[forge-dev] Running the build on Windows

Sat Aug 16 02:51:59 EDT 2014

Hi everybody,

I fixed most of the tests. However, since a couple of days (actually nights
;)) I am struggling with a few tests. I have attached the JUnit reports. I
would appreciate any help from someone with more knowledge in regular
expressions and finite automata (like the author of
http://ocpsoft.org/tutorials/regular-expressions/java-visual-regex-tester/
;)).

My work so far is pushed here:
https://github.com/ivannov/core/tree/FORGE-1966

I would appreciate even some hints on how to debug and find the issue, as I
have run out of ideas.

Thanks a lot!
Ivan

On Sun, Aug 10, 2014 at 11:09 PM, Ivan St. Ivanov <ivan.st.ivanov at gmail.com>
wrote:

> Hi again,
>
> As promised, this is the output of the test that failed. As I already
> wrote, it actually broke with a timeout, as you might guess from the output.
>
> I am working now on fixing the text addon. However, it takes some time, as
> I am not experienced in debugging state machines that calculate their state
> based on pattern matching ;)
>
> Cheers,
> Ivan
>
>
> On Fri, Aug 8, 2014 at 4:03 PM, George Gastaldi <ggastald at redhat.com>
> wrote:
>
>> Hey Ivan,
>>
>> Seeing that the Windows build in our CI always fail because of this I
>> believe that fixing the text addon may be a better choice.
>>
>> Thanks,
>>
>> George
>>
>> Em 08/08/2014, às 09:12, "Ivan St. Ivanov" <ivan.st.ivanov at gmail.com>
>> escreveu:
>>
>> Hi everybody,
>>
>> Sorry for the whole week without news from me, I was in vacation and
>> didn't want to touch a computer ;)
>>
>> So, I updated the JDK on my windows box to the latest version. And now
>> the JVM error while running the javaee build is gone. However, it failed
>> with a timeout in calling a javassist proxy. Usually on Windows tests run
>> extremely slowly, so I believe that the test would run fine if it was a bit
>> faster. I will send you later (today or tomorrow) the exact stack trace.
>>
>> BTW, which issue is more important? The text addon fix or the JUnit test
>> command?
>>
>> Regards,
>> Ivan
>>
>>
>> On Thu, Jul 31, 2014 at 7:36 PM, George Gastaldi <ggastald at redhat.com>
>> wrote:
>>
>>> Right, sounds good.
>>>
>>> Thanks for the help, you are awesome! :)
>>>
>>> George
>>>
>>> Em 31/07/2014, às 12:54, "Ivan St. Ivanov" <ivan.st.ivanov at gmail.com>
>>> escreveu:
>>>
>>> George, I didn't mean tests at all here.
>>>
>>> I meant that whenever you want to use the text addon to parse something
>>> for you, you will have troubles on Windows, because the RegEx patterns in
>>> the productive code do not work on Windows. And the tests are there to
>>> prove that, not to be ignored by us.
>>>
>>> I will create the JIRA issue and will assign it to me. I guess it has
>>> higher prio than JUnit command.
>>>
>>>
>>> On Thu, Jul 31, 2014 at 4:32 PM, George Gastaldi <ggastald at redhat.com>
>>> wrote:
>>>
>>>> Hey Ivan,
>>>>
>>>> When you refer to the "whole text addon is broken" you mean the Scanner
>>>> tests or the Inflector too? I ask this because the Inflector tests pass in
>>>> my local box.
>>>>
>>>> Em 31/07/2014, às 09:55, "Ivan St. Ivanov" <ivan.st.ivanov at gmail.com>
>>>> escreveu:
>>>>
>>>> George, I think that it is not the test that do not work on Windows,
>>>> but the whole text addon is broken. I mean, the patterns that rely on \n
>>>> are there. I fixed easily the YAML patterns, but I stuck in the Groovy ones
>>>> - too few knowledge in RegExp :)
>>>>
>>>>
>>>> On Thu, Jul 31, 2014 at 3:46 PM, George Gastaldi <ggastald at redhat.com>
>>>> wrote:
>>>>
>>>>> Hi Ivan,
>>>>>
>>>>> That would be good if the fix is trivial. Otherwise we can add an
>>>>> Assume.assumeFalse(OperatingSystemUtils.isWindows()); in a @Before method
>>>>> in the scanner tests.
>>>>>
>>>>>
>>>>> Em 31/07/2014, às 09:32, "Ivan St. Ivanov" <ivan.st.ivanov at gmail.com>
>>>>> escreveu:
>>>>>
>>>>> Hey George,
>>>>>
>>>>> Thanks!
>>>>>
>>>>> What do you think, should we fix the text issues on Windows? I think
>>>>> yes. If you agree with me, I will at least create a JIRA for that.
>>>>>
>>>>> Regards,
>>>>> Ivan
>>>>>
>>>>>
>>>>> On Thu, Jul 31, 2014 at 3:23 PM, George Gastaldi <ggastald at redhat.com>
>>>>> wrote:
>>>>>
>>>>>> Hey Ivan,
>>>>>>
>>>>>> Good catch! You're the man!
>>>>>>
>>>>>> In my previous attempt to find out what the problem was, I had to
>>>>>> ignore the text addon tests on Windows to proceed.
>>>>>>
>>>>>> The JVM crash is reported in JIRA also, I think Antonio Goncalves
>>>>>> experienced something like that.
>>>>>>
>>>>>> Let me know if you find somehing else.
>>>>>>
>>>>>> Best Regards,
>>>>>>
>>>>>> George
>>>>>>
>>>>>> Em 31/07/2014, às 08:28, "Ivan St. Ivanov" <ivan.st.ivanov at gmail.com>
>>>>>> escreveu:
>>>>>>
>>>>>> Hey Daniel,
>>>>>>
>>>>>> The idea of this effort is just to make sure Forge tests run locally
>>>>>> on Windows. And then to investigate why the central build fails. I.e. first
>>>>>> clean up your own garden and only then look in the public park ;)
>>>>>>
>>>>>> If you have any special issues with Aesh, please share them (probably
