Probably has something to do with PunBB's search caching.  On PunBB forums where some users have a very large number of posts, this makes it impossible to view their most recent posts.

http://punbb.org/forums/search.php?acti … rch=Submit

Reading the entire log file into memory and exploding it into an array every single time any page is called is a lot harder on the server than logging more often then every 10 seconds.

Well, it's no Mint.

Does PunBB 1.3 provide protection against a user breaking a page with invalid markup?

Jansson wrote:
Mark topics as read                               Miles

Apparently I don't need to port this? smile

6

(4 replies, posted in PunBB 1.2 discussion)

If you're running PHP as a CGI (like Dreamhost allows you to) the config.php doesn't need to be readable by anyone else.

What about a UNIQUE constraint on the online table on (user_id, ident)?

8

(3 replies, posted in Feature requests)

How about <link rel="first/next/prev/last" ...>, then?

9

(3 replies, posted in Feature requests)

PunBB 1.2 uses accesskeys on the Quickpost submit and Jump submit buttons, which is nice, but not terribly useful.  Can more be added, especially on navigation?  I'd kill for "p" and "n" to go to the previous/next page so I could read a thread without having to reach for the mouse to click a tiny page link. smile

Gizzmo wrote:

well one thing you can do is input the code correctly. like :

[quote][url]http://www.example.com/foo/bar/[/url][/quote]
Here is some more text.

maybe there should be a check to see if it nestled correctly and if it isnt give an error

Yeah, I realize that.

I wrote:

I know this isn't really that bad of a bug, because I'm passing it bad BBCode to begin with, but could it return some sort of BBCode parse error instead of vomiting strange markup?

It isn't that bad most of the time when you misnest the other tags.  It's only really annoying when the link shortner comes into play and truncates some of the HTML, sometimes badly screwing up the rest of the page.

The BBCode parser does all sorts of weird things with improperly nested quote and url tags (especially when the automagical linkshortenermajigger comes into effect).  I know this isn't really that bad of a bug, because I'm passing it bad BBCode to begin with, but could it return some sort of BBCode parse error instead of vomiting strange markup? smile

Credit for the find goes to ScifilterX at MacAddict Forums.  Sorry if this has already been posted about, I didn't see anything obvious in the first few pages of this forum.

The code:

[quote][url]http://www.example.com/foo/bar/[/quote]
[/url]Here is some more text.

The result:

<blockquote><div class="incqbox"><p><a href="http://www.example.com/foo/bar/</p></div></blockquote><p>">http://www.example.com/foo/bar/</p></di … kquote><p></a><br /><br />Here is some more text.

See it in action:
Removed by me, Rickard, because it was screwing up the format of the page.

This has become hard for me to debug in Safari with the anchor bug.

Sorry.  I found and fixed that bug three months ago but never posted the fix.

Step 45 of install.txt should actually be:

#
#---------[ 45. REPLACE WITH ]------------------------------------------------
#

    // MOD: MARK TOPICS AS READ - 10 LINES NEW CODE FOLLOW
    $result = $db->query('SELECT forum_id FROM '.$db->prefix.'topics WHERE id='.$id)
        or error('Unable to fetch post info', __FILE__, __LINE__, $db->error());
    $forum_id = $db->result($result);
    $last_read = $pun_user['last_visit'];
    if ($forum_id && !empty($pun_user['read_topics']['f'][$forum_id]) &&
        $pun_user['read_topics']['f'][$forum_id] >= $last_read)
        $last_read = $pun_user['read_topics']['f'][$forum_id];
    if (!empty($pun_user['read_topics']['t'][$id]) &&
        $pun_user['read_topics']['t'][$id] >= $last_read)
        $last_read = $pun_user['read_topics']['t'][$id];

The root of the problem though is that in the old version I accidentally used $pun_user['read_topic'] (but 'topic' should have been plural).

Odds are good that only the surrounding code has changed much.  Adapting the mod should be a snap... I just get tired of the tedium of counting lines for each new point release.

Not at this time.

Sorry for the confusion.  Out of curiosity, did you try installing it via the diff first?  I'd be interested in how well it works when there are other mods installed.

I really detest the current standard mod format.  It'd be awesome if PunBB had some sort of dedicated mod package system (since by design PunBB is lightweight and leaves much additional functionality to mods... but overly modding gets messy).  I haven't looked at PunMod, so I don't know if it's the right solution.  I was brainstorming some ideas for an implementation of a mod manager but I (a) don't know that it would be right for PunBB and (b) wouldn't want to compete with the existing standard.

Smartys is saying that your code is insecure and can be used to compromise your server by a hacker.

You should "scrub" the exclude= parameter to make sure that it doesn't contain any harmful SQL code.

    $exclude = -1;
    if (isset($_GET['exclude'])) {
        $excludes = explode(',', trim($_GET['exclude']));
        $excludes = array_map('intval', $excludes);
    
        if (!empty($excludes))
            $exclude=implode(',', $excludes);
    }

Mark topics as read 1.1.3.

fine in safari 1.3

I'm sorry, this seems to be the same bug Paul reported in a thread a few rows down.  But it seems to be fixed later in connor's post here...

but if the text is wrapped in quote tags

it's fine?
 what is going on?

It seems like it doesn't work if there is any text BEFORE the first {code} tag.

 test

I'm not sure why but I'm having a very hard time posting any code blocks on this forum, even something as simple as [ code ]this[/ code ].  I always get
? The BBCode syntax in the message is incorrect. Missing end tag for [ code ].
Am I doing something wrong?

The link truncation function works on an entitized version of the URL, which can occasionally cause an entity to get cut apart in an invalid way.  For example:

http://www.example.com/form.php?id=7&am … mp;where=4

http://validator.w3.org/check?uri=referer

I know this code is kind of quick and dirty ("Ok, not very pretty :-)") and that it's really a minor issue, but shouldn't it be easy to fix by de-entitizing the displayed URI before truncation and then re-entitizing it?