Changeset - e5a7f8f41370
[Not reviewed]
default
0 3 0
Thomas De Schampheleire - 8 years ago 2018-02-16 22:30:51
thomas.de_schampheleire@nokia.com
issues: backout special whitespace handling

This is essentially a backout of commit 32e1e0745d3c.

That commit checked for whitespace at the beginning of the matched issue
reference, and explicitly retained it in the resulting link text.

The way this was handled is not only suboptimal, e.g. a set of 4 spaces
would still be reduced to 1, but is also not actually necessary: if
whitespace before the issue reference is not required, then it does not need
to be specified in the issue pattern, and if it _is_ required, then a
positive lookbehind assertion can be used instead.
3 files changed with 20 insertions and 7 deletions:
0 comments (0 inline, 0 general)
kallithea/lib/helpers.py
Show inline comments
 
@@ -1111,107 +1111,105 @@ def linkify_others(t, l):
 
            links.append('<a class="message-link" href="%s">%s</a>' % (l, e))
 
        else:
 
            links.append(e)
 

	
 
    return ''.join(links)
 

	
 

	
 
# Global variable that will hold the actual urlify_issues function body.
 
# Will be set on first use when the global configuration has been read.
 
_urlify_issues_f = None
 

	
 

	
 
def urlify_issues(newtext, repo_name):
 
    """Urlify issue references according to .ini configuration"""
 
    global _urlify_issues_f
 
    if _urlify_issues_f is None:
 
        from kallithea import CONFIG
 
        from kallithea.model.db import URL_SEP
 
        assert CONFIG['sqlalchemy.url'] # make sure config has been loaded
 

	
 
        # Build chain of urlify functions, starting with not doing any transformation
 
        tmp_urlify_issues_f = lambda s: s
 

	
 
        issue_pat_re = re.compile(r'issue_pat(.*)')
 
        for k in CONFIG.keys():
 
            # Find all issue_pat* settings that also have corresponding server_link and prefix configuration
 
            m = issue_pat_re.match(k)
 
            if m is None:
 
                continue
 
            suffix = m.group(1)
 
            issue_pat = CONFIG.get(k)
 
            issue_server_link = CONFIG.get('issue_server_link%s' % suffix)
 
            issue_prefix = CONFIG.get('issue_prefix%s' % suffix)
 
            if not issue_pat or not issue_server_link or issue_prefix is None: # issue_prefix can be empty but should be present
 
                log.error('skipping incomplete issue pattern %r: %r -> %r %r', suffix, issue_pat, issue_server_link, issue_prefix)
 
                continue
 

	
 
            # Wrap tmp_urlify_issues_f with substitution of this pattern, while making sure all loop variables (and compiled regexpes) are bound
 
            try:
 
                issue_re = re.compile(issue_pat)
 
            except re.error as e:
 
                log.error('skipping invalid issue pattern %r: %r -> %r %r. Error: %s', suffix, issue_pat, issue_server_link, issue_prefix, str(e))
 
                continue
 

	
 
            log.debug('issue pattern %r: %r -> %r %r', suffix, issue_pat, issue_server_link, issue_prefix)
 

	
 
            def issues_replace(match_obj,
 
                               issue_server_link=issue_server_link, issue_prefix=issue_prefix):
 
                leadingspace = ' ' if match_obj.group().startswith(' ') else ''
 
                issue_id = ''.join(match_obj.groups())
 
                issue_url = issue_server_link.replace('{id}', issue_id)
 
                issue_url = issue_url.replace('{repo}', repo_name)
 
                issue_url = issue_url.replace('{repo_name}', repo_name.split(URL_SEP)[-1])
 
                return (
 
                    '%(leadingspace)s<a class="issue-tracker-link" href="%(url)s">'
 
                    '<a class="issue-tracker-link" href="%(url)s">'
 
                    '%(issue-prefix)s%(id-repr)s'
 
                    '</a>'
 
                    ) % {
 
                     'leadingspace': leadingspace,
 
                     'url': issue_url,
 
                     'id-repr': issue_id,
 
                     'issue-prefix': issue_prefix,
 
                     'serv': issue_server_link,
 
                    }
 
            tmp_urlify_issues_f = (lambda s,
 
                                          issue_re=issue_re, issues_replace=issues_replace, chain_f=tmp_urlify_issues_f:
 
                                   issue_re.sub(issues_replace, chain_f(s)))
 

	
 
        # Set tmp function globally - atomically
 
        _urlify_issues_f = tmp_urlify_issues_f
 

	
 
    return _urlify_issues_f(newtext)
 

	
 

	
 
def render_w_mentions(source, repo_name=None):
 
    """
 
    Render plain text with revision hashes and issue references urlified
 
    and with @mention highlighting.
 
