Re: [Jython-dev] test_httpservers

[Jim B. <jim...@py...>]

Jeff,

I removed the modified copy of test_httpservers.py in favor of the standard
version in lib-python/2.7/test. My experience is that it works fine on
Windows 8.1 and it does not hang the regrtest:

c:\jythondev\jython27>dist\bin\jython.bat dist\Lib\test\test_httpservers.py
test_http_0_9 (__main__.BaseHTTPRequestHandlerTestCase) ... ok
test_http_1_0 (__main__.BaseHTTPRequestHandlerTestCase) ... ok
test_http_1_1 (__main__.BaseHTTPRequestHandlerTestCase) ... ok
test_request_length (__main__.BaseHTTPRequestHandlerTestCase) ... ok
test_with_continue_1_0 (__main__.BaseHTTPRequestHandlerTestCase) ... ok
test_query_arguments (__main__.SimpleHTTPRequestHandlerTestCase) ... ok
test_start_with_double_slash (__main__.SimpleHTTPRequestHandlerTestCase)
... ok
test_command (__main__.BaseHTTPServerTestCase) ... ok
test_handler (__main__.BaseHTTPServerTestCase) ... ok
test_head_keep_alive (__main__.BaseHTTPServerTestCase) ... ok
test_header_close (__main__.BaseHTTPServerTestCase) ... ok
test_internal_key_error (__main__.BaseHTTPServerTestCase) ... ok
test_request_line_trimming (__main__.BaseHTTPServerTestCase) ... ok
test_return_custom_status (__main__.BaseHTTPServerTestCase) ... ok
test_return_header_keep_alive (__main__.BaseHTTPServerTestCase) ... ok
test_send_blank (__main__.BaseHTTPServerTestCase) ... ok
test_version_bogus (__main__.BaseHTTPServerTestCase) ... ok
test_version_digits (__main__.BaseHTTPServerTestCase) ... ok
test_version_invalid (__main__.BaseHTTPServerTestCase) ... ok
test_version_none (__main__.BaseHTTPServerTestCase) ... ok
test_version_none_get (__main__.BaseHTTPServerTestCase) ... ok
test_get (__main__.SimpleHTTPServerTestCase) ... ok
test_head (__main__.SimpleHTTPServerTestCase) ... ok
test_invalid_requests (__main__.SimpleHTTPServerTestCase) ... ok
test_authorization (__main__.CGIHTTPServerTestCase) ... ok
test_headers_and_content (__main__.CGIHTTPServerTestCase) ... ok
test_invaliduri (__main__.CGIHTTPServerTestCase) ... ok
test_no_leading_slash (__main__.CGIHTTPServerTestCase) ... ok
test_os_environ_is_not_altered (__main__.CGIHTTPServerTestCase) ... ok
test_post (__main__.CGIHTTPServerTestCase) ... ok
test_url_collapse_path (__main__.CGIHTTPServerTestCase) ... ok

----------------------------------------------------------------------
Ran 31 tests in 32.974s

OK

(Same results, as would be expected, using the alternative command of start
dist\bin\jython -m test.test_httpservers, per what you reported in
http://bugs.jython.org/issue2109)

It would be worthwhile testing this on Windows 7 as well, so I will work on
getting a personal setup of Windows 7 as a VM later this week, just to see
if the delta is there or perhaps more likely in different network
configuration.

There is an exclude option (-x) for the regrtest, but this option does not
currently work:

$ dist/bin/jython dist/Lib/test/regrtest.py -x test_httpservers
Traceback (most recent call last):
  File "dist/Lib/test/regrtest.py", line 1449, in <module>
    main()
  File "dist/Lib/test/regrtest.py", line 368, in main
    nottests[:0] = args
TypeError: 'set' object does not support item assignment

I have opened a corresponding bug: http://bugs.jython.org/issue2150

- Jim


On Mon, May 19, 2014 at 3:43 AM, Jeff Allen <ja...@fa...> wrote:

> I notice that in recent changes our custom Lib/test/test_httpservers.py
> has been deleted. We need this to deal with issue 2109
> (http://bugs.jython.org/issue2109) until CPython catches up, and without
> it regrtest hangs on my machine.
>
> I can restore the file, but what's puzzling me is that, as we deleted
> it, we presumably think we don't need the skips it contained. But
> without them I get these failures:
>
> test_authorization (__main__.CGIHTTPServerTestCase) ... FAIL
> test_headers_and_content (__main__.CGIHTTPServerTestCase) ... FAIL
>
> test_no_leading_slash (__main__.CGIHTTPServerTestCase) ... FAIL
> test_os_environ_is_not_altered (__main__.CGIHTTPServerTestCase) ... FAIL
> test_post (__main__.CGIHTTPServerTestCase) ... FAIL
>
> Was this a choice to make it fail noisily, as with many other tests
> recently re-enabled? Or do these pass for others?
>
> For speed, and minimal change, I'm going to restore this with the #2109
> changes, but without the skips.
>
> -- Jeff Allen
>
>
> ------------------------------------------------------------------------------
> "Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
> Instantly run your Selenium tests across 300+ browser/OS combos.
> Get unparalleled scalability from the best Selenium testing platform
> available
> Simple to use. Nothing to install. Get started now for free."
> http://p.sf.net/sfu/SauceLabs
> _______________________________________________
> Jython-dev mailing list
> Jyt...@li...
> https://lists.sourceforge.net/lists/listinfo/jython-dev
>