143826 2015-04-16 06:02:13 -0700 Use CommonCrypto for SHA1 and MD5 2015-04-17 10:54:42 -0700 1 1 1 Unclassified WebKit Web Template Framework 528+ (Nightly build) Unspecified Unspecified NEW P2 Normal --- 1 koivisto webkit-unassigned benjamin cmarcelo commit-queue oldest_to_newest 1085895 0 koivisto 2015-04-16 06:02:13 -0700 It is faster. 1085897 1 250918 koivisto 2015-04-16 06:11:18 -0700 Created attachment 250918 patch 1085943 2 250918 andersca 2015-04-16 09:33:06 -0700 Comment on attachment 250918 patch Seems like computeHash could just return the Digest. 1085972 3 koivisto 2015-04-16 10:45:57 -0700 https://trac.webkit.org/r182895 1086328 4 250918 darin 2015-04-17 09:18:47 -0700 Comment on attachment 250918 patch View in context: https://bugs.webkit.org/attachment.cgi?id=250918&action=review > Source/WTF/ChangeLog:22 > + Remove the side effect where computeHash resets the state. No one relies on it. All I see is the comment change and the new CommonCrypto. I don’t see a code change that makes that change for the non-CommonCrypto version. Is the idea that now computeHash might reset the state and might not? I think this is slightly untidy right now. I don’t know what the rules are about what you can and can’t do after calling computeHash and what behavior you should expect. 1086351 5 koivisto 2015-04-17 10:54:42 -0700 > All I see is the comment change and the new CommonCrypto. I don’t see a code > change that makes that change for the non-CommonCrypto version. Is the idea > that now computeHash might reset the state and might not? I think this is > slightly untidy right now. I don’t know what the rules are about what you > can and can’t do after calling computeHash and what behavior you should > expect. It seemed unnecessary to actually change the behavior of the existing implemention. It just calls reset() which sets the state back to initial state. This is not expensive and some sort of state cleanup is needed to not have computation left in memory. On the other hand I didn't want to introduce extra CC_SHA1_Init call to match this unused behavior. CC_SHA1_Final cleans up the state just fine. The new behavior is that you create new