Bug number not shown as resolved in changset view

[anonymous]

When I commit my code to SVN, I include the bug number in the log message. When browsing the source, my log message shows the bug as resolved (strike-through formatting)

But, when I view the changeset, it appears to still be open: (no strike-through)

[Christian Boos]

That's a cacheing issue.

The content of the commit message does participate to the computation of the ETag value (the unique identifier used for detecting a content change), but only the raw content.

That would be quite easy to fix, though.

changeset.py

@@ -194 +194 @@
                                  new_path=new_path, new_rev=new)
         if chgset:
             chgset = repos.get_changeset(new)
+            message = chgset.message or '--'
+            if self.wiki_format_messages:
+                chgset.message = wiki_to_html(message, self.env, req,
+                                              escape_newlines=True)
+            else:
+                chgset.message = html.PRE(message)
             req.check_modified(chgset.date, [
                 diff_options[0],
                 ''.join(diff_options[1]),
@@ -299 +305 @@
                 properties.append({'name': name, 'value': value,
                                    'htmlclass': htmlclass})
 
-            message = chgset.message or '--'
-            if self.wiki_format_messages:
-                message = wiki_to_html(message, self.env, req,
-                                       escape_newlines=True)
-            else:
-                message = html.PRE(message)
             req.hdf['changeset'] = {
                 'revision': chgset.rev,
                 'time': util.format_datetime(chgset.date),
                 'age': util.pretty_timedelta(chgset.date, None, 3600),
                 'author': chgset.author or 'anonymous',
-                'message': message, 'properties': properties
+                'message': chgset.message, 'properties': properties
             }
             oldest_rev = repos.oldest_rev
             if chgset.rev != oldest_rev:

Any comment on this one, Christopher?

[Christopher Lenz]

The idea is good, but I don't think we should be changing the "message" member of the Changeset object, though. That kind of stuff can be very confusing IMHO.

[Christian Boos]

Well, it was for avoiding either to format the message twice or to give an extra arg to the _render_html method. But I see your point, there's no need to abuse the encapsulation of the Changeset object, even if Python makes this tempting :)

[Christian Boos]

2nd iteration of the patch for fixing this issue

[Christian Boos]

I was about to commit the attachment:3116.patch​, then I realized that without the patch, the changeset was also reloaded!

This prompted me to have a closer look, and it seems that my browsers (Firefox 1.5.3/windows, IE 6 something) don't set the expected If-None-Match: header anymore.

This is with trunk, dumped using Firefox/LiveHTTPHeaders (it's not the first load, of course):

http://localhost:8000/devel/changeset/70

GET /devel/changeset/70 HTTP/1.1
Host: localhost:8000
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.0.3) Gecko/20060426 Firefox/1.5.0.3
Accept: text/xml,application/xml,application/xhtml+xml,text/html;q=0.9,text/plain;q=0.8,image/png,*/*;q=0.5
Accept-Language: en-us,en;q=0.5
Accept-Encoding: gzip,deflate
Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7
Keep-Alive: 300
Connection: keep-alive
Cookie: trac_session=b3177286e05a9d0cfa15c5b7; trac_session=ca6b49eb53d04be055b54d53; trac_session=ca6b49eb53d04be055b54d53

HTTP/1.x 200 OK
Server: tracd/0.10dev Python/2.4.2
Date: Fri, 05 May 2006 09:21:19 GMT
Etag: W"anonymous/1130747279/701fdea080309adad112b3d3705f27d3"
Cache-Control: must-revalidate
Expires: Fri, 01 Jan 1999 00:00:00 GMT
Content-Type: text/html;charset=utf-8
Content-Length: 8712
----------------------------------------------------------

Same with 0.9.6dev:

http://localhost:8000/stable/changeset/70

GET /stable/changeset/70 HTTP/1.1
Host: localhost:8000
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.0.3) Gecko/20060426 Firefox/1.5.0.3
Accept: text/xml,application/xml,application/xhtml+xml,text/html;q=0.9,text/plain;q=0.8,image/png,*/*;q=0.5
Accept-Language: en-us,en;q=0.5
Accept-Encoding: gzip,deflate
Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7
Keep-Alive: 300
Connection: keep-alive
Cookie: trac_session=ca6b49eb53d04be055b54d53

HTTP/1.x 200 OK
Server: tracd/0.9.6dev Python/2.4.2
Date: Fri, 05 May 2006 09:40:38 GMT
Cache-Control: must-revalidate
Expires: Fri, 01 Jan 1999 00:00:00 GMT
Content-Type: text/html;charset=utf-8
Content-Length: 8609
Etag: W"anonymous/1130747279/inline-U2"
----------------------------------------------------------

I also verified that the If-None-Match is not sent when I don't use LiveHTTPHeaders, by adding a check in the request_modified method What's going on, is it just me?

[Christian Boos]

Patch integrated in r3310.

My ETag issues were due to the following reasons:

• IE doesn't seem to support the ETag we generate
• Firefox had the cache disable. That was not apparent in the options, though, and I had to do a Reset of the browser.cache.disk.enable and browser.cache.memory.enable settings in the about:config page…