Re: ws2_32/tests: Add hostent struct tests for gethostbyname
Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at http://testbot.winehq.org/JobDetails.pl?Key=17727 Your paranoid android. === WXPPROSP3 (32 bit sock) === sock.c:509: Test failed: oob_server (1ac): unexpectedly at the OOB mark: 0 sock.c:519: Test failed: oob_server (1ac): unexpectedly at the OOB mark: 0
On Sat, Apr 7, 2012 at 4:54 PM, Marvin <testbot(a)testbot.winehq.org> wrote:
... === WXPPROSP3 (32 bit sock) === sock.c:509: Test failed: oob_server (1ac): unexpectedly at the OOB mark: 0 sock.c:519: Test failed: oob_server (1ac): unexpectedly at the OOB mark: 0
Is this an expected failure? Also, if you're unfamiliar with the gethostbyname() code then the place where you would want to change this is in WS_create_he(). If the behavior is opposite on OSes earlier than XP then you may wish to have it behave differently depending on the Windows version. Erich Hoover ehoover(a)mines.edu
On Sun, Apr 8, 2012 at 13:38, Erich E. Hoover <ehoover(a)mines.edu> wrote:
On Sat, Apr 7, 2012 at 4:54 PM, Marvin <testbot(a)testbot.winehq.org> wrote:
... === WXPPROSP3 (32 bit sock) === sock.c:509: Test failed: oob_server (1ac): unexpectedly at the OOB mark: 0 sock.c:519: Test failed: oob_server (1ac): unexpectedly at the OOB mark: 0
Is this an expected failure?
The OOB failure happens randomly because there are time related issues in the tests. We should only send one byte packages and wait for them, that would work in windows and wine. The main problem related to oob is that unix implementation differs from windows.
Also, if you're unfamiliar with the gethostbyname() code then the place where you would want to change this is in WS_create_he().
Thanks, I noticed that while studying the code. As soon as the tests get in I will start working in a solution, I know you have more experience with that code and I wouldn't oppose if you would like to work in it, just tell me.
If the behavior is opposite on OSes earlier than XP then you may wish to have it behave differently depending on the Windows version.
Once I was told that wine should behave only as newer systems to avoid if(os==x) dothis else dothat. I'll see what I can do.
Erich Hoover ehoover(a)mines.edu
Best wishes, Bruno
On Sun, Apr 8, 2012 at 10:54 AM, Bruno Jesus <00cpxxx(a)gmail.com> wrote:
...
If the behavior is opposite on OSes earlier than XP then you may wish to have it behave differently depending on the Windows version.
Once I was told that wine should behave only as newer systems to avoid if(os==x) dothis else dothat. I'll see what I can do.
This is certainly true with tests, in this particular case that might not be true. I would not be surprised to find that older applications rely on the "broken" buggy behavior and new applications reply on the "correct" buggy behavior. It may be that it's best to just go for the "correct" buggy behavior and then handle both situations if someone ever winds up reporting a bug for the "broken" behavior. Erich Hoover ehoover(a)mines.edu
On Sat, Apr 7, 2012 at 19:54, Marvin <testbot(a)testbot.winehq.org> wrote:
Hi, ... http://testbot.winehq.org/JobDetails.pl?Key=17727 ... Your paranoid android.
=== WXPPROSP3 (32 bit sock) === sock.c:509: Test failed: oob_server (1ac): unexpectedly at the OOB mark: 0 sock.c:519: Test failed: oob_server (1ac): unexpectedly at the OOB mark: 0
This is a know random failure. Not affected by my changes. Bruno
participants (3)
-
Bruno Jesus -
Erich E. Hoover -
Marvin