SHA1 instance for each computation. This is how the clients seem to expect it to behave anyway and matches similar MD5 type. 250918 2015-04-16 06:11:18 -0700 2015-04-16 09:33:06 -0700 patch sha1-cc.patch text/plain 6435 koivisto SW5kZXg6IFNvdXJjZS9XVEYvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFNvdXJjZS9XVEYvQ2hh bmdlTG9nCShyZXZpc2lvbiAxODI4ODIpCisrKyBTb3VyY2UvV1RGL0NoYW5nZUxvZwkod29ya2lu ZyBjb3B5KQpAQCAtMSwzICsxLDMwIEBACisyMDE1LTA0LTE2ICBBbnR0aSBLb2l2aXN0byAgPGFu dHRpQGFwcGxlLmNvbT4KKworICAgICAgICBVc2UgQ29tbW9uQ3J5cHRvIGZvciBTSEExIGFuZCBN RDUKKyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTE0Mzgy NgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgIENvbW1v bkNyeXB0byBTSEExIGltcGxlbWVudGF0aW9uIGlzID41eCBmYXN0ZXIgdGhhbiB0aGUgbmFpdmUg V1RGIG9uZS4gVXNlIGl0IHdoZW4gYXZhaWxhYmxlLgorCisgICAgICAgIFRoZXNlIGFyZSBjb3Zl cmVkIGJ5IGV4aXN0aW5nIEFQSSB0ZXN0cy4KKworICAgICAgICAqIHd0Zi9NRDUuY3BwOgorICAg ICAgICAoV1RGOjpNRDU6Ok1ENSk6CisgICAgICAgIChXVEY6Ok1ENTo6YWRkQnl0ZXMpOgorICAg ICAgICAoV1RGOjpNRDU6OmNoZWNrc3VtKToKKyAgICAgICAgKiB3dGYvTUQ1Lmg6CisgICAgICAg ICogd3RmL1NIQTEuY3BwOgorICAgICAgICAoV1RGOjpTSEExOjpTSEExKToKKyAgICAgICAgKFdU Rjo6U0hBMTo6YWRkQnl0ZXMpOgorICAgICAgICAoV1RGOjpTSEExOjpjb21wdXRlSGFzaCk6CisK KyAgICAgICAgICAgIFJlbW92ZSB0aGUgc2lkZSBlZmZlY3Qgd2hlcmUgY29tcHV0ZUhhc2ggcmVz ZXRzIHRoZSBzdGF0ZS4gTm8gb25lIHJlbGllcyBvbiBpdC4KKworICAgICAgICAoV1RGOjpTSEEx OjpoZXhEaWdlc3QpOgorICAgICAgICAoV1RGOjpTSEExOjpjb21wdXRlSGV4RGlnZXN0KToKKyAg ICAgICAgKiB3dGYvU0hBMS5oOgorCiAyMDE1LTA0LTE1ICDQoNGD0YHQu9Cw0L0g0JjQttCx0YPQ u9Cw0YLQvtCyICA8bHJuMTk4NkBnbWFpbC5jb20+CiAKICAgICAgICAgSGFybW9uaXplIGJpbmFy eSBzZW1hcGhvcmUgaWZkZWZzCkluZGV4OiBTb3VyY2UvV1RGL3d0Zi9NRDUuY3BwCj09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT0KLS0tIFNvdXJjZS9XVEYvd3RmL01ENS5jcHAJKHJldmlzaW9uIDE4Mjg0MCkKKysrIFNvdXJj ZS9XVEYvd3RmL01ENS5jcHAJKHdvcmtpbmcgY29weSkKQEAgLTIsNiArMiw3IEBACiAvLyBNb2Rp ZmljYXRpb25zIENvcHlyaWdodCAyMDA2IEdvb2dsZSBJbmMuIEFsbCBSaWdodHMgUmVzZXJ2ZWQK IC8qCiAgKiBDb3B5cmlnaHQgKEMpIDIwMTAgR29vZ2xlIEluYy4gQWxsIHJpZ2h0cyByZXNlcnZl ZC4KKyAqIENvcHlyaWdodCAoQykgMjAxNSBBcHBsZSBJbmMuIEFsbCByaWdodHMgcmVzZXJ2ZWQu CiAgKgogICogUmVkaXN0cmlidXRpb24gYW5kIHVzZSBpbiBzb3VyY2UgYW5kIGJpbmFyeSBmb3Jt