~launchpad-pqm/launchpad/devel

Displaying Paragraphs of Text with ZPT

======================================

To display paragraphs of text in HTML, use fmt:text-to-html. For details,

see <https://launchpad.canonical.com/DisplayingParagraphsOfText>.

Basics

------

    >>> from lp.testing import test_tales

    >>> text = ('This is a paragraph.\n'

    ...         '\n'

    ...         'This is another paragraph.')

    >>> test_tales('foo/fmt:text-to-html', foo=text)

    '<p>This is a paragraph.</p>\n<p>This is another paragraph.</p>'

    >>> text = ('This is a line.\n'

    ...         'This is another line.')

    >>> test_tales('foo/fmt:text-to-html', foo=text)

    '<p>This is a line.<br />\nThis is another line.</p>'

    >>> text = (

    ...     'This is a paragraph that has been hard-wrapped by an e-mail'

    ...     ' application.\n'

    ...     'We used to handle this specially, but we no longer do because it'

    ...     ' was disturbing\n'

    ...     'the display of backtraces. Expected results:\n'

    ...     '* joy\n'

    ...     '* elation'

    ...     )

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p>This is a paragraph that has been hard-wrapped by an e-mail

    application.<br />

    We used to handle this specially, but we no longer do because it was

    disturbing<br />

    the display of backtraces. Expected results:<br />

    * joy<br />

    * elation</p>

    >>> text = (

    ...     " 1. Here's an example\n"

    ...     " 2. where a list is followed by a paragraph.\n"

    ...     "   Leading spaces in a line or paragraph are presented, which "

    ...     "means converting them to  . Trailing spaces are passed "

    ...     "through as-is, which means browsers will ignore them, but "

    ...     "that's fine, they're not important anyway.\n"

    ...     )

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p>&nbsp;1. Here's an example<br />

    &nbsp;2. where a list is followed by a paragraph.<br />

    &nbsp;&nbsp;&nbsp;Leading spaces in a line or paragraph are presented, which means converting them to &amp;nbsp;. Trailing spaces are passed through as-is, which means browsers will ignore them, but that's fine, they're not important anyway.</p>

    >>> text = (

    ...     'This is a little paragraph all by itself. How cute!'

    ...     )

    >>> test_tales('foo/fmt:text-to-html', foo=text)

    '<p>This is a little paragraph all by itself. How cute!</p>'

    >>> text = (

    ...     'Here are two paragraphs with lots of whitespace between them.\n'

    ...     '\n'

    ...     '\n'

    ...     '\n'

    ...     '\n'

    ...     'But they\'re still just two paragraphs.')

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p>Here are two paragraphs with lots of whitespace between them.</p>

    <p>But they're still just two paragraphs.</p>

If a line begins with whitespace, it will not be merged with the

previous line.  This aids in the display of code samples:

    >>> text = (

    ...     'This is a code sample written in Python.\n'

    ...     '    def messageCount(self):\n'

    ...     '        """See IRosettaStats."""\n'

    ...     '        return self.potemplate.messageCount()\n'

    ...     '\n'

    ...     '    def currentCount(self, language=None):\n'

    ...     '        """See IRosettaStats."""\n'

    ...     '        return self.currentCount\n')

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p>This is a code sample written in Python.<br />

    &nbsp;&nbsp;&nbsp;&nbsp;def messageCount(self):<br />

    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;"""See IRosettaStats."""<br />

    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;return self.potemplate<wbr></wbr>.messageCount(<wbr></wbr>)</p>

    <p>&nbsp;&nbsp;&nbsp;&nbsp;def currentCount(self, language=None):<br />

    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;"""See IRosettaStats."""<br />

    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;return self.currentCount</p>

Testing a bunch of URL links.

