hg: jdk7/jsn/jdk: 6621697: Problem with file test/sun/net/www/http/ChunkedInputStream/test.txt
Changeset: de7864ec310c Author:chegar Date: 2008-09-24 17:10 +0100 URL: http://hg.openjdk.java.net/jdk7/jsn/jdk/rev/de7864ec310c 6621697: Problem with file test/sun/net/www/http/ChunkedInputStream/test.txt Reviewed-by: michaelm ! test/sun/net/www/http/ChunkedInputStream/ChunkedEncodingTest.java - test/sun/net/www/http/ChunkedInputStream/test.txt
Re: ChunkedInputStream/test.txt
On Tue, 2008-09-23 at 09:46 -0700, Kelly O'Hair wrote: > Probably true that the test can be re-written, and you are right, a 500K > binary file is not something I want in the repository either. > Unfortunately, some engineers have the historical and bad habit of just > adding in files to workspaces and repositories without thinking about it > much, especially with tests. This habit must stop now, hopefully we can > catch this during the code review phase. And it seems Christopher Hegarty already took care of the rewrite: http://hg.openjdk.java.net/jdk7/jsn/jdk/rev/de7864ec310c I didn't see any code review on list, but basically the byte[] used for handling the request is now filled with data during runtime with a simple for loop stuffing the array in a static block in a simple pattern. Which eliminates the need for the test.txt file (removed in the patch above). Probably a nice idea to pull this into jdk6 also. Cheers, Mark
Re: ChunkedInputStream/test.txt
Mark Wielaard wrote: On Tue, 2008-09-23 at 09:46 -0700, Kelly O'Hair wrote: Probably true that the test can be re-written, and you are right, a 500K binary file is not something I want in the repository either. Unfortunately, some engineers have the historical and bad habit of just adding in files to workspaces and repositories without thinking about it much, especially with tests. This habit must stop now, hopefully we can catch this during the code review phase. And it seems Christopher Hegarty already took care of the rewrite: http://hg.openjdk.java.net/jdk7/jsn/jdk/rev/de7864ec310c I didn't see any code review on list, but basically the byte[] used for handling the request is now filled with data during runtime with a simple for loop stuffing the array in a static block in a simple pattern. Which eliminates the need for the test.txt file (removed in the patch above). Probably a nice idea to pull this into jdk6 also. Cheers, Yes, I've already asked Chris to fix this in OpenJDK 6 too. Regards, -Joe
