You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Michael Haggerty <mh...@alum.mit.edu> on 2009/07/27 18:21:17 UTC
[PATCH v2] Change test suite to read file properties in XML format
James Abbatiello wrote:
> On Mon, Jul 27, 2009 at 11:15 AM, Michael Haggerty<mh...@alum.mit.edu> wrote:
>> [...]
>> But even though SVN outputs '\r\n' on Windows, I still don't understand
>> how the '\r\n' characters get through in the test suite.
>
> This had me stumped for a while too.
> C:>svn --version
> svn, version 1.6.3 (r38063)
> compiled Jun 18 2009, 12:57:17
> ...
>
> C:>svn proplist --verbose --xml svntest
> <?xml version="1.0"?>
> <properties>
> <target
> path="svntest">
> <property
> name="svn:ignore">*.pyc
> *.o
> *~
> .*~
>
> </property>
> </target>
> </properties>
>
> The '\r\n' sequences are being folded to '\n'. But then the XML
> parser turns '
\n' into '\r\n' in the final output.
Yuck. I wonder whether the output contains ' \n' or ' \r\n'
(i.e., does the output contain one or two extraneous CR characters?
In any case, I'm defeated. Attached is a patch like the previous one,
except that it also smashes all EOL combinations into '\n'. Feedback is
welcome, especially from Windows users.
Michael
[[[
Read svn properties in XML rather than text format in test suite.
This makes the routine shorter and more robust to strange property
values and to format changes in the text output format.
* subversion/tests/cmdline/svntest/tree.py
(get_props): Read svn properties via "svn proplist -v" using the
"--xml" option.
]]]
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2376090
Re: [PATCH v2] Change test suite to read file properties in XML
format
Posted by Julian Foad <ju...@btopenworld.com>.
On Mon, 2009-07-27 at 20:21 +0200, Michael Haggerty wrote:
> James Abbatiello wrote:
> > On Mon, Jul 27, 2009 at 11:15 AM, Michael Haggerty<mh...@alum.mit.edu> wrote:
> >> [...]
> >> But even though SVN outputs '\r\n' on Windows, I still don't understand
> >> how the '\r\n' characters get through in the test suite.
> >
> > This had me stumped for a while too.
> > C:>svn --version
> > svn, version 1.6.3 (r38063)
> > compiled Jun 18 2009, 12:57:17
> > ...
> >
> > C:>svn proplist --verbose --xml svntest
> > <?xml version="1.0"?>
> > <properties>
> > <target
> > path="svntest">
> > <property
> > name="svn:ignore">*.pyc
> > *.o
> > *~
> > .*~
> >
> > </property>
> > </target>
> > </properties>
> >
> > The '\r\n' sequences are being folded to '\n'. But then the XML
> > parser turns '
\n' into '\r\n' in the final output.
>
> Yuck. I wonder whether the output contains ' \n' or ' \r\n'
> (i.e., does the output contain one or two extraneous CR characters?
>
> In any case, I'm defeated. Attached is a patch like the previous one,
> except that it also smashes all EOL combinations into '\n'. Feedback is
> welcome, especially from Windows users.
>
> Michael
>
> [[[
> Read svn properties in XML rather than text format in test suite.
>
> This makes the routine shorter and more robust to strange property
> values and to format changes in the text output format.
Please state the main goal of making this part of the test suite
compatible with the output of pre-1.6 svn, and the cvs2svn project's
interest in this. And this change doesn't really make it shorter if you
ignore the blank lines.
Thanks.
- Julian
> * subversion/tests/cmdline/svntest/tree.py
> (get_props): Read svn properties via "svn proplist -v" using the
> "--xml" option.
> ]]]
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2376142
Re: [PATCH v3] Change test suite to read file properties in XML
format
Posted by James Abbatiello <ab...@gmail.com>.
On Tue, Jul 28, 2009 at 4:47 AM, Michael Haggerty<mh...@alum.mit.edu> wrote:
> This time I hand-tested against the output that you listed in your
> previous email, and it seems to work. Third time's the charm?
Success! It works for me.
--
James Abbatiello
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2376398
Re: [PATCH v3] Change test suite to read file properties in XML
format
Posted by Julian Foad <ju...@btopenworld.com>.
Michael Haggerty wrote:
> [[[
> Read svn properties in XML rather than text format in test suite.
>
> This makes the routine more robust to strange property values. It also
> makes it immune to changes in the text output format, as happened
> between svn 1.5 and 1.6. (This is useful for the cvs2svn project,
> which uses the svntest infrastructure but doesn't want to care what
> version of svn the user has installed.)