    >>> text = (

    ...     'https://launchpad.net/ is the new Launchpad site\n'

    ...     'http://example.com/something?foo=bar&hum=baz\n'

    ...     'You can check the PPC md5sums at '

    ...     'ftp://ftp.ubuntu.com/ubuntu/dists/breezy/main/installer-powerpc'

    ...     '/current/images/MD5SUMS\n'

    ...     'irc://irc.freenode.net/#launchpad\n'

    ...     '\n'

    ...     'I have a Jabber account (jabber:foo@jabber.example.com)\n'

    ...     'Foo Bar <mailto:foo.bar@example.net>')

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a rel="nofollow" href="https://launchpad.net/">https:/<wbr></wbr>/launchpad.<wbr></wbr>net/</a> is the new Launchpad site<br />

    <a rel="nofollow" href="http://example.com/something?foo=bar&amp;hum=baz">http://<wbr></wbr>example.<wbr></wbr>com/something?<wbr></wbr>foo=bar&amp;<wbr></wbr>hum=baz</a><br />

    You can check the PPC md5sums at <a rel="nofollow" href="ftp://ftp.ubuntu.com/ubuntu/dists/breezy/main/installer-powerpc/current/images/MD5SUMS">ftp://ftp.<wbr></wbr>ubuntu.<wbr></wbr>com/ubuntu/<wbr></wbr>dists/breezy/<wbr></wbr>main/installer-<wbr></wbr>powerpc/<wbr></wbr>current/<wbr></wbr>images/<wbr></wbr>MD5SUMS</a><br />

    <a rel="nofollow" href="irc://irc.freenode.net/#launchpad">irc://irc.<wbr></wbr>freenode.<wbr></wbr>net/#launchpad</a></p>

    <p>I have a Jabber account (<a rel="nofollow" href="jabber:foo@jabber.example.com">jabber:<wbr></wbr>foo@jabber.<wbr></wbr>example.<wbr></wbr>com</a>)<br />

    Foo Bar &lt;<a rel="nofollow" href="mailto:foo.bar@example.net">mailto:<wbr></wbr>foo.bar@<wbr></wbr>example.<wbr></wbr>net</a>&gt;</p>

URL linkification

-----------------

fmt:text-to-html knows how to linkify URLs:

    >>> text = (

    ...     'http://localhost:8086/bar/baz/foo.html\n'

    ...     'ftp://localhost:8086/bar/baz/foo.bar.html\n'

    ...     'sftp://localhost:8086/bar/baz/foo.bar.html.\n'

    ...     'http://localhost:8086/bar/baz/foo.bar.html;\n'

    ...     'news://localhost:8086/bar/baz/foo.bar.html:\n'

    ...     'http://localhost:8086/bar/baz/foo.bar.html?\n'

    ...     'http://localhost:8086/bar/baz/foo.bar.html,\n'

    ...     '<http://localhost:8086/bar/baz/foo.bar.html>\n'

    ...     '<http://localhost:8086/bar/baz/foo.bar.html>,\n'

    ...     '<http://localhost:8086/bar/baz/foo.bar.html>.\n'

    ...     '<http://localhost:8086/bar/baz/foo.bar.html>;\n'

    ...     '<http://localhost:8086/bar/baz/foo.bar.html>:\n'

    ...     '<http://localhost:8086/bar/baz/foo.bar.html>?\n'

    ...     '(http://localhost:8086/bar/baz/foo.bar.html)\n'

    ...     '(http://localhost:8086/bar/baz/foo.bar.html),\n'

    ...     '(http://localhost:8086/bar/baz/foo.bar.html).\n'

    ...     '(http://localhost:8086/bar/baz/foo.bar.html);\n'

    ...     '(http://localhost:8086/bar/baz/foo.bar.html):\n'

    ...     'http://localhost/bar/baz/foo.bar.html?a=b&b=a\n'

    ...     'http://localhost/bar/baz/foo.bar.html?a=b&b=a.\n'

    ...     'http://localhost/bar/baz/foo.bar.html?a=b&b=a,\n'

    ...     'http://localhost/bar/baz/foo.bar.html?a=b&b=a;\n'

    ...     'http://localhost/bar/baz/foo.bar.html?a=b&b=a:\n'

    ...     'http://localhost/bar/baz/foo.bar.html?'

