluZm8uZGlzcGxheV9u YW1lKToKKyAgICAgICAgKiBTY3JpcHRzL3dlYmtpdHB5L2NvbW1vbi9zeXN0ZW0vcGxhdGZvcm1p bmZvX21vY2sucHk6CisgICAgICAgIChNb2NrUGxhdGZvcm1JbmZvLmRpc3BsYXlfbmFtZSk6Cisg ICAgICAgICogU2NyaXB0cy93ZWJraXRweS9sYXlvdXRfdGVzdHMvcG9ydC9iYXNlLnB5OgorICAg ICAgICAoUG9ydC5kZWZhdWx0X2NoaWxkX3Byb2Nlc3Nlcyk6CisgICAgICAgICogU2NyaXB0cy93 ZWJraXRweS9sYXlvdXRfdGVzdHMvcG9ydC9iYXNlX3VuaXR0ZXN0LnB5OgorICAgICAgICAoUG9y dFRlc3QudGVzdF9kZWZhdWx0X2NoaWxkX3Byb2Nlc3Nlcyk6CisgICAgICAgIChQb3J0VGVzdC50 ZXN0X3ByZXR0eV9wYXRjaF9vc19lcnJvcik6CisKIDIwMTEtMTItMDUgIEFkYW0gQmFydGggIDxh YmFydGhAd2Via2l0Lm9yZz4KIAogICAgICAgICBbQ2hyb21pdW1dIFJlbW92ZSBuby1sb25nZXIt dXNlZCBmb3J3YXJkaW5nIGhlYWRlcnMgaW4gdGhlIFdlYktpdCBBUEkKZGlmZiAtLWdpdCBhL1Rv b2xzL1NjcmlwdHMvcnVuLXdlYmtpdC10ZXN0cyBiL1Rvb2xzL1NjcmlwdHMvcnVuLXdlYmtpdC10 ZXN0cwppbmRleCBhMDNmYzE3MjE1NWE1YmRiNjBlNWNmN2EyMTU2NWVmNDMyMjQ2ZDA3Li4wZTVm ZTFmNzFhYThlYTc4NDUxYTQ1OWFkM2E2MGY3OGZlZmQ5OGE5IDEwMDc1NQotLS0gYS9Ub29scy9T Y3JpcHRzL3J1bi13ZWJraXQtdGVzdHMKKysrIGIvVG9vbHMvU2NyaXB0cy9ydW4td2Via2l0LXRl c3RzCkBAIC03Niw3ICs3Niw3IEBAIHN1YiBwbGF0Zm9ybUlzUmVhZHlGb3JQYXJhbGxlbFRlc3Rp bmcoKQogICAgICMgV2UgYmVsaWV2ZSBhbGwgcGxhdGZvcm1zIGFyZSByZWFkeSBmb3IgZGVmYXVs dCBwYXJhbGxlbCB0ZXN0aW5nIGV4Y2VwdAogICAgICMgUXQsIGFzIFF0IHJ1bnMgbW9yZSB0aGFu IG9uZSBidWlsZC1zbGF2ZSBwZXItc2VydmVyLgogICAgICMgT3NzeSBoYXMgYXNrZWQgbWUgdG8g YmxhY2tsaXN0IFF0IGZvciBub3cuCi0gICAgcmV0dXJuICFpc1F0KCkgJiYgIWlzQXBwbGVNYWNX ZWJLaXQoKTsKKyAgICByZXR1cm4gIWlzUXQoKTsKIH0KIAogbXkgJGhhcm5lc3NOYW1lID0gIm9s ZC1ydW4td2Via2l0LXRlc3RzIjsKZGlmZiAtLWdpdCBhL1Rvb2xzL1NjcmlwdHMvd2Via2l0cHkv Y29tbW9uL2hvc3QucHkgYi9Ub29scy9TY3JpcHRzL3dlYmtpdHB5L2NvbW1vbi9ob3N0LnB5Cmlu ZGV4IDZkNDEwZjg0ODI1ODNjZDBiMTc5NmZhMjEyNmQ2YjY5ZmQ1Y2M4ZWMuLjZjY2RlNzNlYWM2 YmQyYzk3Zjk4ZDRmMjExMDFmMjQyZmY3ZDhhNjQgMTAwNjQ0Ci0tLSBhL1Rvb2xzL1NjcmlwdHMv