cywgd2l0aCBvciB3aXRob3V0CiAgKiBtb2RpZmljYXRpb24sIGFyZSBwZXJtaXR0ZWQgcHJvdmlk ZWQgdGhhdCB0aGUgZm9sbG93aW5nIGNvbmRpdGlvbnMgYXJlCkBAIC01OCw2ICs1OSwyNSBAQAog CiBuYW1lc3BhY2UgV1RGIHsKIAorI2lmIFBMQVRGT1JNKENPQ09BKQorCitNRDU6Ok1ENSgpCit7 CisgICAgQ0NfTUQ1X0luaXQoJm1fY29udGV4dCk7Cit9CisKK3ZvaWQgTUQ1OjphZGRCeXRlcyhj b25zdCB1aW50OF90KiBpbnB1dCwgc2l6ZV90IGxlbmd0aCkKK3sKKyAgICBDQ19NRDVfVXBkYXRl KCZtX2NvbnRleHQsIGlucHV0LCBsZW5ndGgpOworfQorCit2b2lkIE1ENTo6Y2hlY2tzdW0oRGln ZXN0JiBoYXNoKQoreworICAgIENDX01ENV9GaW5hbChoYXNoLmRhdGEoKSwgJm1fY29udGV4dCk7 Cit9CisgICAgCisjZWxzZQorCiAvLyBOb3RlOiB0aGlzIGNvZGUgaXMgaGFybWxlc3Mgb24gbGl0 dGxlLWVuZGlhbiBtYWNoaW5lcy4KIAogc3RhdGljIHZvaWQgdG9MaXR0bGVFbmRpYW4odWludDhf dCogYnVmLCB1bnNpZ25lZCBsb25ncykKQEAgLTI2NCw0ICsyODQsNiBAQCB2b2lkIE1ENTo6Y2hl Y2tzdW0oRGlnZXN0JiBkaWdlc3QpCiAgICAgbWVtc2V0KG1faW4sIDAsIHNpemVvZihtX2luKSk7 CiB9CiAKKyNlbmRpZgorCiB9IC8vIG5hbWVzcGFjZSBXVEYKSW5kZXg6IFNvdXJjZS9XVEYvd3Rm L01ENS5oCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT0KLS0tIFNvdXJjZS9XVEYvd3RmL01ENS5oCShyZXZpc2lvbiAxODI4 NDApCisrKyBTb3VyY2UvV1RGL3d0Zi9NRDUuaAkod29ya2luZyBjb3B5KQpAQCAtMzQsNiArMzQs MTAgQEAKICNpbmNsdWRlIDxhcnJheT4KICNpbmNsdWRlIDx3dGYvVmVjdG9yLmg+CiAKKyNpZiBQ TEFURk9STShDT0NPQSkKKyNpbmNsdWRlIDxDb21tb25DcnlwdG8vQ29tbW9uRGlnZXN0Lmg+Cisj ZW5kaWYKKwogbmFtZXNwYWNlIFdURiB7CiAKIGNsYXNzIE1ENSB7CkBAIC01Niw5ICs2MCwxMyBA QCBwdWJsaWM6CiAgICAgV1RGX0VYUE9SVF9QUklWQVRFIHZvaWQgY2hlY2tzdW0oRGlnZXN0Jik7 CiAKIHByaXZhdGU6CisjaWYgUExBVEZPUk0oQ09DT0EpCisgICAgQ0NfTUQ1X0NUWCBtX2NvbnRl eHQ7CisjZWxzZQogICAgIHVpbnQzMl90IG1fYnVmWzRdOwogICAgIHVpbnQzMl90IG1fYml0c1sy XTsKICAgICB1aW50OF90IG1faW5bNjRdOworI2VuZGlmCiB9OwogCiB9IC8vIG5hbWVzcGFjZSBX VEYKSW5kZXg6IFNvdXJjZS9XVEYvd3RmL1NIQTEuY3BwCj09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFNvdXJjZS9X VEYvd3RmL1NIQTEuY3BwCShyZXZpc2lvbiAxODI4NDApCisrKyBTb3VyY2UvV1RGL3d0Zi9TSEEx LmNwcAkod29ya2luZyBjb3B5KQpAQCAtMSw1ICsxLDYgQEAKIC8qCiAgKiBDb3B5cmlnaHQgKEMp IDIwMTEgR29vZ2xlIEluYy4gQWxsIHJpZ2h0cyByZXNlcnZlZC4KKyAqIENvcHlyaWdodCAoQykg MjAxNSBBcHBsZSBJbmMuIEFsbCByaWdodHMgcmVzZXJ2ZWQuCiAgKgogICogUmVkaXN0cmlidXRp b24gYW5kIHVzZSBpbiBzb3VyY2UgYW5kIGJpbmFyeSBmb3Jtcywgd2l0aCBvciB3aXRob3V0CiAg