    ...         'a=b&b=a:b;c@d_e%f~g#h,j!k-l+m$n*o\'p\n'

    ...     'http://www.searchtools.com/test/urls/(parens).html\n'

    ...     'http://www.searchtools.com/test/urls/-dash.html\n'

    ...     'http://www.searchtools.com/test/urls/_underscore.html\n'

    ...     'http://www.searchtools.com/test/urls/period.x.html\n'

    ...     'http://www.searchtools.com/test/urls/!exclamation.html\n'

    ...     'http://www.searchtools.com/test/urls/~tilde.html\n'

    ...     'http://www.searchtools.com/test/urls/*asterisk.html\n'

    ...     'irc://irc.freenode.net/launchpad\n'

    ...     'irc://irc.freenode.net/%23launchpad,isserver\n'

    ...     'mailto:noreply@launchpad.net\n'

    ...     'jabber:noreply@launchpad.net\n'

    ...     'http://localhost/foo?xxx&\n'

    ...     'http://localhost?testing=[square-brackets-in-query]\n'

    ... )

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a rel="nofollow" href="http://localhost:8086/bar/baz/foo.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>html</a><br />

    <a rel="nofollow" href="ftp://localhost:8086/bar/baz/foo.bar.html">ftp://localhost<wbr></wbr>:8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a><br />

    <a rel="nofollow" href="sftp://localhost:8086/bar/baz/foo.bar.html">sftp://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>.<br />

    <a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>;<br />

    <a rel="nofollow" href="news://localhost:8086/bar/baz/foo.bar.html">news://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>:<br />

    <a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>?<br />

    <a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>,<br />

    &lt;<a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>&gt;<br />

    &lt;<a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>&gt;,<br />

    &lt;<a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>&gt;.<br />

    &lt;<a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>&gt;;<br />

    &lt;<a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>&gt;:<br />

    &lt;<a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>&gt;?<br />

    (<a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>)<br />

    (<a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>),<br />

    (<a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>).<br />

    (<a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>);<br />

    (<a rel="nofollow" href="http://localhost:8086/bar/baz/foo.bar.html">http://<wbr></wbr>localhost:<wbr></wbr>8086/bar/<wbr></wbr>baz/foo.<wbr></wbr>bar.html</a>):<br />

    <a rel="nofollow" href="http://localhost/bar/baz/foo.bar.html?a=b&amp;b=a">http://<wbr></wbr>localhost/<wbr></wbr>bar/baz/<wbr></wbr>foo.bar.<wbr></wbr>html?a=<wbr></wbr>b&amp;b=a</a><br />

    <a rel="nofollow" href="http://localhost/bar/baz/foo.bar.html?a=b&amp;b=a">http://<wbr></wbr>localhost/<wbr></wbr>bar/baz/<wbr></wbr>foo.bar.<wbr></wbr>html?a=<wbr></wbr>b&amp;b=a</a>.<br />

    <a rel="nofollow" href="http://localhost/bar/baz/foo.bar.html?a=b&amp;b=a">http://<wbr></wbr>localhost/<wbr></wbr>bar/baz/<wbr></wbr>foo.bar.<wbr></wbr>html?a=<wbr></wbr>b&amp;b=a</a>,<br />

    <a rel="nofollow" href="http://localhost/bar/baz/foo.bar.html?a=b&amp;b=a">http://<wbr></wbr>localhost/<wbr></wbr>bar/baz/<wbr></wbr>foo.bar.<wbr></wbr>html?a=<wbr></wbr>b&amp;b=a</a>;<br />

    <a rel="nofollow" href="http://localhost/bar/baz/foo.bar.html?a=b&amp;b=a">http://<wbr></wbr>localhost/<wbr></wbr>bar/baz/<wbr></wbr>foo.bar.<wbr></wbr>html?a=<wbr></wbr>b&amp;b=a</a>:<br />

