Please review the fix for JDK-8160941 . Webrev: http://cr.openjdk.java.net/~rstevens/8160941/webrev/
The problem: when copy-pasting (or drag-and-dropping) multiple files, the data in the transferable for the flavor "text/uri-list" concats the first two paths. If you for example copy: /Users/robin/Desktop/file1.txt /Users/robin/Desktop/file2.txt /Users/robin/Desktop/file3.txt the data in the transferable is /Users/robin/Desktop/file1.txt/Users/robin/Desktop/file2.txt /Users/robin/Desktop/file3.txt while the expected data is /Users/robin/Desktop/file1.txt /Users/robin/Desktop/file2.txt /Users/robin/Desktop/file3.txt This is also what you can observe when running the manual testcase. As there was already a manual testcase available (which fails on JDK8), the patch does not include one. Thanks, Robin On Fri, Jul 8, 2016 at 9:53 AM, Alexandr Scherbatiy < alexandr.scherba...@oracle.com> wrote: > On 7/7/2016 10:07 PM, Robin Stevens wrote: > > Thanks Alexander for creating the issue in the bug tracker and hosting the > webrev. > Do I need to send a new review request to the list with the official bug > number in the subject, or is the current email thread sufficient ? > > > Yes. You can just reply on this email, correct the bug id and provide > the webrev link. > > Thanks, > Alexandr. > > > Robin > > On Thu, Jul 7, 2016 at 9:36 AM, Alexandr Scherbatiy < > alexandr.scherba...@oracle.com> wrote: > >> >> The issue is recorded under id JDK-8160941 "text/uri-list" dataflavor >> concats the first two strings >> https://bugs.openjdk.java.net/browse/JDK-8160941 >> >> The webrev is uploaded to >> http://cr.openjdk.java.net/~alexsch/robin.stevens/8160941/webrev.00 >> >> Thanks, >> Alexandr. >> >> On 7/7/2016 10:10 AM, Robin Stevens wrote: >> >> Hello, >> >> the backport of the fix for >> https://bugs.openjdk.java.net/browse/JDK-8136763 looks incorrect. >> The corresponding manual test case fails on jdk8. >> >> The problem: when copy-pasting (or drag-and-dropping) multiple files, the >> data in the transferable for the flavor "text/uri-list" concats the first >> two paths. >> >> If you for example copy: >> /Users/robin/Desktop/file1.txt >> /Users/robin/Desktop/file2.txt >> /Users/robin/Desktop/file3.txt >> >> the data in the transferable is >> >> /Users/robin/Desktop/file1.txt/Users/robin/Desktop/file2.txt >> /Users/robin/Desktop/file3.txt >> >> while the expected data is >> >> /Users/robin/Desktop/file1.txt >> /Users/robin/Desktop/file2.txt >> /Users/robin/Desktop/file3.txt >> >> >> This is also what you can observe when running the manual testcase. >> >> I have logged this in the bug database as issue JI-9041413. >> >> Attached you find a webrev with the proposed patch. >> The manual testcase succeeds with this patch, and fails without. >> At the bottom of this email, you also find the output of hg diff. >> >> The patch does not include a testcase, as there is already one available. >> >> Regards, >> >> Robin >> >> >> hg diff output: >> >> diff -r 0844fa517c35 >> src/macosx/classes/sun/lwawt/macosx/CDataTransferer.java >> --- a/src/macosx/classes/sun/lwawt/macosx/CDataTransferer.java Tue Jul >> 05 11:03:13 2016 -0700 >> +++ b/src/macosx/classes/sun/lwawt/macosx/CDataTransferer.java Thu Jul >> 07 08:48:37 2016 +0200 >> @@ -148,6 +148,7 @@ >> StringBuilder sb = new StringBuilder(); >> if(strings.length > 0) { >> sb.append(strings[0]); >> + sb.append(separator); >> for(int i = 1; i < strings.length; i++) { >> sb.append(strings[i]); >> sb.append(separator); >> >> >> > >