    """
 
    s = safe_unicode(source)
 
    s = urlify_text(s, repo_name=repo_name)
 
    return literal('<div class="formatted-fixed">%s</div>' % s)
 

	
 

	
 
def short_ref(ref_type, ref_name):
 
    if ref_type == 'rev':
 
        return short_id(ref_name)
 
    return ref_name
 

	
 

	
 
def link_to_ref(repo_name, ref_type, ref_name, rev=None):
 
    """
 
    Return full markup for a href to changeset_home for a changeset.
 
    If ref_type is branch it will link to changelog.
 
    ref_name is shortened if ref_type is 'rev'.
 
    if rev is specified show it too, explicitly linking to that revision.
 
    """
 
    txt = short_ref(ref_type, ref_name)
 
    if ref_type == 'branch':
 
        u = url('changelog_home', repo_name=repo_name, branch=ref_name)
 
    else:
 
        u = url('changeset_home', repo_name=repo_name, revision=ref_name)
 
    l = link_to(repo_name + '#' + txt, u)
 
    if rev and ref_type != 'rev':
 
        l = literal('%s (%s)' % (l, link_to(short_id(rev), url('changeset_home', repo_name=repo_name, revision=rev))))
 
    return l
 

	
kallithea/lib/paster_commands/template.ini.mako
Show inline comments
 
@@ -222,96 +222,99 @@ git_path = git
 
<%text>## hide all refs in changelog switch this to --branches --tags</%text>
 
#git_rev_filter = --branches --tags
 

	
 
<%text>## RSS feed options</%text>
 
rss_cut_off_limit = 256000
 
rss_items_per_page = 10
 
rss_include_diff = false
 

	
 
<%text>## options for showing and identifying changesets</%text>
 
show_sha_length = 12
 
show_revision_number = false
 

	
 
<%text>## Canonical URL to use when creating full URLs in UI and texts.</%text>
 
<%text>## Useful when the site is available under different names or protocols.</%text>
 
<%text>## Defaults to what is provided in the WSGI environment.</%text>
 
#canonical_url = https://kallithea.example.com/repos
 

	
 
<%text>## gist URL alias, used to create nicer urls for gist. This should be an</%text>
 
<%text>## url that does rewrites to _admin/gists/<gistid>.</%text>
 
<%text>## example: http://gist.example.com/{gistid}. Empty means use the internal</%text>
 
<%text>## Kallithea url, ie. http[s]://kallithea.example.com/_admin/gists/<gistid></%text>
 
gist_alias_url =
 

	
 
<%text>## white list of API enabled controllers. This allows to add list of</%text>
 
<%text>## controllers to which access will be enabled by api_key. eg: to enable</%text>
 
<%text>## api access to raw_files put `FilesController:raw`, to enable access to patches</%text>
 
<%text>## add `ChangesetController:changeset_patch`. This list should be "," separated</%text>
 
<%text>## Syntax is <ControllerClass>:<function>. Check debug logs for generated names</%text>
 
<%text>## Recommended settings below are commented out:</%text>
 
api_access_controllers_whitelist =
 
#    ChangesetController:changeset_patch,
 
#    ChangesetController:changeset_raw,
 
#    FilesController:raw,
 
#    FilesController:archivefile
 

	
 
<%text>## default encoding used to convert from and to unicode</%text>
 
<%text>## can be also a comma separated list of encoding in case of mixed encodings</%text>
 
default_encoding = utf8
 

	
 
<%text>## issue tracker for Kallithea (leave blank to disable, absent for default)</%text>
 
#bugtracker = https://bitbucket.org/conservancy/kallithea/issues
 

	
 