    <a rel="nofollow" href="http://localhost/bar/baz/foo.bar.html?a=b&amp;b=a:b;c@d_e%f~g#h,j!k-l+m$n*o'p">http://<wbr></wbr>localhost/<wbr></wbr>bar/baz/<wbr></wbr>foo.bar.<wbr></wbr>html?a=<wbr></wbr>b&amp;b=a:b;<wbr></wbr>c@d_e%f~<wbr></wbr>g#h,j!k-<wbr></wbr>l+m$n*o'<wbr></wbr>p</a><br />

    <a rel="nofollow" href="http://www.searchtools.com/test/urls/(parens).html">http://<wbr></wbr>www.searchtools<wbr></wbr>.com/test/<wbr></wbr>urls/(parens)<wbr></wbr>.html</a><br />

    <a rel="nofollow" href="http://www.searchtools.com/test/urls/-dash.html">http://<wbr></wbr>www.searchtools<wbr></wbr>.com/test/<wbr></wbr>urls/-dash.<wbr></wbr>html</a><br />

    <a rel="nofollow" href="http://www.searchtools.com/test/urls/_underscore.html">http://<wbr></wbr>www.searchtools<wbr></wbr>.com/test/<wbr></wbr>urls/_underscor<wbr></wbr>e.html</a><br />

    <a rel="nofollow" href="http://www.searchtools.com/test/urls/period.x.html">http://<wbr></wbr>www.searchtools<wbr></wbr>.com/test/<wbr></wbr>urls/period.<wbr></wbr>x.html</a><br />

    <a rel="nofollow" href="http://www.searchtools.com/test/urls/!exclamation.html">http://<wbr></wbr>www.searchtools<wbr></wbr>.com/test/<wbr></wbr>urls/!exclamati<wbr></wbr>on.html</a><br />

    <a rel="nofollow" href="http://www.searchtools.com/test/urls/~tilde.html">http://<wbr></wbr>www.searchtools<wbr></wbr>.com/test/<wbr></wbr>urls/~tilde.<wbr></wbr>html</a><br />

    <a rel="nofollow" href="http://www.searchtools.com/test/urls/*asterisk.html">http://<wbr></wbr>www.searchtools<wbr></wbr>.com/test/<wbr></wbr>urls/*asterisk.<wbr></wbr>html</a><br />

    <a rel="nofollow" href="irc://irc.freenode.net/launchpad">irc://irc.<wbr></wbr>freenode.<wbr></wbr>net/launchpad</a><br />

    <a rel="nofollow" href="irc://irc.freenode.net/%23launchpad,isserver">irc://irc.<wbr></wbr>freenode.<wbr></wbr>net/%23launchpa<wbr></wbr>d,isserver</a><br />

    <a rel="nofollow" href="mailto:noreply@launchpad.net">mailto:<wbr></wbr>noreply@<wbr></wbr>launchpad.<wbr></wbr>net</a><br />

    <a rel="nofollow" href="jabber:noreply@launchpad.net">jabber:<wbr></wbr>noreply@<wbr></wbr>launchpad.<wbr></wbr>net</a><br />

    <a rel="nofollow" href="http://localhost/foo?xxx&amp;">http://<wbr></wbr>localhost/<wbr></wbr>foo?xxx&amp;</a><br />

    <a rel="nofollow" href="http://localhost?testing=[square-brackets-in-query]">http://<wbr></wbr>localhost?<wbr></wbr>testing=<wbr></wbr>[square-<wbr></wbr>brackets-<wbr></wbr>in-query]</a></p>

The fmt:text-to-html formatter leaves a number of non-URIs unlinked:

    >>> text = (

    ...     'nothttp://launchpad.net/\n'

    ...     'http::No-cache=True\n')

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p>nothttp:<wbr></wbr>//launchpad.<wbr></wbr>net/<br />

    http::No-cache=True</p>

Bug references

--------------

fmt:text-to-html is also smart enough to convert bug references into

links:

    >>> text = (

    ...     'bug 123\n'

    ...     'bug    123\n'

    ...     'bug #123\n'

    ...     'bug number 123\n'

    ...     'bug number. 123\n'

    ...     'bug num 123\n'

    ...     'bug num. 123\n'

    ...     'bug no 123\n'

