Blog "title" in the URL

[Alec Thomas]

A lot of blog entries seem to have a portion of the title inserted into the URL, after a little munging. For ​example.

You could add an extra magic formatting symbol to the Page Name Format String (eg. %@) which substitutes the munged page title.

I've attached a quick patch which does this, unfortunately if you update the entry title after it's been substituted into the page name, the new title won't be substituted. Dunno what to do about that, if anything.

[Alec Thomas]

Implementation of concept

[John Hampton]

(In [656]) * Now depends on Tags >= 0.3 and < 0.5

• Wrapped macro output in unicode(x, 'utf-8'), Closes #306
• Added patch from alect to support using the title in the URL, Closes #313
• Added patch from coderanger adding a NoFloatFormatter so that the blog won't render macros that support it, like the TocMacro.

[John Hampton]

For the time being, I don't think we should worry about the case where someone changes the title. There isn't a good method of renaming wikipages in trac and that's what would be required to support that. Also, if a link suddenly changes out from under someone, that'll probably be fairly surprising.

[Alec Thomas]

Got another patch which fixes one bug and cleans up my patch a bit. The bug occurred when you use something like [[BlogPost(tag=(devtodo,news),pagename="DevTodo/News/%Y-%m-%d-%@",readonly=1, link="Post new news item")]]

The strftime symbols were not being replaced, as substitution only occurred if a pagename was not provided.

The other fixes are truncating pagenames with %@ substitution to 60 characters, as that is the maximum size of the page title, and stripping leading and trailing hyphens (which can occur if you have a title like "I like cheese!!").

blog/new_blog.py

@@ -114 +114 @@
                                           '%Y/%m/%d/%H.%M')
         wikitext = req.args.get('text', '')
         blogtitle = req.args.get('blogtitle', '')
-        urltitle = re.sub(r'[^\w]+', '-', blogtitle).lower() 
-        while '-' in urltitle and len(urltitle) > 32: 
-            urltitle = '-'.join(urltitle.split('-')[:-1]) 
-        pagename = req.args.get('pagename', time.strftime(pg_name_fmt)) 
-        if '%@' in pagename and urltitle: 
+        pagename = req.args.get('pagename', pg_name_fmt) 
+        pagename = time.strftime(pagename)
+        if '%@' in pagename and blogtitle: 
+            urltitle = re.sub(r'[^\w]+', '-', blogtitle).lower() 
             pagename = pagename.replace('%@', urltitle) 
+            while '-' in pagename and len(pagename) > 60: 
+                pagename = '-'.join(pagename.split('-')[:-1]) 
+            pagename.strip('-')
         comment = req.args.get('comment', '')
         readonly = int(req.args.has_key('readonly'))
         edit_rows = int(req.args.get('edite_rows', 20))

[Alec Thomas]

Whoops, stupid.

Index: blog/new_blog.py
===================================================================
--- blog/new_blog.py    (revision 657)
+++ blog/new_blog.py    (working copy)
@@ -114,12 +114,14 @@
                                           '%Y/%m/%d/%H.%M')
         wikitext = req.args.get('text', '')
         blogtitle = req.args.get('blogtitle', '')
-        urltitle = re.sub(r'[^\w]+', '-', blogtitle).lower() 
-        while '-' in urltitle and len(urltitle) > 32: 
-            urltitle = '-'.join(urltitle.split('-')[:-1]) 
-        pagename = req.args.get('pagename', time.strftime(pg_name_fmt)) 
-        if '%@' in pagename and urltitle: 
+        pagename = req.args.get('pagename', pg_name_fmt) 
+        pagename = time.strftime(pagename)
+        if '%@' in pagename and blogtitle: 
+            urltitle = re.sub(r'[^\w]+', '-', blogtitle).lower() 
             pagename = pagename.replace('%@', urltitle) 
+            while '-' in pagename and len(pagename) > 60: 
+                pagename = '-'.join(pagename.split('-')[:-1]) 
+            pagename = pagename.strip('-')
         comment = req.args.get('comment', '')
         readonly = int(req.args.has_key('readonly'))
         edit_rows = int(req.args.get('edite_rows', 20))

(in plain text for ease of copy and pasting)

[Alec Thomas]

Fixed in r658.