About the twill Web scripting language ()

Python | 5 Oct 2011 15:44

Want to crontribute

I'm wondering if the source code for twill is available in a public repository. I'm thinking I could contribute to the project, but not sure if it's the best thing just to submit patches. I'd be interested for instance in porting the application to Python 3 and adding a couple of new features for which I had to use workarounds.

If development has been frozen we could create a fork if the license allows, given the last release was performed almost 4 years ago, asit would be a pity to let such a good project die.

_______________________________________________
twill mailing list
twill@...
http://lists.idyll.org/listinfo/twill

headers

C. Titus Brown | 5 Oct 2011 15:49

Re: Want to crontribute

github.com/ctb/twill/.  Go to town :).

cheers,
--titus

On Wed, Oct 05, 2011 at 06:44:45AM -0700, Python wrote:
> I'm wondering if the source code for twill is available in a public repository. I'm thinking I could
contribute to the project, but not sure if it's the best thing just to submit patches. I'd be interested for
instance in porting the application to Python 3 and adding a couple of new features for which I had to use workarounds.
> 
> If development has been frozen we could create a fork if the license allows, given the last release was
performed almost 4 years ago, asit would be a pity to let such a good project die.
> _______________________________________________
> twill mailing list
> twill@...
> http://lists.idyll.org/listinfo/twill

--

-- 
C. Titus Brown, ctb@...

headers

Adam Victor Nazareth Brandizzi | 5 Oct 2011 18:22

Re: Want to crontribute

On Wed, Oct 5, 2011 at 10:49 AM, C. Titus Brown <ctb@...> wrote:
> github.com/ctb/twill/.  Go to town :).

How sad, I was really having fun wit the Darcs repo... :) No problem,
however, I bet it will be more accessible now.

--

-- 
Adam Victor Nazareth Brandizzi
http://brandizzi.googlepages.com/

headers

Jonathan Rogers | 5 Oct 2011 21:49

meta http-equiv refresh content syntax

I have removed Twill's ancient copy of mechanize and made it work with a
regular, current distribution. One test currently fails because of a
disagreement about how to interpret the content attribute of a meta
http-equiv="refresh" element:

<meta http-equiv="refresh" content="0;'URL=/login'">

The test script (test-equiv-refresh.twill) expects the above to cause a
load of "/login". It seems that's what the ancient internal version of
mechanize did, but a current version tries to load "URL=/login" which
results in a 404. This is exactly the same behavior I get in Firefox,
Chromium, and Epiphany.

If the test is correct and typical browsers are not, what's the real
authority? I'm having trouble even finding a formal definition of the
syntax the content attribute is supposed to have.

--

-- 
Jonathan Ross Rogers

headers

Stephen McInerney | 5 Oct 2011 21:55

Re: meta http-equiv refresh content syntax

Excellent work Jonathan.

Stephen

> Date: Wed, 5 Oct 2011 15:49:28 -0400
> From: jrogers-+z4E91SeU6Qg+LkckTzjXg@public.gmane.org
> To: twill-zFNWeKVCUJJD5a/XDn3G2A@public.gmane.org
> Subject: [twill] meta http-equiv refresh content syntax
>
> I have removed Twill's ancient copy of mechanize and made it work with a
> regular, current distribution. One test currently fails because of a
> disagreement about how to interpret the content attribute of a meta
> http-equiv="refresh" element:
>
> <meta http-equiv="refresh" content="0;'URL=/login'">
>
> The test script (test-equiv-refresh.twill) expects the above to cause a
> load of "/login". It seems that's what the ancient internal version of
> mechanize did, but a current version tries to load "URL=/login" which
> results in a 404. This is exactly the same behavior I get in Firefox,
> Chromium, and Epiphany.
>
> If the test is correct and typical browsers are not, what's the real
> authority? I'm having trouble even finding a formal definition of the
> syntax the content attribute is supposed to have.
>
> --
> Jonathan Ross Rogers
>
> _______________________________________________
> twill mailing list
> twill-zFNWeKVCUJI@public.gmane.orgyll.org
> http://lists.idyll.org/listinfo/twill

_______________________________________________
twill mailing list
twill@...
http://lists.idyll.org/listinfo/twill

headers

Jonathan Rogers | 6 Oct 2011 00:12

Re: meta http-equiv refresh content syntax

Leon Robbins wrote:
> Why is
>
> URL=/login
>
> bounded by single quotes? The examples on this Wikipedia page omit them:

That's exactly what I've been trying to figure out. The HTML in question
is in "tests/twilltestserver.py". TwillTest.test_refresh3() in that
module has the single-quoted value. Test script
"tests/test-equiv-refresh.twill" expects that value to cause Twill to
reload "/login". It seems that's what it did using the old, internal
version of mechanize.