    ...     'bug report 123\n'

    ...     'bug no. 123\n'

    ...     'bug#123\n'

    ...     'bug-123\n'

    ...     'bug-report-123\n'

    ...     'bug=123\n'

    ...     'bug\n'

    ...     '#123\n'

    ...     'debug #52\n')

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="/bugs/123" class="bug-link">bug 123</a><br />

    <a href="/bugs/123" class="bug-link">bug    123</a><br />

    <a href="/bugs/123" class="bug-link">bug #123</a><br />

    <a href="/bugs/123" class="bug-link">bug number 123</a><br />

    bug number. 123<br />

    <a href="/bugs/123" class="bug-link">bug num 123</a><br />

    <a href="/bugs/123" class="bug-link">bug num. 123</a><br />

    <a href="/bugs/123" class="bug-link">bug no 123</a><br />

    <a href="/bugs/123" class="bug-link">bug report 123</a><br />

    <a href="/bugs/123" class="bug-link">bug no. 123</a><br />

    bug#123<br />

    <a href="/bugs/123" class="bug-link">bug-123</a><br />

    <a href="/bugs/123" class="bug-link">bug-report-123</a><br />

    <a href="/bugs/123" class="bug-link">bug=123</a><br />

    <a href="/bugs/123" class="bug-link">bug<br /> #123</a><br />

    debug #52</p>

    >>> text = (

    ...     'bug 123\n'

    ...     'bug 123\n')

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="/bugs/123" class="bug-link">bug 123</a><br />

    <a href="/bugs/123" class="bug-link">bug 123</a></p>

    >>> text = (

    ...     'bug 1234\n'

    ...     'bug 123\n')

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="/bugs/1234" class="bug-link">bug 1234</a><br />

    <a href="/bugs/123" class="bug-link">bug 123</a></p>

    >>> text = 'bug 0123\n'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="/bugs/123" class="bug-link">bug 0123</a></p>

We linkify bugs that are in the Ubuntu convention for referring to bugs in

Debian changelogs.

    >>> text = 'LP: #123.\n'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p>LP: <a href="/bugs/123" class="bug-link">#123</a>.</p>

Works with multiple bugs:

    >>> text = 'LP: #123, #2.\n'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p>LP: <a href="/bugs/123" class="bug-link">#123</a>, <a href="/bugs/2" class="bug-link">#2</a>.</p>

And with lower case 'lp' too:

    >>> text = 'lp: #123, #2.\n'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p>lp: <a href="/bugs/123" class="bug-link">#123</a>, <a href="/bugs/2" class="bug-link">#2</a>.</p>

Even line breaks cannot stop the power of bug linking:

    >>> text = 'LP:  #123,\n#2.\n'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p>LP:  <a href="/bugs/123" class="bug-link">#123</a>,<br />

    <a href="/bugs/2" class="bug-link">#2</a>.</p>

To check a private bug, we need to log in and set a bug to be private.

    >>> from zope.component import getUtility

    >>> from lp.bugs.interfaces.bug import IBugSet

    >>> bugset = getUtility(IBugSet)

    >>> firefox_crashes = bugset.get(6)

    >>> login("test@canonical.com")

    >>> current_user = getUtility(ILaunchBag).user

    >>> firefox_crashes.setPrivate(True, current_user)

    True

Bug.setPrivate adds all indirect subscribers to the bug as direct

subscribers, but we want to see what the bug looks like if we're not a

subscriber.

    >>> firefox_crashes.unsubscribe(current_user, current_user)

A private bug is still linked as no check is made on the actual bug.