d2Via2l0cHkvY29tbW9uL2hvc3QucHkKKysrIGIvVG9vbHMvU2NyaXB0cy93ZWJraXRweS9jb21t b24vaG9zdC5weQpAQCAtNTIsNyArNTIsNyBAQCBjbGFzcyBIb3N0KG9iamVjdCk6CiAgICAgICAg IHNlbGYuZXhlY3V0aXZlID0gZXhlY3V0aXZlLkV4ZWN1dGl2ZSgpCiAgICAgICAgIHNlbGYuZmls ZXN5c3RlbSA9IGZpbGVzeXN0ZW0uRmlsZVN5c3RlbSgpCiAgICAgICAgIHNlbGYudXNlciA9IHVz ZXIuVXNlcigpCi0gICAgICAgIHNlbGYucGxhdGZvcm0gPSBwbGF0Zm9ybWluZm8uUGxhdGZvcm1J bmZvKCkKKyAgICAgICAgc2VsZi5wbGF0Zm9ybSA9IHBsYXRmb3JtaW5mby5QbGF0Zm9ybUluZm8o c2VsZi5leGVjdXRpdmUpCiAgICAgICAgIHNlbGYud29ya3NwYWNlID0gd29ya3NwYWNlLldvcmtz cGFjZShzZWxmLmZpbGVzeXN0ZW0sIHNlbGYuZXhlY3V0aXZlKQogICAgICAgICBzZWxmLndlYiA9 IHdlYi5XZWIoKQogCmRpZmYgLS1naXQgYS9Ub29scy9TY3JpcHRzL3dlYmtpdHB5L2NvbW1vbi9z eXN0ZW0vZXhlY3V0aXZlX21vY2sucHkgYi9Ub29scy9TY3JpcHRzL3dlYmtpdHB5L2NvbW1vbi9z eXN0ZW0vZXhlY3V0aXZlX21vY2sucHkKaW5kZXggMGJjM2M1Njc3YzBhYjVmMWM5ZjQ2YWQ2ZDdm YjMwMjkxN2FlMDkyMC4uYzNhNGM0ZDVmYTg3N2ZlYWYyYmFiZGVmODUwYTJjMWVjZGE5MDgxNCAx MDA2NDQKLS0tIGEvVG9vbHMvU2NyaXB0cy93ZWJraXRweS9jb21tb24vc3lzdGVtL2V4ZWN1dGl2 ZV9tb2NrLnB5CisrKyBiL1Rvb2xzL1NjcmlwdHMvd2Via2l0cHkvY29tbW9uL3N5c3RlbS9leGVj dXRpdmVfbW9jay5weQpAQCAtNzQsNiArNzQsOSBAQCBjbGFzcyBNb2NrRXhlY3V0aXZlKG9iamVj dCk6CiAgICAgICAgICAgICByYWlzZSBTY3JpcHRFcnJvcigiTU9DSyBTY3JpcHRFcnJvciIsIG91 dHB1dD1vdXRwdXQpCiAgICAgICAgIHJldHVybiBvdXRwdXQKIAorICAgIGRlZiBjcHVfY291bnQo c2VsZik6CisgICAgICAgIHJldHVybiAyCisKIAogY2xhc3MgTW9ja0V4ZWN1dGl2ZTIob2JqZWN0 KToKICAgICBkZWYgX19pbml0X18oc2VsZiwgb3V0cHV0PScnLCBleGl0X2NvZGU9MCwgZXhjZXB0 aW9uPU5vbmUsCmRpZmYgLS1naXQgYS9Ub29scy9TY3JpcHRzL3dlYmtpdHB5L2NvbW1vbi9zeXN0 ZW0vcGxhdGZvcm1pbmZvLnB5IGIvVG9vbHMvU2NyaXB0cy93ZWJraXRweS9jb21tb24vc3lzdGVt L3BsYXRmb3JtaW5mby5weQppbmRleCBjYzM3MGJhZTFhZmE2OTgyYmI2MDFjMDU2ZTU5NWJiNDE4 NjdiNDZiLi4xMDBlMjY0MGE4OGU1ZjY1YmVkMTBkNTQ4N2ZiMDJmMzNkYzdmNjFlIDEwMDY0NAot