trac-admin's wiki dump feature fails on hierarchical pages

[peter.fischer@…]

I have hierarchcal wikipages in one of my trac wikis e.g.: http://trac.srv/infrastruktur/wiki/TerraSelEinkauf/IngramMicro

​http://projects.edgewall.com/trac/wiki/WikiPageNames does allow this:

5. The use of slash ( / ) is permitted to create a hierarchy inside the wiki.

Error description: if I want to dump all wiki pages using trac-admin's wiki dump function, the following occurs:

apache ~% mkdir test
apache ~% cd test
apache ~/test% mkdir dumpdestination
apache ~/test% trac-admin /var/lib/trac/infrastruktur wiki dump dumpdestination
 TracReports => dumpdestination/TracReports
[...lots of wiki root level pages being dumped correctly...]
 TelefonListe => dumpdestination/TelefonListe
 TerraSelEinkauf => dumpdestination/TerraSelEinkauf
 TerraSelEinkauf/IngramMicro => dumpdestination/TerraSelEinkauf/IngramMicro
Wiki dump failed: [Errno 20] Not a directory: 'dumpdestination/TerraSelEinkauf/IngramMicro'
apache ~/test%

The bug consists of three errors:

• Dumping hierarchical wikipages using trac-admin's wiki dump function does not work.
• trac-admin breaks after error, so all pages after the first hierarchical one are ommitted, too.
• trac-admin does not inform about its failure, so a backup script can not decide about success or error:

   TerraSelEinkauf/IngramMicro => dumpdestination/TerraSelEinkauf/IngramMicro
  Wiki dump failed: [Errno 20] Not a directory: 'dumpdestination/TerraSelEinkauf/IngramMicro'
  apache ~/test% echo $?
  0
  

I use gentoo linux and stock trac 0.8.

[osamu.hashimoto@…]

Replace trac-admin function with following code.

I test it on Windows XP.

    def _do_wiki_export(self, page,filename=''):
        data=self.db_execsql("SELECT text FROM wiki "
                             " WHERE name='%s'"
                             " ORDER BY version DESC LIMIT 1" % page)
        text = data[0][0]
        if not filename:
            print text
        else:
            if os.path.isfile(filename):
                raise Exception("File '%s' exists" % filename)
            if not os.path.isdir(os.path.dirname(filename)):
                os.makedirs(os.path.dirname(filename))
            f = open(filename,'w')
            f.write(text)
            f.close()

    def _do_wiki_dump(self, dir):
        pages = self.get_wiki_list()
        for p in pages:
            lp=unicode( p,"utf-8").encode(encoding)
            dst = os.path.join(dir,lp)
            print " %s => %s" % (lp, dst)
            self._do_wiki_export(p, dst)

    def _do_wiki_load(self, dir,cursor=None, ignore=[], prefix=""):
        for page in os.listdir(dir):
            if page in ignore:
                continue
            entry = os.path.join(dir, page)
	    if not prefix == "":
                page = prefix + '/' + page 
            if os.path.isfile(entry):
                print " %s => %s" % (entry, page)
                self._do_wiki_import(entry,unicode(page,encoding).encode("utf-8"),  cursor)
            if os.path.isdir(entry):
                self._do_wiki_load(entry,None,[],page)

[Mark Rowe]

I believe that the code you suggest will break when you have both FooBar and FooBar/BarFoo as you cannot have a file + directory of the same name. A possible solution may be to output the files with their names URL-encoded. This would result in files named FooBar and FooBar%2FBarFoo being created for the case mentioned above.

[Mark Rowe]

This is fixed in [1504].

[Matthew Good]

#1628 has been marked as a duplicate of this ticket.