23077 2009-01-02 08:23:16 -0800 Make application cache use SQLite built-in user_version 2009-01-05 07:21:14 -0800 1 1 1 Unclassified WebKit WebCore Misc. 528+ (Nightly build) All All RESOLVED FIXED P3 Normal --- 1 ap ap andersca darin ddkilzer mjs oldest_to_newest 104085 0 ap 2009-01-02 08:23:16 -0800 Anders says we should use built-in SQLite versioning instead of SchemaVersion table. 104086 1 26370 ap 2009-01-02 08:26:58 -0800 Created attachment 26370 proposed patch 104092 2 26370 darin 2009-01-02 09:28:41 -0800 Comment on attachment 26370 proposed patch > + if (static_cast<int>(sizeof(userVersionSQL)) < numBytes) > + ASSERT_NOT_REACHED(); This is a strange way to write an assertion. Are you just trying to avoid the unused local variable warning? Is there a better idiom for this sort of thing? It seems to come up often. r=me 104098 3 ap 2009-01-02 09:43:12 -0800 Yes, that was the reason. Other ways to write this are to put the left part of variable definition/assignment under ifndef NDEBUG, or to add UNUSED_PARAM() for the variable. I'm not sure if either is better. 104154 4 darin 2009-01-02 12:46:49 -0800 (In reply to comment #3) > Yes, that was the reason. Other ways to write this are to put the left part of > variable definition/assignment under ifndef NDEBUG, or to add UNUSED_PARAM() > for the variable. I'm not sure if either is better. This comes up often enough that I think we should add a macro just for this purpose: ASSERT_RESULT(variableName, result); It could use UNUSED_PARAM in NDEBUG and plain old ASSERT in debug. An advantage of this over ASSERT_NOT_REACHED is that the assertion message can be more specific. And also, clarity. It's way better than the #ifndef NDEBUG and the explicit use of UNUSED_PARAM for something that's not a "PARAM". What do you think? 104170 5 ap 2009-01-02 13:31:49 -0800 Sounds good to me. I'll make a patch (do you think that adding this to wtf/UnusedParam.h would be acceptable? or should I rename it to Unused.h?) 104175 6 darin 2009-01-02 14:00:05 -0800 (In reply to comment #5) > I'll make a patch (do you think that adding this to > wtf/UnusedParam.h would be acceptable? I think this is part of Assertions.h. The fact that it depends on UnusedParam.h is an implementation detail. > or should I rename it to Unused.h?) That might be good to do eventually. 104176 7 darin 2009-01-02 14:03:08 -0800 (In reply to comment #6) > I think this is part of Assertions.h. The fact that it depends on UnusedParam.h > is an implementation detail. And whether it actually uses UnusedParam.h or not is an open question in my opinion. 