KiBtb2RpZmljYXRpb24sIGFyZSBwZXJtaXR0ZWQgcHJvdmlkZWQgdGhhdCB0aGUgZm9sbG93aW5n IGNvbmRpdGlvbnMgYXJlCkBAIC0yOCwxMCArMjksNiBAQAogICogT0YgVEhJUyBTT0ZUV0FSRSwg RVZFTiBJRiBBRFZJU0VEIE9GIFRIRSBQT1NTSUJJTElUWSBPRiBTVUNIIERBTUFHRS4KICAqLwog Ci0vLyBBIHN0cmFpZ2h0Zm9yd2FyZCBTSEEtMSBpbXBsZW1lbnRhdGlvbiBiYXNlZCBvbiBSRkMg MzE3NC4KLS8vIGh0dHA6Ly93d3cuaWV0Zi5vcmcvcmZjL3JmYzMxNzQudHh0Ci0vLyBUaGUgbmFt ZXMgb2YgZnVuY3Rpb25zIGFuZCB2YXJpYWJsZXMgKHN1Y2ggYXMgImEiLCAiYiIsIGFuZCAiZiIp IGZvbGxvdyBub3RhdGlvbnMgaW4gUkZDIDMxNzQuCi0KICNpbmNsdWRlICJjb25maWcuaCIKICNp bmNsdWRlICJTSEExLmgiCiAKQEAgLTQyLDYgKzM5LDI5IEBACiAKIG5hbWVzcGFjZSBXVEYgewog CisjaWYgUExBVEZPUk0oQ09DT0EpCisKK1NIQTE6OlNIQTEoKQoreworICAgIENDX1NIQTFfSW5p dCgmbV9jb250ZXh0KTsKK30KKwordm9pZCBTSEExOjphZGRCeXRlcyhjb25zdCB1aW50OF90KiBp bnB1dCwgc2l6ZV90IGxlbmd0aCkKK3sKKyAgICBDQ19TSEExX1VwZGF0ZSgmbV9jb250ZXh0LCBp bnB1dCwgbGVuZ3RoKTsKK30KKwordm9pZCBTSEExOjpjb21wdXRlSGFzaChEaWdlc3QmIGhhc2gp Cit7CisgICAgQ0NfU0hBMV9GaW5hbChoYXNoLmRhdGEoKSwgJm1fY29udGV4dCk7Cit9CisKKyNl bHNlCisKKy8vIEEgc3RyYWlnaHRmb3J3YXJkIFNIQS0xIGltcGxlbWVudGF0aW9uIGJhc2VkIG9u IFJGQyAzMTc0LgorLy8gaHR0cDovL3d3dy5pZXRmLm9yZy9yZmMvcmZjMzE3NC50eHQKKy8vIFRo ZSBuYW1lcyBvZiBmdW5jdGlvbnMgYW5kIHZhcmlhYmxlcyAoc3VjaCBhcyAiYSIsICJiIiwgYW5k ICJmIikgZm9sbG93IG5vdGF0aW9ucyBpbiBSRkMgMzE3NC4KKwogc3RhdGljIGlubGluZSB1aW50 MzJfdCBmKGludCB0LCB1aW50MzJfdCBiLCB1aW50MzJfdCBjLCB1aW50MzJfdCBkKQogewogICAg IEFTU0VSVCh0ID49IDAgJiYgdCA8IDgwKTsKQEAgLTEwNCwyNSArMTI0LDYgQEAgdm9pZCBTSEEx Ojpjb21wdXRlSGFzaChEaWdlc3QmIGRpZ2VzdCkKICAgICByZXNldCgpOwogfQogCi1DU3RyaW5n IFNIQTE6OmhleERpZ2VzdChjb25zdCBEaWdlc3QmIGRpZ2VzdCkKLXsKLSAgICBjaGFyKiBzdGFy dCA9IDA7Ci0gICAgQ1N0cmluZyByZXN1bHQgPSBDU3RyaW5nOjpuZXdVbmluaXRpYWxpemVkKDQw LCBzdGFydCk7Ci0gICAgY2hhciogYnVmZmVyID0gc3RhcnQ7Ci0gICAgZm9yIChzaXplX3QgaSA9 IDA7IGkgPCBoYXNoU2l6ZTsgKytpKSB7Ci0gICAgICAgIHNucHJpbnRmKGJ1ZmZlciwgMywgIiUw MlgiLCBkaWdlc3QuYXQoaSkpOwotICAgICAgICBidWZmZXIgKz0gMjsKLSAgICB9Ci0gICAgcmV0 dXJuIHJlc3VsdDsKLX0KLQotQ1N0cmluZyBTSEExOjpjb21wdXRlSGV4RGlnZXN0KCkKLXsKLSAg ICBEaWdlc3QgZGlnZXN0OwotICAgIGNvbXB1dGVIYXNoKGRpZ2VzdCk7Ci0gICAgcmV0dXJuIGhl eERpZ2VzdChkaWdlc3QpOwotfQotCiB2b2lkIFNIQTE6OmZpbmFsaXplKCkKIHsKICAgICBBU1NF