LS0gYS9Ub29scy9TY3JpcHRzL3dlYmtpdHB5L2NvbW1vbi9zeXN0ZW0vcGxhdGZvcm1pbmZvLnB5 CisrKyBiL1Rvb2xzL1NjcmlwdHMvd2Via2l0cHkvY29tbW9uL3N5c3RlbS9wbGF0Zm9ybWluZm8u cHkKQEAgLTMxLDYgKzMxLDggQEAgaW1wb3J0IHBsYXRmb3JtCiAKICMgV2UgdXNlIHRoaXMgaW5z dGVhZCBvZiBjYWxscyB0byBwbGF0Zm9ybSBkaXJlY3RseSB0byBhbGxvdyBtb2NraW5nLgogY2xh c3MgUGxhdGZvcm1JbmZvKG9iamVjdCk6CisgICAgZGVmIF9faW5pdF9fKHNlbGYsIGV4ZWN1dGl2 ZSk6CisgICAgICAgIHNlbGYuX2V4ZWN1dGl2ZSA9IGV4ZWN1dGl2ZQogCiAgICAgZGVmIGRpc3Bs YXlfbmFtZShzZWxmKToKICAgICAgICAgIyBwbGF0Zm9ybS5wbGF0Zm9ybSgpIHJldHVybnMgRGFy d2luIGluZm9ybWF0aW9uIGZvciBNYWMsIHdoaWNoIGlzIGp1c3QgY29uZnVzaW5nLgpAQCAtNDEs MyArNDMsOSBAQCBjbGFzcyBQbGF0Zm9ybUluZm8ob2JqZWN0KToKICAgICAgICAgIyBMaW51eC0y LjYuMTgtMTk0LjMuMS5lbDUtaTY4Ni13aXRoLXJlZGhhdC01LjUtRmluYWwKICAgICAgICAgIyBX aW5kb3dzLTIwMDhTZXJ2ZXJSMi02LjEuNzYwMAogICAgICAgICByZXR1cm4gcGxhdGZvcm0ucGxh dGZvcm0oKQorCisgICAgZGVmIHRvdGFsX2J5dGVzX21lbW9yeShzZWxmKToKKyAgICAgICAgc3lz dGVtX25hbWUgPSBwbGF0Zm9ybS5zeXN0ZW0oKQorICAgICAgICBpZiBzeXN0ZW1fbmFtZSA9PSAi RGFyd2luIjoKKyAgICAgICAgICAgIHJldHVybiBpbnQoc2VsZi5fZXhlY3V0aXZlLnJ1bl9jb21t YW5kKFsic3lzY3RsIiwgIi1uIiwgImh3Lm1lbXNpemUiXSkpCisgICAgICAgIHJldHVybiBOb25l CmRpZmYgLS1naXQgYS9Ub29scy9TY3JpcHRzL3dlYmtpdHB5L2NvbW1vbi9zeXN0ZW0vcGxhdGZv cm1pbmZvX21vY2sucHkgYi9Ub29scy9TY3JpcHRzL3dlYmtpdHB5L2NvbW1vbi9zeXN0ZW0vcGxh dGZvcm1pbmZvX21vY2sucHkKaW5kZXggNWJkNzRlZDI5YzBkZWU5YjVlMmI5MTZiNmUxODg4MDI1 ZDc5ZjYzZS4uZDcxMzgxMWZjNzM0MjM0MjQ4NjRmZmYxZTk2NGNmMzViOTExZmY4OCAxMDA2NDQK LS0tIGEvVG9vbHMvU2NyaXB0cy93ZWJraXRweS9jb21tb24vc3lzdGVtL3BsYXRmb3JtaW5mb19t b2NrLnB5CisrKyBiL1Rvb2xzL1NjcmlwdHMvd2Via2l0cHkvY29tbW9uL3N5c3RlbS9wbGF0Zm9y bWluZm9fbW9jay5weQpAQCAtMzAsMyArMzAsNiBAQAogY2xhc3MgTW9ja1BsYXRmb3JtSW5mbyhv YmplY3QpOgogICAgIGRlZiBkaXNwbGF5X25hbWUoc2VsZik6CiAgICAgICAgIHJldHVybiAiTW9j a1BsYXRmb3JtIDEuMCIKKworICAgIGRlZiB0b3RhbF9ieXRlc19tZW1vcnkoc2VsZik6CisgICAg