104407 8 ap 2009-01-05 00:56:23 -0800 Committed revision 39597. I'll make a separate patch for ASSERT_RESULT. 104425 9 darin 2009-01-05 07:21:14 -0800 (In reply to comment #8) > Committed revision 39597. I'll make a separate patch for ASSERT_RESULT. I was also thinking we might want a similar ASSERT_UNUSED_PARAM function for an assertion on a parameter that is otherwise unused. Sorry to continue hijacking this bug report for this discussion. 26370 2009-01-02 08:26:58 -0800 2009-01-02 09:28:41 -0800 proposed patch UserVersion.txt text/plain 3507 ap SW5kZXg6IFdlYkNvcmUvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFdlYkNvcmUvQ2hhbmdlTG9n CShyZXZpc2lvbiAzOTU0OSkKKysrIFdlYkNvcmUvQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBA IC0xLDMgKzEsMTggQEAKKzIwMDktMDEtMDIgIEFsZXhleSBQcm9za3VyeWFrb3YgIDxhcEB3ZWJr aXQub3JnPgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAg IGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD0yMzA3NworICAgICAgICBN YWtlIGFwcGxpY2F0aW9uIGNhY2hlIHVzZSBTUUxpdGUgYnVpbHQtaW4gdXNlcl92ZXJzaW9uCisK KyAgICAgICAgKiBsb2FkZXIvYXBwY2FjaGUvQXBwbGljYXRpb25DYWNoZVN0b3JhZ2UuY3BwOgor ICAgICAgICAoV2ViQ29yZTo6QXBwbGljYXRpb25DYWNoZVN0b3JhZ2U6OnZlcmlmeVNjaGVtYVZl cnNpb24pOiBDaGFuZ2VkIHRvIHRyYWNrIHZlcnNpb25zIHdpdGgKKyAgICAgICAgUFJBR01BIHVz ZXJfdmVyc2lvbi4KKworICAgICAgICAqIGxvYWRlci9hcHBjYWNoZS9BcHBsaWNhdGlvbkNhY2hl UmVzb3VyY2UuaDogKFdlYkNvcmU6OkFwcGxpY2F0aW9uQ2FjaGVSZXNvdXJjZTo6KToKKyAgICAg ICAgVXNlZCB0aGUgb2NjYXNpb24gdG8gZ2V0IHJpZCBvZiB0aGUgaG9sZSBpbiBiaXRtYXNrIHRo YXQgd2FzIGEgbGVmdG92ZXIgZnJvbSBvcHBvcnR1bmlzdGljCisgICAgICAgIGNhY2hlIGVudHJp ZXMuCisKIDIwMDktMDEtMDIgIEFsZXhleSBQcm9za3VyeWFrb3YgIDxhcEB3ZWJraXQub3JnPgog CiAgICAgICAgIFJldmlld2VkIGJ5IERhcmluIEFkbGVyLgpJbmRleDogV2ViQ29yZS9sb2FkZXIv YXBwY2FjaGUvQXBwbGljYXRpb25DYWNoZVJlc291cmNlLmgKPT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gV2ViQ29y ZS9sb2FkZXIvYXBwY2FjaGUvQXBwbGljYXRpb25DYWNoZVJlc291cmNlLmgJKHJldmlzaW9uIDM5 NTQ3KQorKysgV2ViQ29yZS9sb2FkZXIvYXBwY2FjaGUvQXBwbGljYXRpb25DYWNoZVJlc291cmNl LmgJKHdvcmtpbmcgY29weSkKQEAgLTQwLDggKzQwLDcgQEAgcHVibGljOgogICAgICAgICBFeHBs aWNpdCA9IDEgPDwgMiwKICAgICAgICAgRm9yZWlnbiA9IDEgPDwgMywKICAgICAgICAgRmFsbGJh Y2sgPSAxIDw8IDQsCi0gICAgICAgIC8vIEJpdCA1IHdhcyB1c2VkIGZvciBvcHBvcnR1bmlzdGlj IGVudHJpZXMsIHdoaWNoIHdlcmUgcmVtb3ZlZCBmcm9tIHRoZSBzcGVjLgotICAgICAgICBEeW5h bWljID0gMSA8PCA2CisgICAgICAgIER5bmFtaWMgPSAxIDw8IDUKICAgICB9OwogICAgICAgICAK ICAgICBzdGF0aWMgUGFzc1JlZlB0cjxBcHBsaWNhdGlvbkNhY2hlUmVzb3VyY2U+IGNyZWF0ZShj b25zdCBLVVJMJiB1cmwsIGNvbnN0IFJlc291cmNlUmVzcG9uc2UmIHJlc3BvbnNlLCB1bnNpZ25l