<%text>## issue tracking mapping for commits messages</%text>
 
<%text>## comment out issue_pat, issue_server, issue_prefix to enable</%text>
 

	
 
<%text>## pattern to get the issues from commit messages</%text>
 
<%text>## default one used here is #<numbers> with a regex passive group for `#`</%text>
 
<%text>## {id} will be all groups matched from this pattern</%text>
 
<%text>## To require mandatory whitespace before the issue pattern, use:</%text>
 
<%text>## (?:^|(?<=\s)) before the actual pattern, and for mandatory whitespace</%text>
 
<%text>## behind the issue pattern, use (?:$|(?=\s)) after the actual pattern</%text>
 

	
 
issue_pat = #(\d+)
 

	
 
<%text>## server url to the issue, each {id} will be replaced with match</%text>
 
<%text>## fetched from the regex and {repo} is replaced with full repository name</%text>
 
<%text>## including groups {repo_name} is replaced with just name of repo</%text>
 

	
 
issue_server_link = https://issues.example.com/{repo}/issue/{id}
 

	
 
<%text>## prefix to add to link to indicate it's an url</%text>
 
<%text>## #314 will be replaced by <issue_prefix><id></%text>
 

	
 
issue_prefix = #
 

	
 
<%text>## issue_pat, issue_server_link, issue_prefix can have suffixes to specify</%text>
 
<%text>## multiple patterns, to other issues server, wiki or others</%text>
 
<%text>## below an example how to create a wiki pattern</%text>
 
# wiki-some-id -> https://wiki.example.com/some-id
 

	
 
#issue_pat_wiki = (?:wiki-)(.+)
 
#issue_server_link_wiki = https://wiki.example.com/{id}
 
#issue_prefix_wiki = WIKI-
 

	
 
<%text>## alternative return HTTP header for failed authentication. Default HTTP</%text>
 
<%text>## response is 401 HTTPUnauthorized. Currently Mercurial clients have trouble with</%text>
 
<%text>## handling that. Set this variable to 403 to return HTTPForbidden</%text>
 
auth_ret_code =
 

	
 
<%text>## locking return code. When repository is locked return this HTTP code. 2XX</%text>
 
<%text>## codes don't break the transactions while 4XX codes do</%text>
 
lock_ret_code = 423
 

	
 
<%text>## allows to change the repository location in settings page</%text>
 
allow_repo_location_change = True
 

	
 
<%text>## allows to setup custom hooks in settings page</%text>
 
allow_custom_hooks_settings = True
 

	
 
<%text>## extra extensions for indexing, space separated and without the leading '.'.</%text>
 
# index.extensions =
 
#    gemfile
 
#    lock
 

	
 
<%text>## extra filenames for indexing, space separated</%text>
 
# index.filenames =
 
#    .dockerignore
 
#    .editorconfig
 
#    INSTALL
kallithea/tests/other/test_libs.py
Show inline comments
 
@@ -366,112 +366,124 @@ class TestLibs(TestController):
 
       ""),
 
      ("deadbeefcafe 123412341234",
 
       """<a class="changeset_hash" href="/repo_name/changeset/deadbeefcafe">deadbeefcafe</a> <a class="changeset_hash" href="/repo_name/changeset/123412341234">123412341234</a>""",
 
       ""),
 
      ("We support * markup for *bold* markup of *single or multiple* words, "
 
       "*a bit @like http://slack.com*. "
 
       "The first * must come after whitespace and not be followed by whitespace, "
 
       "contain anything but * and newline until the next *, "
 
       "which must not come after whitespace "
 
       "and not be followed by * or alphanumerical *characters*.",
 
       """We support * markup for <b>*bold*</b> markup of <b>*single or multiple*</b> words, """
 
       """<b>*a bit <b>@like</b> <a href="http://slack.com">http://slack.com</a>*</b>. """
 
       """The first * must come after whitespace and not be followed by whitespace, """
 
       """contain anything but * and newline until the next *, """
 
       """which must not come after whitespace """
 