ICAgIHJldHVybiAyICogMTAyNCAqIDEwMjQgKiAxMDI0ICAjIDJHQiBpcyBhIHJlYXNvbmFibGUg YW1vdW50IG9mIHJhbSB0byBtb2NrLgpkaWZmIC0tZ2l0IGEvVG9vbHMvU2NyaXB0cy93ZWJraXRw eS9sYXlvdXRfdGVzdHMvcG9ydC9iYXNlLnB5IGIvVG9vbHMvU2NyaXB0cy93ZWJraXRweS9sYXlv dXRfdGVzdHMvcG9ydC9iYXNlLnB5CmluZGV4IDA1NDFhNGZlNGVkMDEwNjc2NmVmNWExN2JiZmYz MDk0YmY1NDhmY2EuLmFlMjMyOTUxMTNkM2VjYTc2MWVkZWI3MTA2M2JkOGI3MmJkM2U3NTMgMTAw NzU1Ci0tLSBhL1Rvb2xzL1NjcmlwdHMvd2Via2l0cHkvbGF5b3V0X3Rlc3RzL3BvcnQvYmFzZS5w eQorKysgYi9Ub29scy9TY3JpcHRzL3dlYmtpdHB5L2xheW91dF90ZXN0cy9wb3J0L2Jhc2UucHkK QEAgLTE2NCw3ICsxNjQsMTcgQEAgY2xhc3MgUG9ydChvYmplY3QpOgogCiAgICAgZGVmIGRlZmF1 bHRfY2hpbGRfcHJvY2Vzc2VzKHNlbGYpOgogICAgICAgICAiIiJSZXR1cm4gdGhlIG51bWJlciBv ZiBEdW1wUmVuZGVyVHJlZSBpbnN0YW5jZXMgdG8gdXNlIGZvciB0aGlzIHBvcnQuIiIiCi0gICAg ICAgIHJldHVybiBzZWxmLl9leGVjdXRpdmUuY3B1X2NvdW50KCkKKyAgICAgICAgY3B1X2NvdW50 ID0gc2VsZi5fZXhlY3V0aXZlLmNwdV9jb3VudCgpCisgICAgICAgICMgTWFrZSBzdXJlIHdlIGhh dmUgZW5vdWdoIHJhbSB0byBzdXBwb3J0IHRoYXQgbWFueSBpbnN0YW5jZXM6CisgICAgICAgIHRv dGFsX21lbW9yeSA9IHNlbGYuaG9zdC5wbGF0Zm9ybS50b3RhbF9ieXRlc19tZW1vcnkoKQorICAg ICAgICBpZiB0b3RhbF9tZW1vcnk6CisgICAgICAgICAgICBieXRlc19wZXJfZHJ0ID0gMzAwICog MTAyNCAqIDEwMjQgICMgQXNzdW1lIGVhY2ggRFJUIG5lZWRzIDMwME1CIHRvIHJ1bi4KKyAgICAg ICAgICAgIHN1cHBvcnRhYmxlX2luc3RhbmNlcyA9IHRvdGFsX21lbW9yeSAvIGJ5dGVzX3Blcl9k cnQKKyAgICAgICAgICAgIGlmIHN1cHBvcnRhYmxlX2luc3RhbmNlcyA8IGNwdV9jb3VudDoKKyAg ICAgICAgICAgICAgICAjIEZJWE1FOiBUaGUgUHJpbnRlciBpc24ndCBpbml0aWFsaXplZCB3aGVu IHRoaXMgaXMgY2FsbGVkLCBzbyB1c2luZyBfbG9nIHdvdWxkIGp1c3Qgc2hvdyBhbiB1bml0aWFs aXplZCBsb2dnZXIgZXJyb3IuCisgICAgICAgICAgICAgICAgcHJpbnQgIlRoaXMgbWFjaGluZSBj b3VsZCBzdXBwb3J0ICVzIGNoaWxkIHByb2Nlc3NlcywgYnV0IG9ubHkgaGFzIGVub3VnaCBtZW1v cnkgZm9yICVzLiIgJSAoY3B1X2NvdW50LCBzdXBwb3J0YWJsZV9pbnN0YW5jZXMpCisgICAgICAg