ZCB0eXBlLCBQYXNzUmVmUHRyPFNoYXJlZEJ1ZmZlcj4gYnVmZmVyID0gU2hhcmVkQnVmZmVyOjpj cmVhdGUoKSkKSW5kZXg6IFdlYkNvcmUvbG9hZGVyL2FwcGNhY2hlL0FwcGxpY2F0aW9uQ2FjaGVT dG9yYWdlLmNwcAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09Ci0tLSBXZWJDb3JlL2xvYWRlci9hcHBjYWNoZS9BcHBsaWNh dGlvbkNhY2hlU3RvcmFnZS5jcHAJKHJldmlzaW9uIDM5NTQ4KQorKysgV2ViQ29yZS9sb2FkZXIv YXBwY2FjaGUvQXBwbGljYXRpb25DYWNoZVN0b3JhZ2UuY3BwCSh3b3JraW5nIGNvcHkpCkBAIC0y OTgsMzAgKzI5OCwzMSBAQCBib29sIEFwcGxpY2F0aW9uQ2FjaGVTdG9yYWdlOjpleGVjdXRlU1FM CiAgICAgcmV0dXJuIHJlc3VsdDsKIH0KIAotc3RhdGljIGNvbnN0IGludCBTY2hlbWFWZXJzaW9u ID0gMzsKK3N0YXRpYyBjb25zdCBpbnQgc2NoZW1hVmVyc2lvbiA9IDM7CiAgICAgCiB2b2lkIEFw cGxpY2F0aW9uQ2FjaGVTdG9yYWdlOjp2ZXJpZnlTY2hlbWFWZXJzaW9uKCkKIHsKLSAgICBpZiAo bV9kYXRhYmFzZS50YWJsZUV4aXN0cygiU2NoZW1hVmVyc2lvbiIpKSB7Ci0gICAgICAgIGludCB2 ZXJzaW9uID0gU1FMaXRlU3RhdGVtZW50KG1fZGF0YWJhc2UsICJTRUxFQ1QgdmVyc2lvbiBmcm9t IFNjaGVtYVZlcnNpb24iKS5nZXRDb2x1bW5JbnQoMCk7Ci0gICAgICAgIAotICAgICAgICBpZiAo dmVyc2lvbiA9PSBTY2hlbWFWZXJzaW9uKQotICAgICAgICAgICAgcmV0dXJuOwotICAgIH0KLSAg ICAKKyAgICBpbnQgdmVyc2lvbiA9IFNRTGl0ZVN0YXRlbWVudChtX2RhdGFiYXNlLCAiUFJBR01B IHVzZXJfdmVyc2lvbiIpLmdldENvbHVtbkludCgwKTsKKyAgICBpZiAodmVyc2lvbiA9PSBzY2hl bWFWZXJzaW9uKQorICAgICAgICByZXR1cm47CisKICAgICBtX2RhdGFiYXNlLmNsZWFyQWxsVGFi bGVzKCk7CiAKLSAgICBTUUxpdGVUcmFuc2FjdGlvbiBjcmVhdGVTY2hlbWFWZXJzaW9uVGFibGUo bV9kYXRhYmFzZSk7Ci0gICAgY3JlYXRlU2NoZW1hVmVyc2lvblRhYmxlLmJlZ2luKCk7CisgICAg Ly8gVXBkYXRlIHVzZXIgdmVyc2lvbi4KKyAgICBTUUxpdGVUcmFuc2FjdGlvbiBzZXREYXRhYmFz ZVZlcnNpb24obV9kYXRhYmFzZSk7CisgICAgc2V0RGF0YWJhc2VWZXJzaW9uLmJlZ2luKCk7CisK KyAgICBjaGFyIHVzZXJWZXJzaW9uU1FMWzMyXTsKKyAgICBpbnQgbnVtQnl0ZXMgPSBzbnByaW50 Zih1c2VyVmVyc2lvblNRTCwgc2l6ZW9mKHVzZXJWZXJzaW9uU1FMKSwgIlBSQUdNQSB1c2VyX3Zl cnNpb249JWQiLCBzY2hlbWFWZXJzaW9uKTsKKyAgICBpZiAoc3RhdGljX2Nhc3Q8aW50PihzaXpl b2YodXNlclZlcnNpb25TUUwpKSA8IG51bUJ5dGVzKQorICAgICAgICBBU1NFUlRfTk9UX1JFQUNI RUQoKTsKIAotICAgIGV4ZWN1dGVTUUxDb21tYW5kKCJDUkVBVEUgVEFCTEUgU2NoZW1hVmVyc2lv biAodmVyc2lvbiBJTlRFR0VSIE5PVCBOVUxMKSIpOwotICAgIFNRTGl0ZVN0YXRlbWVudCBzdGF0 ZW1lbnQobV9kYXRhYmFzZSwgIklOU0VSVCBJTlRPIFNjaGVtYVZlcnNpb24gKHZlcnNpb24pIFZB TFVFUyAoPykiKTsKKyAgICBTUUxpdGVTdGF0ZW1lbnQgc3RhdGVtZW50KG1fZGF0YWJhc2UsIHVz ZXJWZXJzaW9uU1FMKTsKICAgICBpZiAoc3RhdGVtZW50LnByZXBhcmUoKSAhPSBTUUxSZXN1bHRP aykKICAgICAgICAgcmV0dXJuOwogICAgIAotICAgIHN0YXRlbWVudC5iaW5kSW50NjQoMSwgU2No ZW1hVmVyc2lvbik7CiAgICAgZXhlY3V0ZVN0YXRlbWVudChzdGF0ZW1lbnQpOwotICAgIGNyZWF0 ZVNjaGVtYVZlcnNpb25UYWJsZS5jb21taXQoKTsKKyAgICBzZXREYXRhYmFzZVZlcnNpb24uY29t bWl0KCk7CiB9CiAgICAgCiB2b2lkIEFwcGxpY2F0aW9uQ2FjaGVTdG9yYWdlOjpvcGVuRGF0YWJh c2UoYm9vbCBjcmVhdGVJZkRvZXNOb3RFeGlzdCkK