Hi Marcel,
Actually, I'm not seeing the problem on my system (Linux). The tests in NetworkTests amd the WebClient tests are green, regardless of whether the "Enable IPv6" preference is true or false.
I do see intermittent error failures in WebClientServerTest. Usually all tests pass, but I have seen failures in #testStreaming and #testPersistentAutRedirectSession so far. It is very intermittent, but I know that the new resolver primitives introduce some new possibilities for concurrency issue, so there may be some problems in that regard.
But the failures you see are presumably consistent, and are much different from what I am seeing on my system.
I have to suspect platform differences at this point. I think that you are using OS X (is that right?) and I expect that the network resolver may behave quite differently on different platforms, so we'll need to collect a few more data points to see where the issues are.
Can others please try running WebClientServerTest both with "Enable IPv6" true and with it false, and report back on whether the tests pass on your Unix/Linux/Windows/OS X system?
Thanks, Dave
On Wed, Apr 22, 2020 at 11:31:30AM -0400, David T. Lewis wrote:
Oh my, that's going to take some tidying up.
And I expect that in order to get full coverage, we will need to arrange for tests that switch the preference to cover both the old and the new network settings when we run the full test suite.
I can't look at it right now but will do so soon-ish.
Dave
On Wed, Apr 22, 2020 at 03:56:37PM +0200, Marcel Taeumel wrote:
Hi Dave! :-)
Now we have to fix those tests that suddenly got nervous and fail or time out -- at least on my machine:
YELLOW - FAIL
WebClientServerTest>>#testBasicAuth WebClientServerTest>>#testBerarerAuth WebClientServerTest>>#testContentTypeCharsetEncoding WebClientServerTest>>#testDigestAuth WebClientServerTest>>#testHtmlSubmit WebClientServerTest>>#testHttpDelete WebClientServerTest>>#testInvalidCookies WebClientServerTest>>#testNestedAction
RED - ERROR
ArbitraryObjectSocketTestCase>>#testBasics SocketTest>>#testClientConnect SocketTest>>#testDataReceive SocketTest>>#testDataSending SocketTest>>#testLocalAddress SocketTest>>#testLocalPort SocketTest>>#testPeerName SocketTest>>#testReceiveTimeout SocketTest>>#testRemoteAddress SocketTest>>#testRemotePort SocketTest>>#testSendTimeout SocketTest>>#testServerAccept SocketTest>>#testStringFromAddress SocketTest>>#testUDP SqueakSSLTest>>#testGooglePopStream StringSocketTestCase>>#testBasics StringSocketTestCase>>#testBogusInput1 StringSocketTestCase>>#testBogusInput2 WebClientServerTest>>#testListenOnInterface
;o)
Best, Marcel Am 07.04.2020 17:12:43 schrieb commits@source.squeak.org commits@source.squeak.org: David T. Lewis uploaded a new version of Network to project The Trunk: http://source.squeak.org/trunk/Network-dtl.240.mcz
==================== Summary ====================
Name: Network-dtl.240 Author: dtl Time: 6 April 2020, 10:33:07.751386 pm UUID: a8f00e08-7398-4d4a-8cd2-09cedcada729 Ancestors: Network-eem.239
This update enables the Squeak new networking and IPv6 support by default. The update is in the package postscript.
IPv6 network support for the VM (SocketPlugin) and image were implemented by Ian Piumarta circa 2007. This included IPv6 support and a more general image-side representation of network addressing. In 2012, this work was merged into Squeak trunk, with a preference to switch between the old networking support or the new networking support with IPv6. At that time, there were concerns with performance of NetNameResolver on some Windows platforms, and for this reason the "Enable IPv6 and new network support" preference was left disabled for official Squeak distributions.
It now 2020, and operating systems have improved over the years, so it is time to enable the new networking preference by default.
=============== Diff against Network-eem.239 ===============
Item was changed:
- (PackageInfo named: 'Network') postscript: '"Enable IPv6 and new network support by default"
- NetNameResolver enableIPv6: true
- '!
- (PackageInfo named: 'Network') postscript: '"Rebuild CSNonAtom."
- MailAddressTokenizer cleanUp: false'!