       """and not be followed by * or alphanumerical <b>*characters*</b>.""",
 
       "-"),
 
      ("HTML escaping: <abc> 'single' \"double\" &pointer",
 
       "HTML escaping: &lt;abc&gt; &#39;single&#39; &quot;double&quot; &amp;pointer",
 
       "-"),
 
      # tags are covered by test_tag_extractor
 
    ])
 
    def test_urlify_test(self, sample, expected, url_):
 
        expected = self._quick_url(expected,
 
                                   tmpl="""<a href="%s">%s</a>""", url_=url_)
 
        fake_url = FakeUrlGenerator(changeset_home='/%(repo_name)s/changeset/%(revision)s')
 
        with mock.patch('kallithea.config.routing.url', fake_url):
 
            from kallithea.lib.helpers import urlify_text
 
            assert urlify_text(sample, 'repo_name', stylize=True) == expected
 

	
 
    @parametrize('sample,expected', [
 
      ("deadbeefcafe @mention, and http://foo.bar/ yo",
 
       """<a class="changeset_hash" href="/repo_name/changeset/deadbeefcafe">deadbeefcafe</a>"""
 
       """<a class="message-link" href="#the-link"> <b>@mention</b>, and </a>"""
 
       """<a href="http://foo.bar/">http://foo.bar/</a>"""
 
       """<a class="message-link" href="#the-link"> yo</a>"""),
 
    ])
 
    def test_urlify_link(self, sample, expected):
 
        fake_url = FakeUrlGenerator(changeset_home='/%(repo_name)s/changeset/%(revision)s')
 
        with mock.patch('kallithea.config.routing.url', fake_url):
 
            from kallithea.lib.helpers import urlify_text
 
            assert urlify_text(sample, 'repo_name', link_='#the-link') == expected
 

	
 
    @parametrize('issue_pat,issue_server,issue_prefix,sample,expected', [
 
        (r'#(\d+)', 'http://foo/{repo}/issue/{id}', '#',
 
            'issue #123 and issue#456',
 
            """issue <a class="issue-tracker-link" href="http://foo/repo_name/issue/123">#123</a> and """
 
            """issue<a class="issue-tracker-link" href="http://foo/repo_name/issue/456">#456</a>"""),
 
        # following test case shows the result of a backward incompatible change that was made: the
 
        # space between 'issue' and '#123' is removed, because the space is part of the pattern.
 
        (r'(?:\s*#)(\d+)', 'http://foo/{repo}/issue/{id}', '#',
 
            'issue #123 and issue#456',
 
            """issue <a class="issue-tracker-link" href="http://foo/repo_name/issue/123">#123</a> and """
 
            """issue<a class="issue-tracker-link" href="http://foo/repo_name/issue/123">#123</a> and """
 
            """issue<a class="issue-tracker-link" href="http://foo/repo_name/issue/456">#456</a>"""),
 
        # to require whitespace before the issue reference, one may be tempted to use \b...
 
        (r'\bPR(\d+)', 'http://foo/{repo}/issue/{id}', '#',
 
            'issue PR123 and issuePR456',
 
            """issue <a class="issue-tracker-link" href="http://foo/repo_name/issue/123">#123</a> and """
 
            """issuePR456"""),
 
        # following test case shows that \b does not work well in combination with '#': the expectations
 
        # ... but it turns out that \b does not work well in combination with '#': the expectations
 
        # are reversed from what is actually happening.
 
        (r'\b#(\d+)', 'http://foo/{repo}/issue/{id}', '#',
 
            'issue #123 and issue#456',
 
            """issue #123 and """
 
            """issue<a class="issue-tracker-link" href="http://foo/repo_name/issue/456">#456</a>"""),
 
        (r'[ \t]#(\d+)', 'http://foo/{repo}/issue/{id}', '#',
 
            'issue #123 and issue#456',
 
        # ... so maybe try to be explicit? Unfortunately the whitespace before the issue
 
        # reference is not retained, again, because it is part of the pattern.
 
        (r'(?:^|\s)#(\d+)', 'http://foo/{repo}/issue/{id}', '#',
 
            '#15 and issue #123 and issue#456',
 
            """<a class="issue-tracker-link" href="http://foo/repo_name/issue/15">#15</a> and """
 
            """issue<a class="issue-tracker-link" href="http://foo/repo_name/issue/123">#123</a> and """
 
