85111 2012-04-27 16:42:58 -0700 PerfTestRunner.computeStatistics incorrectly calculates min, max and median 2012-04-30 08:53:55 -0700 1 1 1 Unclassified WebKit New Bugs 528+ (Nightly build) Unspecified Unspecified RESOLVED FIXED P2 Normal --- 1 webkit.review.bot tomz rniwa tomz oldest_to_newest 612016 0 webkit.review.bot 2012-04-27 16:42:58 -0700 PerfTestRunner.computeStatistics incorrectly calculates min, max and median Requested by tomz on #webkit. 612017 1 tomz 2012-04-27 16:44:40 -0700 The data.sort() in this routine is doing an alphabetic sort. The dataset [1 3 7 10] shows the problem. 612021 2 139308 tomz 2012-04-27 16:49:34 -0700 Created attachment 139308 Patch 612024 3 tomz 2012-04-27 16:55:16 -0700 I'm not sure where tests for this should go but we most *certainly* need them for the statistics package. I'll be happy to write some. 612034 4 rniwa 2012-04-27 17:09:20 -0700 Unfortunately we don't have any tests for this, and that's why we have this bug. Maybe we can add something in LayoutTests/fast/harness/ that pulls this file? 612037 5 rniwa 2012-04-27 17:11:51 -0700 Now it makes sense that some of min/max/stdev looked bogus. But this problem should go away once we start importing all values. Still nice to fix it here thouh. 612173 6 tomz 2012-04-27 23:30:42 -0700 In that case, could you r+ this and I'll get the tests in ASAP? I'd rather not hold off while we're continuing to generate incorrect data. I tested the fix manually. 612213 7 139308 rniwa 2012-04-28 02:26:42 -0700 Comment on attachment 139308 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=139308&action=review > PerformanceTests/ChangeLog:9 > + Reviewed by NOBODY (OOPS!). This line should appear before the description. 612440 8 tomz 2012-04-29 11:21:29 -0700 Committed r115598: <http://trac.webkit.org/changeset/115598> 139308 2012-04-27 16:49:34 -0700 2012-04-30 08:53:55 -0700 Patch bug-85111-20120427164932.patch text/plain 1385 tomz U3VidmVyc2lvbiBSZXZpc2lvbjogMTE1NTE0CmRpZmYgLS1naXQgYS9QZXJmb3JtYW5jZVRlc3Rz L0NoYW5nZUxvZyBiL1BlcmZvcm1hbmNlVGVzdHMvQ2hhbmdlTG9nCmluZGV4IGI5MGE4MTMyNDVh NjA1NTc4YzRkNmNmMmU0ZGU3ZjEyNzY2YWIyMmIuLmVlY2IzNzMzNzE0NDNkYmJmMWZhZDYyZjk0 ZDc2NWMxOWY5MGRkMTggMTAwNjQ0Ci0tLSBhL1BlcmZvcm1hbmNlVGVzdHMvQ2hhbmdlTG9nCisr KyBiL1BlcmZvcm1hbmNlVGVzdHMvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTUgQEAKKzIwMTItMDQt MjcgIFRvbSBaYWtyYWpzZWsgIDx0b216QGNvZGVhdXJvcmEub3JnPgorCisgICAgICAgIFBlcmZU ZXN0UnVubmVyLmNvbXB1dGVTdGF0aXN0aWNzIGluY29ycmVjdGx5IGNhbGN1bGF0ZXMgbWluLCBt YXggYW5kIG1lZGlhbgorICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5j Z2k/aWQ9ODUxMTEKKworICAgICAgICBUaGUgc29ydCBvZiB0aGUgZGF0YSBpbnB1dCB3YXMgYmVp bmcgZG9uZSBhbHBoYWJldGljYWxseS4KKyAgICAgICAgU28gSSBwcm92aWRlZCBhIG51bWVyaWMg Y29tcGFyZSBmdW5jdGlvbi4KKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4K KworICAgICAgICAqIHJlc291cmNlcy9ydW5uZXIuanM6CisKIDIwMTItMDQtMTcgIElseWEgVGlr aG9ub3Zza3kgIDxsb2lzbG9AY2hyb21pdW0ub3JnPgogCiAgICAgICAgIFdlYiBJbnNwZWN0b3I6 IHRoZSBmaXggZm9yIHRlc3QgYWZ0ZXIgbWFzc2l2ZSByZW5hbWUgaW4gcjExNDI3MS4KZGlmZiAt LWdpdCBhL1BlcmZvcm1hbmNlVGVzdHMvcmVzb3VyY2VzL3J1bm5lci5qcyBiL1BlcmZvcm1hbmNl VGVzdHMvcmVzb3VyY2VzL3J1bm5lci5qcwppbmRleCAxYmMyYTEwYjY5ZWFkNTllMGJkNmY5OTdj NDA0ODIwZmFjYWY2ODcwLi43MTBkOTNiN2I2MWQyYTQ1NDFlODE1M2Y2OTFjOWUzNGNjMzYzMjFh IDEwMDY0NAotLS0gYS9QZXJmb3JtYW5jZVRlc3RzL3Jlc291cmNlcy9ydW5uZXIuanMKKysrIGIv UGVyZm9ybWFuY2VUZXN0cy9yZXNvdXJjZXMvcnVubmVyLmpzCkBAIC01Miw3ICs1Miw3IEBAIFBl cmZUZXN0UnVubmVyLmNvbXB1dGVTdGF0aXN0aWNzID0gZnVuY3Rpb24gKHRpbWVzKSB7CiAgICAg dmFyIGRhdGEgPSB0aW1lcy5zbGljZSgpOwogCiAgICAgLy8gQWRkIHZhbHVlcyBmcm9tIHRoZSBz bWFsbGVzdCB0byB0aGUgbGFyZ2VzdCB0byBhdm9pZCB0aGUgbG9zcyBvZiBzaWduaWZpY2FuY2UK LSAgICBkYXRhLnNvcnQoKTsKKyAgICBkYXRhLnNvcnQoZnVuY3Rpb24oYSxiKXtyZXR1cm4gYS1i O30pOwogCiAgICAgdmFyIG1pZGRsZSA9IE1hdGguZmxvb3IoZGF0YS5sZW5ndGggLyAyKTsKICAg ICB2YXIgcmVzdWx0ID0gewo=