    >>> text = 'bug 6\n'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="/bugs/6" class="bug-link">bug 6</a></p>

FAQ references

--------------

FAQ references are global, and also linkified:

    >>> text = (

    ...     'faq 1\n'

    ...     'faq #2\n'

    ...     'faq-2\n'

    ...     'faq=2\n'

    ...     'faq item 1\n'

    ...     'faq  number  2\n'

    ...     )

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="http://answers.launchpad.dev/ubuntu/+faq/1">faq 1</a><br />

    <a href="http://answers.launchpad.dev/ubuntu/+faq/2">faq #2</a><br />

    <a href="http://answers.launchpad.dev/ubuntu/+faq/2">faq-2</a><br />

    <a href="http://answers.launchpad.dev/ubuntu/+faq/2">faq=2</a><br />

    <a href="http://answers.launchpad.dev/ubuntu/+faq/1">faq item 1</a><br />

    <a href="http://answers.launchpad.dev/ubuntu/+faq/2">faq number 2</a></p>

Except, that is, when the FAQ doesn't exist:

    >>> text = (

    ...     'faq 999\n'

    ...     )

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p>faq 999</p>

Branch references

-----------------

Branch references are linkified:

    >>> text = (

    ...     'lp:~foo/bar/baz\n'

    ...     'lp:~foo/bar/bug-123\n'

    ...     'lp:~foo/+junk/baz\n'

    ...     'lp:~foo/ubuntu/jaunty/evolution/baz\n'

    ...     'lp:foo/bar\n'

    ...     'lp:foo\n'

    ...     'lp:foo,\n'

    ...     'lp:foo/bar.\n'

    ...     'lp:foo/bar/baz\n'

    ...     'lp:///foo\n'

    ...     'lp:/foo\n')

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="/+branch/~foo/bar/baz" class="...">lp:~foo/bar/baz</a><br />

    <a href="/+branch/~foo/bar/bug-123" class="...">lp:~foo/bar/bug-123</a><br />

    <a href="/+branch/~foo/+junk/baz" class="...">lp:~foo/+junk/baz</a><br />

    <a href="/+branch/~foo/ubuntu/jaunty/evolution/baz" class="...">lp:~foo/ubuntu/jaunty/evolution/baz</a><br />

    <a href="/+branch/foo/bar" class="...">lp:foo/bar</a><br />

    <a href="/+branch/foo" class="...">lp:foo</a><br />

    <a href="/+branch/foo" class="...">lp:foo</a>,<br />

    <a href="/+branch/foo/bar" class="...">lp:foo/bar</a>.<br />

    <a href="/+branch/foo/bar/baz" class="...">lp:foo/bar/baz</a><br />

    <a href="/+branch/foo" class="...">lp:///foo</a><br />

    <a href="/+branch/foo" class="...">lp:/foo</a></p>

Text that looks like a branch reference, but is followed only by digits is

treated as a link to a bug.

    >>> text = 'lp:1234'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="/bugs/1234" class="bug-link">lp:1234</a></p>

We are even smart enough to notice the trailing punctuation gunk and separate

that from the link.

    >>> text = 'lp:1234,'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="/bugs/1234" class="bug-link">lp:1234</a>,</p>

OOPS references

---------------

fmt:text-to-html is also smart enough to convert OOPS references into

links. However, it only does this if the logged in person is a member of the

Launchpad Developers team.

XXX 2006-08-23 jamesh

We explicitly cal set_developer_in_launchbag_before_traversal() here.

If this event handler is not called, then the "developer" attribute in

the launchbag is not updated.  Normally it would be called during the

request before traversal, but we aren't doing publication traversal in

this test.

  https://launchpad.net/bugs/30746

When not logged in as a privileged user, no link:

    >>> from lp.services.webapp.launchbag import (

    ...     set_developer_in_launchbag_before_traversal)

    >>> login('test@canonical.com')

    >>> set_developer_in_launchbag_before_traversal(None)

    >>> getUtility(ILaunchBag).developer

    False

    >>> text = 'OOPS-38C23'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p>OOPS-38C23</p>

After login, a link:

    >>> login('foo.bar@canonical.com')

    >>> set_developer_in_launchbag_before_traversal(None)

    >>> getUtility(ILaunchBag).developer

    True

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-38C23">OOPS-38C23</a></p>

OOPS references can take a number of forms:

    >>> text = 'OOPS-38C23'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-38C23">OOPS-38C23</a></p>

    >>> text = 'OOPS-123abcdef'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-123abcdef">OOPS-123abcdef</a></p>

    >>> text = 'OOPS-abcdef123'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-abcdef123">OOPS-abcdef123</a></p>

If the configuration value doesn't end with a slash, we won't add one. This

lets us configure the URL to use query parameters.