>>>>>> in another thread). I also had some, but I thought they were resolved. The
>>>>>> thing is that I don't use Windows at home any more (switched to Fedora), so
>>>>>> I can't catch problems with the rate I used to do before ;)
>>>>>>
>>>>>> Cheers,
>>>>>> Ivan
>>>>>>
>>>>>>
>>>>>> On Thu, Jul 31, 2014 at 12:54 PM, Daniel Cunha <danielsoro at gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>>> Hi Ivan,
>>>>>>>
>>>>>>> I don't know if aesh work ok on windows. In my last test I had many
>>>>>>> problem and that can be the answer about forge don't run ok on windows.
>>>>>>>
>>>>>>>
>>>>>>> On Thu, Jul 31, 2014 at 4:45 AM, Ivan St. Ivanov <
>>>>>>> ivan.st.ivanov at gmail.com> wrote:
>>>>>>>
>>>>>>>> Hey everybody,
>>>>>>>>
>>>>>>>> As we discussed yesterday, a promised to run the build locally on
>>>>>>>> my Windows computer to see if we have any issues there, fix them and only
>>>>>>>> then address the failing builds on the build server.
>>>>>>>>
>>>>>>>> I did that, and I have the following observations:
>>>>>>>>
>>>>>>>> 1) The text addon tests failed, mostly because the patterns that
>>>>>>>> recognize the different language elements for YAML, Groovy, JavaScript and
>>>>>>>> HTML are built with the assumption that the end of line character is \n.
>>>>>>>> You know, on Windows it is \r\n. So there were some tests failures. I
>>>>>>>> decided to fix them and only then to continue. But after it took me more
>>>>>>>> than two hours debugging and trying without much progress (fixed only YAML
>>>>>>>> and most of the Groovy) I gave up. I then cloned the repo with git's
>>>>>>>> autocrlf set to false and the tests passed. Nevertheless, we should fix
>>>>>>>> these I guess, if we want our text addon to work on Windows
>>>>>>>>
>>>>>>>> 2) The JavaEE addon failed with a JVM crash (I attached the
>>>>>>>> generated dump files from the three runs that I did). I tried to run it
>>>>>>>> both as part of the build of the whole Forge core, as well as individually.
>>>>>>>> In both cases it failed. I will install a newer version of the JDK and will
>>>>>>>> try again.
>>>>>>>>
>>>>>>>> That is for the moment from my side. Unfortunately I spent all the
>>>>>>>> time debugging the text addon, so didn't spend any effort on the JUnit
>>>>>>>> command and no email for utility methods so far from me.
>>>>>>>>
>>>>>>>> Cheers,
>>>>>>>> Ivan
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> forge-dev mailing list
>>>>>>>> forge-dev at lists.jboss.org
>>>>>>>> https://lists.jboss.org/mailman/listinfo/forge-dev
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Daniel Cunha (soro) <http://www.cejug.net>
>>>>>>> Blog: http://www.danielsoro.com.br
>>>>>>> Twitter: https://twitter.com/dvlc_
>>>>>>> GitHub: https://github.com/danielsoro
>>>>>>> LinkedIn:  http://www.linkedin.com/in/danielvlcunha
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> forge-dev mailing list
>>>>>>> forge-dev at lists.jboss.org
>>>>>>> https://lists.jboss.org/mailman/listinfo/forge-dev
>>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> forge-dev mailing list
>>>>>> forge-dev at lists.jboss.org
>>>>>> https://lists.jboss.org/mailman/listinfo/forge-dev
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> forge-dev mailing list
>>>>>> forge-dev at lists.jboss.org
>>>>>> https://lists.jboss.org/mailman/listinfo/forge-dev
>>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> forge-dev mailing list
>>>>> forge-dev at lists.jboss.org
>>>>> https://lists.jboss.org/mailman/listinfo/forge-dev
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> forge-dev mailing list
>>>>> forge-dev at lists.jboss.org
>>>>> https://lists.jboss.org/mailman/listinfo/forge-dev
>>>>>
>>>>
>>>> _______________________________________________
>>>> forge-dev mailing list
>>>> forge-dev at lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/forge-dev
>>>>
>>>>
>>>> _______________________________________________
>>>> forge-dev mailing list
>>>> forge-dev at lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/forge-dev
>>>>
>>>
>>> _______________________________________________
>>> forge-dev mailing list
>>> forge-dev at lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/forge-dev
>>>
>>>
>>> _______________________________________________
>>> forge-dev mailing list
>>> forge-dev at lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/forge-dev
>>>
>>
>> _______________________________________________
>> forge-dev mailing list
>> forge-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/forge-dev
>>
>>
>> _______________________________________________
>> forge-dev mailing list
>> forge-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/forge-dev
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/forge-dev/attachments/20140816/c9bcd4b7/attachment-0001.html 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: failing-tests.zip
Type: application/zip
Size: 2877 bytes
Desc: not available
Url : http://lists.jboss.org/pipermail/forge-dev/attachments/20140816/c9bcd4b7/attachment-0001.zip 