UlQobV9jdXJzb3IgPCA2NCk7CkBAIC0xOTUsNCArMTk2LDI1IEBAIHZvaWQgU0hBMTo6cmVzZXQo KQogICAgIG1lbXNldChtX2J1ZmZlciwgMCwgc2l6ZW9mKG1fYnVmZmVyKSk7CiB9CiAKKyNlbmRp ZgorCitDU3RyaW5nIFNIQTE6OmhleERpZ2VzdChjb25zdCBEaWdlc3QmIGRpZ2VzdCkKK3sKKyAg ICBjaGFyKiBzdGFydCA9IDA7CisgICAgQ1N0cmluZyByZXN1bHQgPSBDU3RyaW5nOjpuZXdVbmlu aXRpYWxpemVkKDQwLCBzdGFydCk7CisgICAgY2hhciogYnVmZmVyID0gc3RhcnQ7CisgICAgZm9y IChzaXplX3QgaSA9IDA7IGkgPCBoYXNoU2l6ZTsgKytpKSB7CisgICAgICAgIHNucHJpbnRmKGJ1 ZmZlciwgMywgIiUwMlgiLCBkaWdlc3QuYXQoaSkpOworICAgICAgICBidWZmZXIgKz0gMjsKKyAg ICB9CisgICAgcmV0dXJuIHJlc3VsdDsKK30KKworQ1N0cmluZyBTSEExOjpjb21wdXRlSGV4RGln ZXN0KCkKK3sKKyAgICBEaWdlc3QgZGlnZXN0OworICAgIGNvbXB1dGVIYXNoKGRpZ2VzdCk7Cisg ICAgcmV0dXJuIGhleERpZ2VzdChkaWdlc3QpOworfQorCiB9IC8vIG5hbWVzcGFjZSBXVEYKSW5k ZXg6IFNvdXJjZS9XVEYvd3RmL1NIQTEuaAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBTb3VyY2UvV1RGL3d0Zi9T SEExLmgJKHJldmlzaW9uIDE4Mjg0MCkKKysrIFNvdXJjZS9XVEYvd3RmL1NIQTEuaAkod29ya2lu ZyBjb3B5KQpAQCAtMzUsNiArMzUsMTAgQEAKICNpbmNsdWRlIDx3dGYvVmVjdG9yLmg+CiAjaW5j bHVkZSA8d3RmL3RleHQvQ1N0cmluZy5oPgogCisjaWYgUExBVEZPUk0oQ09DT0EpCisjaW5jbHVk ZSA8Q29tbW9uQ3J5cHRvL0NvbW1vbkRpZ2VzdC5oPgorI2VuZGlmCisKIG5hbWVzcGFjZSBXVEYg ewogCiBjbGFzcyBTSEExIHsKQEAgLTYzLDcgKzY3LDYgQEAgcHVibGljOgogICAgIC8vIHR5cGUg Zm9yIGNvbXB1dGluZyBTSEExIGhhc2gKICAgICB0eXBlZGVmIHN0ZDo6YXJyYXk8dWludDhfdCwg aGFzaFNpemU+IERpZ2VzdDsKIAotICAgIC8vIGNvbXB1dGVIYXNoIGhhcyBhIHNpZGUgZWZmZWN0 IG9mIHJlc2V0dGluZyB0aGUgc3RhdGUgb2YgdGhlIG9iamVjdC4KICAgICBXVEZfRVhQT1JUX1BS SVZBVEUgdm9pZCBjb21wdXRlSGFzaChEaWdlc3QmKTsKICAgICAKICAgICAvLyBHZXQgYSBoZXgg aGFzaCBmcm9tIHRoZSBkaWdlc3QuCkBAIC03Myw2ICs3Niw5IEBAIHB1YmxpYzoKICAgICBXVEZf RVhQT1JUX1BSSVZBVEUgQ1N0cmluZyBjb21wdXRlSGV4RGlnZXN0KCk7CiAKIHByaXZhdGU6Cisj aWYgUExBVEZPUk0oQ09DT0EpCisgICAgQ0NfU0hBMV9DVFggbV9jb250ZXh0OworI2Vsc2UKICAg ICB2b2lkIGZpbmFsaXplKCk7CiAgICAgdm9pZCBwcm9jZXNzQmxvY2soKTsKICAgICB2b2lkIHJl c2V0KCk7CkBAIC04MSw2ICs4Nyw3IEBAIHByaXZhdGU6CiAgICAgc2l6ZV90IG1fY3Vyc29yOyAv LyBOdW1iZXIgb2YgYnl0ZXMgZmlsbGVkIGluIG1fYnVmZmVyICgwLTY0KS4KICAgICB1aW50NjRf dCBtX3RvdGFsQnl0ZXM7IC8vIE51bWJlciBvZiBieXRlcyBhZGRlZCBzbyBmYXIuCiAgICAgdWlu dDMyX3QgbV9oYXNoWzVdOworI2VuZGlmCiB9OwogCiB9IC8vIG5hbWVzcGFjZSBXVEYK