>
> * subversion/tests/cmdline/svntest/tree.py
> (get_props): Read svn properties via "svn proplist -v" using the
> "--xml" option.
> ]]]
Nice log message. Thanks.
I have confirmed this version (patch v3) works in a trunk build (r38481)
on my Linux system. +1 to commit.
- Julian
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2376259
[PATCH v3] Change test suite to read file properties in XML format
Posted by Michael Haggerty <mh...@alum.mit.edu>.
Thanks, everybody, for the feedback.
James Abbatiello wrote:
> Sorry, it still doesn't work.
>>>> svntest.tree.get_props(["svntest"])
> {u'svntest': {u'svn:ignore': u'*.pyc\r\n*.o\r\n*~\r\n.*~\r\n\r\n'}}
>
> You do the folding at the top of the function. As you noted,
> wait_on_pipe() already does basically the same thing so that by the
> time the strings get to you they end with " \n". And rstrip isn't
> going to match against encoded XML entities. You're either going to
> have to match against " \n" at the end of the line (ick) or wait
> until after the XML parser has had a shot at it and adjust the line
> endings then.
Yes, of course you are right again.
This time I hand-tested against the output that you listed in your
previous email, and it seems to work. Third time's the charm?
Julian Foad wrote:
> Please state the main goal of making this part of the test suite
> compatible with the output of pre-1.6 svn, and the cvs2svn project's
> interest in this. And this change doesn't really make it shorter if you
> ignore the blank lines.
Suggestion incorporated.
Hyrum K. Wright wrote:
>> I don't think there is anything precluding the use of "svn proplist
>> > --xml" for svntest's property-checking functions.
>
> Julian Foad wrote
> Nor do I, just as long as we make sure to still test the "standard"
> output somewhere in the testsuite.
I just intentionally broke the code for "svn proplist -v" (without
--xml) and several test cases in prop_test.py broke. So there is at
least some level of testing aside from tree.get_props().
David Glasser wrote:
> The Python versions we support ship with an XML parser, right?
The xml.dom.minidom parser (which I'm using) was added in Python 2.0.
Michael
[[[
Read svn properties in XML rather than text format in test suite.
This makes the routine more robust to strange property values. It also
makes it immune to changes in the text output format, as happened
between svn 1.5 and 1.6. (This is useful for the cvs2svn project,
which uses the svntest infrastructure but doesn't want to care what
version of svn the user has installed.)
* subversion/tests/cmdline/svntest/tree.py
(get_props): Read svn properties via "svn proplist -v" using the
"--xml" option.
]]]
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2376241
Re: [PATCH v2] Change test suite to read file properties in XML
format
Posted by James Abbatiello <ab...@gmail.com>.
On Mon, Jul 27, 2009 at 2:21 PM, Michael Haggerty<mh...@alum.mit.edu> wrote:
> James Abbatiello wrote:
>> The '\r\n' sequences are being folded to '\n'. But then the XML
>> parser turns '
\n' into '\r\n' in the final output.
>
> Yuck. I wonder whether the output contains ' \n' or ' \r\n'
> (i.e., does the output contain one or two extraneous CR characters?
>>> stdin, stdout, stderr, kid = svntest.main.open_pipe(["svn", "proplist", "--verbose", "--xml", "svntest"])
>>> stdout.read()
'<?xml version="1.0"?>\r\n<properties>\r\n<target\r\n
path="svntest">\r\n<property\r\n
name="svn:ignore">*.pyc \r\n*.o \r\n*~ \r\n.*~ \r\n \r\n</property>\r\n</target>\r\n</properties>\r\n'
> In any case, I'm defeated. Attached is a patch like the previous one,
> except that it also smashes all EOL combinations into '\n'. Feedback is
> welcome, especially from Windows users.
Sorry, it still doesn't work.
>>> svntest.tree.get_props(["svntest"])
{u'svntest': {u'svn:ignore': u'*.pyc\r\n*.o\r\n*~\r\n.*~\r\n\r\n'}}
You do the folding at the top of the function. As you noted,
wait_on_pipe() already does basically the same thing so that by the
time the strings get to you they end with " \n". And rstrip isn't
going to match against encoded XML entities. You're either going to
have to match against " \n" at the end of the line (ick) or wait
until after the XML parser has had a shot at it and adjust the line
endings then.
--
James Abbatiello
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2376110