Bug 157072

Created attachment 277546 [details] Patch

Attachment 277546 [details] did not pass style-queue: ERROR: Source/WebCore/Modules/indexeddb/client/IDBConnectionToServer.cpp:414: Extra space before ( in function call [whitespace/parens] [4] ERROR: Source/WebCore/Modules/indexeddb/client/IDBConnectionProxy.h:64: Extra space before ( in function call [whitespace/parens] [4] ERROR: Source/WebCore/Modules/indexeddb/client/IDBConnectionToServer.h:126: Extra space before ( in function call [whitespace/parens] [4] ERROR: Source/WebCore/Modules/indexeddb/client/IDBConnectionToServer.h:146: Extra space before ( in function call [whitespace/parens] [4] ERROR: Source/WebCore/Modules/indexeddb/IDBFactory.cpp:162: Extra space before ( in function call [whitespace/parens] [4] ERROR: Source/WebCore/Modules/indexeddb/client/IDBConnectionProxy.cpp:85: Extra space before ( in function call [whitespace/parens] [4] ERROR: Source/WebCore/Modules/indexeddb/IDBFactory.h:64: Extra space before ( in function call [whitespace/parens] [4] Total errors found: 7 in 42 files If any of these errors are false positives, please file a bug against check-webkit-style.

Comment on attachment 277546 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=277546&action=review > Source/WTF/wtf/HexNumber.h:117 > + return (c >= '0' && c <= '9') || (c >= 'A' && c <= 'F') || (c >= 'a' && c <= 'f'); We can mask the characters then compare to reduce the number of operations done here. > Source/WTF/wtf/HexNumber.h:120 > +inline char hexDigit(int i) "unchecked" here would match uncheckedHexDigitValue > Source/WTF/wtf/HexNumber.h:125 > + return (i >= 10) ? i - 10 + 'A' : i += '0'; +=? This only needs to be +. Probably optimized out, though. > Source/WebCore/platform/FileSystem.cpp:150 > + if (i + 2 < length) >= > Source/WebCore/platform/FileSystem.cpp:168 > + if (i + 5 < length) >=

Comment on attachment 277546 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=277546&action=review > Source/WebCore/platform/posix/FileSystemPOSIX.cpp:260 > +String lastComponentOfPathIgnoringTrailingSlash(const String& path) This is not implemented in GTK. Move to FileSystem.cpp.

Created attachment 277548 [details] Patch for EWS + landing

Attachment 277548 [details] did not pass style-queue: ERROR: Source/WebCore/Modules/indexeddb/client/IDBConnectionToServer.cpp:414: Extra space before ( in function call [whitespace/parens] [4] ERROR: Source/WebCore/Modules/indexeddb/client/IDBConnectionProxy.h:64: Extra space before ( in function call [whitespace/parens] [4] ERROR: Source/WebCore/Modules/indexeddb/client/IDBConnectionToServer.h:126: Extra space before ( in function call [whitespace/parens] [4] ERROR: Source/WebCore/Modules/indexeddb/client/IDBConnectionToServer.h:146: Extra space before ( in function call [whitespace/parens] [4] ERROR: Source/WebCore/Modules/indexeddb/IDBFactory.cpp:162: Extra space before ( in function call [whitespace/parens] [4] ERROR: Source/WebCore/Modules/indexeddb/client/IDBConnectionProxy.cpp:85: Extra space before ( in function call [whitespace/parens] [4] ERROR: Source/WebCore/Modules/indexeddb/IDBFactory.h:64: Extra space before ( in function call [whitespace/parens] [4] Total errors found: 7 in 42 files If any of these errors are false positives, please file a bug against check-webkit-style.

Comment on attachment 277548 [details] Patch for EWS + landing Clearing flags on attachment: 277548 Committed r200163: <http://trac.webkit.org/changeset/200163>

All reviewed patches have been landed. Closing bug.

Welp, EFL doesn't use FileSystemGlib, apparently... *sigh* Working on a fix.

EFL build fix attempted in http://trac.webkit.org/changeset/200168

I think that instead of lastComponentOfPathIgnoringTrailingSlash() introduction it would be better to add a possibility to get list of relative paths from listDirectory(). Right now, listDirectory() implementations (posix and glib) iterate through file names in a directory and convert them to absolute. After that, you extract that original file name from absolute path with lastComponentOfPathIgnoringTrailingSlash(). Looks like the easiest way is to add a flag to listDirectory to control if absolute or relative paths are requested. If you like this proposal I can prepare the patch.

(In reply to comment #11) > If you like this proposal I can prepare the patch. If you think you can make this cleaner then it never hurts to prepare a patch. Once we see an implementation, the worst we could do is reject it. I would believe this could be done cleaner, though.

(In reply to comment #12) > (In reply to comment #11) > > If you like this proposal I can prepare the patch. > If you think you can make this cleaner then it never hurts to prepare a > patch. Once we see an implementation, the worst we could do is reject it. > I would believe this could be done cleaner, though. Just please do it in a new bug. It's harmful to discuss future plans in an already-resolved bug.