ICAgICByZXR1cm4gbWluKHN1cHBvcnRhYmxlX2luc3RhbmNlcywgY3B1X2NvdW50KQorICAgICAg ICByZXR1cm4gY3B1X2NvdW50CiAKICAgICBkZWYgZGVmYXVsdF93b3JrZXJfbW9kZWwoc2VsZik6 CiAgICAgICAgIGlmIHNlbGYuX211bHRpcHJvY2Vzc2luZ19pc19hdmFpbGFibGU6CmRpZmYgLS1n aXQgYS9Ub29scy9TY3JpcHRzL3dlYmtpdHB5L2xheW91dF90ZXN0cy9wb3J0L2Jhc2VfdW5pdHRl c3QucHkgYi9Ub29scy9TY3JpcHRzL3dlYmtpdHB5L2xheW91dF90ZXN0cy9wb3J0L2Jhc2VfdW5p dHRlc3QucHkKaW5kZXggNTYzMzY3Y2NiMWI2YjExZDdjNTE5MDM3NjdhYTI3YTYxNzdmMDEzYi4u OThjMjUzYzRhZTg0YmI1MmY4ZTYxNTUyMjE1Y2M4MmFjMTg4NjY1MCAxMDA2NDQKLS0tIGEvVG9v bHMvU2NyaXB0cy93ZWJraXRweS9sYXlvdXRfdGVzdHMvcG9ydC9iYXNlX3VuaXR0ZXN0LnB5Cisr KyBiL1Rvb2xzL1NjcmlwdHMvd2Via2l0cHkvbGF5b3V0X3Rlc3RzL3BvcnQvYmFzZV91bml0dGVz dC5weQpAQCAtMzQsNyArMzQsNyBAQCBpbXBvcnQgdW5pdHRlc3QKIGZyb20gd2Via2l0cHkuY29t bW9uLnN5c3RlbS5leGVjdXRpdmUgaW1wb3J0IEV4ZWN1dGl2ZSwgU2NyaXB0RXJyb3IKIGZyb20g d2Via2l0cHkuY29tbW9uLnN5c3RlbSBpbXBvcnQgZXhlY3V0aXZlX21vY2sKIGZyb20gd2Via2l0 cHkuY29tbW9uLnN5c3RlbS5maWxlc3lzdGVtX21vY2sgaW1wb3J0IE1vY2tGaWxlU3lzdGVtCi1m cm9tIHdlYmtpdHB5LmNvbW1vbi5zeXN0ZW0gaW1wb3J0IG91dHB1dGNhcHR1cmUKK2Zyb20gd2Vi a2l0cHkuY29tbW9uLnN5c3RlbS5vdXRwdXRjYXB0dXJlIGltcG9ydCBPdXRwdXRDYXB0dXJlCiBm cm9tIHdlYmtpdHB5LmNvbW1vbi5zeXN0ZW0ucGF0aCBpbXBvcnQgYWJzcGF0aF90b191cmkKIGZy b20gd2Via2l0cHkudGhpcmRwYXJ0eS5tb2NrIGltcG9ydCBNb2NrCiBmcm9tIHdlYmtpdHB5LnRv b2wubW9ja3Rvb2wgaW1wb3J0IE1vY2tPcHRpb25zCkBAIC01Myw2ICs1MywyMSBAQCBjbGFzcyBQ b3J0VGVzdCh1bml0dGVzdC5UZXN0Q2FzZSk6CiAgICAgICAgIGhvc3QgPSBob3N0IG9yIE1vY2tI b3N0KCkKICAgICAgICAgcmV0dXJuIFBvcnQoaG9zdCwgKiprd2FyZ3MpCiAKKyAgICBkZWYgdGVz dF9kZWZhdWx0X2NoaWxkX3Byb2Nlc3NlcyhzZWxmKToKKyAgICAgICAgcG9ydCA9IHNlbGYubWFr ZV9wb3J0KCkKKyAgICAgICAgc2VsZi5hc3NlcnRFcXVhbChwb3J0LmRlZmF1bHRfY2hpbGRfcHJv Y2Vzc2VzKCksIDIpCisgICAgICAgIGJ5dGVzX2Zvcl9kcnQgPSAzMDAgKiAxMDI0ICogMTAyNAor