Using a current version of mechanize, no reload occurs. Normal browsers
like Firefox and Chromium try to reload "/URL=/login" itself as a URL,
which is a non-existent page. I haven't been able to find a clear
specification for the format of the content attribute of the meta
element. It's doesn't seem to be specified in either HTML or HTTP and I
suspect it's just supposed to work how Netscape implemented it. Since
Firefox is a current Mozilla browser, it may be the most authoritative.

>
> http://en.wikipedia.org/wiki/Meta_refresh
>
> What happens if you leave them out too?
>

Without the single quotes, everything works as expected. That's what
TwillTest.test_refresh2() does.

--

-- 
Jonathan Ross Rogers

headers

Jonathan Rogers | 6 Oct 2011 00:18

Re: Syntax warning when installing twill

Christoph Zwerschke wrote:
> When installing twill, I get a syntax warning about the following
> statement in twill/other_packages/_mechanize_dist/_msiecookiejar.py:
> 
>                 assert (sig in ("HASH", "LEAK",
>                                 self.padding, "\x00\x00\x00\x00"),
>                         "unrecognized MSIE index.dat record: %s" %
>                         binary_to_str(sig))
> 
> The comma at the beginning of the first line should in fact be at the
> end of the second line:
> 
>                 assert sig in ("HASH", "LEAK",
>                                 self.padding, "\x00\x00\x00\x00"), (
>                         "unrecognized MSIE index.dat record: %s" %
>                         binary_to_str(sig))

That bug is in the ancient version of mechanize included with Twill. I
have ripped that out and made Twill work with a standard, current
version of mechanize. Please try out my branch and see if it works for you:

<URL:https://github.com/JonathanRRogers/twill>

--

-- 
Jonathan Ross Rogers

headers

Jonathan Rogers | 6 Oct 2011 00:21

Re: Want to crontribute

C. Titus Brown wrote:
> github.com/ctb/twill/.  Go to town :).
> 

Please have a look at my branch which addresses several issues I've had
using it with current versions of mechanize and wsgi_intercept:

<URL:https://github.com/JonathanRRogers/twill>
--

-- 
Jonathan Ross Rogers

headers

Adam Victor Nazareth Brandizzi | 18 Oct 2011 16:45

[RFC] Use of non-string values with ${} notation

Hello, all!

consider the following twill script:

    run 'value = 123'
    go http://www.google.com
    fv 1 q ${value}
    submit
    code 200

If I run this script, I get this error:

    AT LINE: test.twill:3
    ** UNHANDLED EXCEPTION: cannot concatenate 'str' and 'int' objects

The solution is simple: just call str() on the value:

    fv 1 q ${str(value)}

(I could quote the value when setting it, too, but It is not always
possible or desirable)

However, I get into this situation a lot of times and find it pretty
annoying. I wonder: does anyone here get in this situation too? I have
patched twill to stringify all values; would this change be desirable
for everyone?

Thanks in advance!

--

-- 
Adam Victor Nazareth Brandizzi
http://brandizzi.googlepages.com/

Mon	Tue	Wed	Thu	Fri	Sat	Sun

					1	2
3	4	5	6	7	8	9
10	11	12	13	14	15	16
17	18	19	20	21	22	23
24	25	26	27	28	29	30
31

1	November 2012
3	October 2012
1	April 2012
4	March 2012
3	February 2012
9	October 2011
6	September 2011
1	July 2011
1	June 2011
1	April 2011
14	February 2011
6	January 2011
4	November 2010
6	October 2010
1	August 2010
8	July 2010
5	June 2010
13	May 2010
3	April 2010
16	March 2010
19	February 2010
9	January 2010
7	December 2009
6	November 2009
6	October 2009
9	September 2009
3	July 2009
6	June 2009
6	May 2009
18	April 2009
14	March 2009
8	February 2009
4	January 2009
8	December 2008
2	November 2008
11	October 2008
13	September 2008
4	August 2008
19	July 2008
10	April 2008
5	March 2008
4	February 2008
26	January 2008
25	December 2007
12	November 2007
16	October 2007
22	September 2007
23	August 2007
28	July 2007
13	June 2007
4	May 2007
13	April 2007
26	March 2007
9	February 2007
24	January 2007
10	December 2006
29	November 2006
36	October 2006
16	September 2006
41	August 2006
34	July 2006
18	June 2006
46	May 2006
32	April 2006
123	March 2006
49	February 2006
17	January 2006
42	December 2005
90	November 2005
8	October 2005