    >>> from lp.services.config import config

    >>> oops_root_url = """

    ...     [launchpad]

    ...     oops_root_url: http://foo/bar

    ...     """

    >>> config.push('oops_root_url', oops_root_url)

    >>> text = 'OOPS-38C23'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p><a href="http://foo/barOOPS-38C23">OOPS-38C23</a></p>

    >>> config_data = config.pop('oops_root_url')

Check against false positives:

    >>> text = 'OOPS code'

    >>> print test_tales('foo/fmt:text-to-html', foo=text)

    <p>OOPS code</p>

Reset login information.

    >>> login('test@canonical.com')

    >>> set_developer_in_launchbag_before_traversal(None)

    >>> getUtility(ILaunchBag).developer

    False

Regex helper functions

----------------------

The _substitute_matchgroup_for_spaces() static method is part of the

fmt:text-to-html code.  It is a helper for writing regular expressions where

we want to replace a variable number of spaces with the same number of

  entities.

    >>> from lp.app.browser.stringformatter import FormattersAPI

    >>> import re

    >>> matchobj = re.match('foo(.*)bar', 'fooX Ybar')

    >>> matchobj.groups()

    ('X Y',)

    >>> FormattersAPI._substitute_matchgroup_for_spaces(matchobj)

    '   '

The _linkify_substitution() static method is used for converting bug

references or URLs into links.  It uses the named matchgroups 'bug' and

'bugnum' when it is dealing with bugs, and 'url' when it is dealing with URLs.

First, let's try a match of nothing it understands.  This is a bug, so we get

an AssertionError.

    >>> matchobj = re.match(

    ...     ('(?P<bug>xxx)?(?P<faq>www)?(?P<url>yyy)?(?P<oops>zzz)?'

    ...      '(?P<lpbranchurl>www)?(?P<clbug>vvv)?'),

    ...     'fish')

    >>> sorted(matchobj.groupdict().items())

    [('bug', None),

     ('clbug', None),

     ('faq', None),

      ('lpbranchurl', None),

      ('oops', None),

      ('url', None)]

    >>> FormattersAPI._linkify_substitution(matchobj)

    Traceback (most recent call last):

    ...

    AssertionError: Unknown pattern matched.

When we have a URL, the URL is made into a link.  A quote is added to the

url to demonstrate quoting in the HTML attribute.

    >>> matchobj = re.match('(?P<bug>xxx)?(?P<url>y"y)?', 'y"y')

    >>> sorted(matchobj.groupdict().items())

    [('bug', None), ('url', 'y"y')]

    >>> FormattersAPI._linkify_substitution(matchobj)

    '<a rel="nofollow" href="y&quot;y">y"y</a>'

When we have a bug reference, the 'bug' group is used as the text of the link,

and the 'bugnum' is used to look up the bug.

    >>> matchobj = re.match(

    ...     '(?P<bug>xxxx)?(?P<bugnum>2)?(?P<url>yyy)?', 'xxxx2')

    >>> sorted(matchobj.groupdict().items())

    [('bug', 'xxxx'), ('bugnum', '2'), ('url', None)]

    >>> FormattersAPI._linkify_substitution(matchobj)

    '<a href="/bugs/2" class="bug-link">xxxx</a>'

When the bugnum doesn't match any bug, we still get a link, but get a message

in the link's title.

    >>> matchobj = re.match(

    ...     '(?P<bug>xxxx)?(?P<bugnum>2000)?(?P<url>yyy)?', 'xxxx2000')

    >>> sorted(matchobj.groupdict().items())

    [('bug', 'xxxx'), ('bugnum', '2000'), ('url', None)]

    >>> FormattersAPI._linkify_substitution(matchobj)

    '<a href="/bugs/2000" class="bug-link">xxxx</a>'