Latest SVN Changes forum section not updating
- Thread starter Ethelred
- Start date
matt.chugg
Moderator
hmm
line 21, column 50: XML parsing error: <unknown>:21:50: undefined entity
<title>10936: The Nostril of the Serpent is item n°5645, not 5545.</title>
it looks like the title has been htmlentitised, but that isn't valid xml, since the main use of the rss feed is for on the forum, i'd like to make sure it works as html however wrapping the title in cdata tags results in the cdata tags being encoded and breaking the feed again.
As a quick fix i'm going to modify the commit message in my database, and come up with a better fix later!
Edit: done.
Whilst this is a bug in MY code, this isn't really a kolmafia bug, could someone move this somewhere better please!
line 21, column 50: XML parsing error: <unknown>:21:50: undefined entity
<title>10936: The Nostril of the Serpent is item n°5645, not 5545.</title>
it looks like the title has been htmlentitised, but that isn't valid xml, since the main use of the rss feed is for on the forum, i'd like to make sure it works as html however wrapping the title in cdata tags results in the cdata tags being encoded and breaking the feed again.
As a quick fix i'm going to modify the commit message in my database, and come up with a better fix later!
Edit: done.
Whilst this is a bug in MY code, this isn't really a kolmafia bug, could someone move this somewhere better please!
Last edited:
Ah, sorry about that. I'm not sure how I should report bugs/problems with the forums, support tools, etc. Maybe some different categories. Or maybe I just overlooked already existing methods.
matt.chugg
Moderator
Not your fault slyz, to be fair, either the svntomysql script should handle it when getting the latest information, or more properly the feed generator should account for it and not encode them to html entities. probably. i'll have to double check the rss spec, but it isn't really intended for html I don't think.
Yeah, appears to be broken due to the ™ in Drizzlers™.
My guess? <?xml version="1.0" encoding="ISO-8859-1"?> should probably be changed to <?xml version="1.0" encoding="UTF-8"?> in the RSS feed.
Last edited:
So yeah, I read up on the rest of the thread instead of just replying to the issue at hand and I realized the old issue never got properly fixed. The two are separate issues, the most recent one due to encoding.
To fix the old issue, you need to escape the ampersand symbol in the HTML character entity. XML only has 5 entities < > & ' " (<, >, &, ', "), the XML parser got confused when it saw the non-XML defined character entity. It's a good idea to replace all instances of any of these reserved characters with their XML entities. In this case, ° would become ° but when the XML gets parsed, it goes back to looking like °.
To fix the more recent issue, see my post above, the XML encoding should be utf-8 to allow for the ™ symbol (which was not a HTML entity in this case).
To fix the old issue, you need to escape the ampersand symbol in the HTML character entity. XML only has 5 entities < > & ' " (<, >, &, ', "), the XML parser got confused when it saw the non-XML defined character entity. It's a good idea to replace all instances of any of these reserved characters with their XML entities. In this case, ° would become ° but when the XML gets parsed, it goes back to looking like °.
To fix the more recent issue, see my post above, the XML encoding should be utf-8 to allow for the ™ symbol (which was not a HTML entity in this case).
Winterbay
Active member
Well, I think the problem is that the RSS-bot is choking not that the forum is so even if we change the forum the bot still won't post things. And the bot is not run by fewyn...
Ah, right! Well my post was intended for matt.chugg but as you replied I thought you were assuming responsibility for it.
matt.chugg
Moderator
Reference to undefined entity 'Atilde'.
Line: 27 Character: 62
<title>11509: Add support for Summon Rad Libs. Add papier-mâché glob. Explicitly add the plural of...</title>
The problem is in the title element, which is encoded, but the xml standards don't allow anything except > < & and I "thought" wasn't allowed cdata encoding, however i've shouted at the rssbuilder class and told it to cdata the title anyway, lets see what happens!
The bot IS run by Fewyn, however the data it uses is run by me, in this instance, the data was the issue, and the bot was CORRECT to do nothing because the xml wasn't up to scratch. it would I presume have carried on when the ã went out of the last 10 posts leaving us a gap, the cdata fix seems to work.
Last edited:
matt.chugg
Moderator
Aslo, you can send me a message when it breaks! i'll see quicker like that, although i'd like to point out that uptime has been fairly good!
I dont' remember the issue in July, possibly missed that completly! sorry bout that! it should handle any future characters ok now though.
If there are any missing logs, you can still see them here: http://kolmafia.us/showthread.php?4145-SVN-Log-Searching
I dont' remember the issue in July, possibly missed that completly! sorry bout that! it should handle any future characters ok now though.
If there are any missing logs, you can still see them here: http://kolmafia.us/showthread.php?4145-SVN-Log-Searching
Last edited:
For reference, I described an alternative way of fixing the issue you described here, but glad to hear it should be fixed for good now.