            """issue#456"""),
 
        # ... instead, use lookbehind assertions.
 
        (r'(?:^|(?<=\s))#(\d+)', 'http://foo/{repo}/issue/{id}', '#',
 
            '#15 and issue #123 and issue#456',
 
            """<a class="issue-tracker-link" href="http://foo/repo_name/issue/15">#15</a> and """
 
            """issue <a class="issue-tracker-link" href="http://foo/repo_name/issue/123">#123</a> and """
 
            """issue#456"""),
 
        (r'(?:pullrequest|pull request|PR|pr) ?#?(\d+)', 'http://foo/{repo}/issue/{id}', 'PR#',
 
            'fixed with pullrequest #1, pull request#2, PR 3, pr4',
 
            """fixed with <a class="issue-tracker-link" href="http://foo/repo_name/issue/1">PR#1</a>, """
 
            """<a class="issue-tracker-link" href="http://foo/repo_name/issue/2">PR#2</a>, """
 
            """<a class="issue-tracker-link" href="http://foo/repo_name/issue/3">PR#3</a>, """
 
            """<a class="issue-tracker-link" href="http://foo/repo_name/issue/4">PR#4</a>"""),
 
        (r'#(\d+)', 'http://foo/{repo}/issue/{id}', 'PR',
 
            'interesting issue #123',
 
            """interesting issue <a class="issue-tracker-link" href="http://foo/repo_name/issue/123">PR123</a>"""),
 
        (r'BUG\d{5}', 'https://bar/{repo}/{id}', 'BUG',
 
            'silly me, I did not parenthesize the {id}, BUG12345.',
 
            """silly me, I did not parenthesize the {id}, <a class="issue-tracker-link" href="https://bar/repo_name/">BUG</a>."""),
 
        (r'BUG(\d{5})', 'https://bar/{repo}/', 'BUG',
 
            'silly me, the URL does not contain {id}, BUG12345.',
 
            """silly me, the URL does not contain {id}, <a class="issue-tracker-link" href="https://bar/repo_name/">BUG12345</a>."""),
 
        (r'(PR-\d+)', 'http://foo/{repo}/issue/{id}', '',
 
            'interesting issue #123, err PR-56',
 
            """interesting issue #123, err <a class="issue-tracker-link" href="http://foo/repo_name/issue/PR-56">PR-56</a>"""),
 
        (r'#(\d+)', 'http://foo/{repo}/issue/{id}', '#',
 
            "some 'standard' text with apostrophes",
 
            """some &#39;standard&#39; text with apostrophes"""),
 
        (r'#(\d+)', 'http://foo/{repo}/issue/{id}', '#',
 
            "some 'standard' issue #123",
 
            """some &#39;standard&#39; issue <a class="issue-tracker-link" href="http://foo/repo_name/issue/123">#123</a>"""),
 
        (r'#(\d+)', 'http://foo/{repo}/issue/{id}', '#',
 
            'an issue   #123       with extra whitespace',
 
            """an issue   <a class="issue-tracker-link" href="http://foo/repo_name/issue/123">#123</a>       with extra whitespace"""),
 
        # Note: whitespace is squashed
 
        (r'(?:\s*#)(\d+)', 'http://foo/{repo}/issue/{id}', '#',
 
            'an issue   #123       with extra whitespace',
 
            """an issue <a class="issue-tracker-link" href="http://foo/repo_name/issue/123">#123</a>       with extra whitespace"""),
 
        # invalid issue pattern
 
        (r'(PR\d+', 'http://foo/{repo}/issue/{id}', '',
 
            'PR135',
 
            """PR135"""),
 
    ])
 
    def test_urlify_issues(self, issue_pat, issue_server, issue_prefix, sample, expected):
 
        from kallithea.lib.helpers import urlify_text
 
        config_stub = {
 
            'sqlalchemy.url': 'foo',
 
            'issue_pat': issue_pat,
 
            'issue_server_link': issue_server,
 
            'issue_prefix': issue_prefix,
 
        }
 
        # force recreation of lazy function
 
        with mock.patch('kallithea.lib.helpers._urlify_issues_f', None):
0 comments (0 inline, 0 general)