CisgICAgICAgIHBvcnQuaG9zdC5wbGF0Zm9ybS50b3RhbF9ieXRlc19tZW1vcnkgPSBsYW1iZGE6 IGJ5dGVzX2Zvcl9kcnQKKyAgICAgICAgZXhwZWN0ZWRfc3Rkb3V0ID0gIlRoaXMgbWFjaGluZSBj b3VsZCBzdXBwb3J0IDIgY2hpbGQgcHJvY2Vzc2VzLCBidXQgb25seSBoYXMgZW5vdWdoIG1lbW9y eSBmb3IgMS5cbiIKKyAgICAgICAgY2hpbGRfcHJvY2Vzc2VzID0gT3V0cHV0Q2FwdHVyZSgpLmFz c2VydF9vdXRwdXRzKHNlbGYsIHBvcnQuZGVmYXVsdF9jaGlsZF9wcm9jZXNzZXMsICgpLCBleHBl Y3RlZF9zdGRvdXQ9ZXhwZWN0ZWRfc3Rkb3V0KQorICAgICAgICBzZWxmLmFzc2VydEVxdWFsKGNo aWxkX3Byb2Nlc3NlcywgMSkKKworICAgICAgICBwb3J0Lmhvc3QucGxhdGZvcm0udG90YWxfYnl0 ZXNfbWVtb3J5ID0gbGFtYmRhOiBieXRlc19mb3JfZHJ0IC0gMQorICAgICAgICBleHBlY3RlZF9z dGRvdXQgPSAiVGhpcyBtYWNoaW5lIGNvdWxkIHN1cHBvcnQgMiBjaGlsZCBwcm9jZXNzZXMsIGJ1 dCBvbmx5IGhhcyBlbm91Z2ggbWVtb3J5IGZvciAwLlxuIgorICAgICAgICBjaGlsZF9wcm9jZXNz ZXMgPSBPdXRwdXRDYXB0dXJlKCkuYXNzZXJ0X291dHB1dHMoc2VsZiwgcG9ydC5kZWZhdWx0X2No aWxkX3Byb2Nlc3NlcywgKCksIGV4cGVjdGVkX3N0ZG91dD1leHBlY3RlZF9zdGRvdXQpCisgICAg ICAgIHNlbGYuYXNzZXJ0RXF1YWwoY2hpbGRfcHJvY2Vzc2VzLCAwKQorCiAgICAgZGVmIHRlc3Rf Zm9ybWF0X3dkaWZmX291dHB1dF9hc19odG1sKHNlbGYpOgogICAgICAgICBvdXRwdXQgPSAiT1VU UFVUICVzICVzICVzIiAlIChQb3J0Ll9XRElGRl9ERUwsIFBvcnQuX1dESUZGX0FERCwgUG9ydC5f V0RJRkZfRU5EKQogICAgICAgICBodG1sID0gc2VsZi5tYWtlX3BvcnQoKS5fZm9ybWF0X3dkaWZm X291dHB1dF9hc19odG1sKG91dHB1dCkKQEAgLTgyLDcgKzk3LDcgQEAgY2xhc3MgUG9ydFRlc3Qo dW5pdHRlc3QuVGVzdENhc2UpOgogCiAgICAgZGVmIHRlc3RfcHJldHR5X3BhdGNoX29zX2Vycm9y KHNlbGYpOgogICAgICAgICBwb3J0ID0gc2VsZi5tYWtlX3BvcnQoZXhlY3V0aXZlPWV4ZWN1dGl2 ZV9tb2NrLk1vY2tFeGVjdXRpdmUyKGV4Y2VwdGlvbj1PU0Vycm9yKSkKLSAgICAgICAgb2MgPSBv dXRwdXRjYXB0dXJlLk91dHB1dENhcHR1cmUoKQorICAgICAgICBvYyA9IE91dHB1dENhcHR1cmUo KQogICAgICAgICBvYy5jYXB0dXJlX291dHB1dCgpCiAgICAgICAgIHNlbGYuYXNzZXJ0RXF1YWwo cG9ydC5wcmV0dHlfcGF0Y2hfdGV4dCgicGF0Y2gudHh0IiksCiAgICAgICAgICAgICAgICAgICAg ICAgICAgcG9ydC5fcHJldHR5X3BhdGNoX2Vycm9yX2h0bWwpCg==