[Carpesimia]

I know this response comes pretty late, but I just inherited a forum running TFSEO on 3.8.7, and my boss wanted a sitemap immediately. I downloaded this package and tested it out. Here is what I found, and I hope it helps you...

1) The sitemap ran for awhile and then just stopped dead. It was missing a function called convert_int_to_utf8(). This could be the issue many of you reported. I added this function, and the sitemap ran to completion each time.

2) The sitemap generator is stupid, and grabs all forums and threads. Including private forums. So I added an "excuded_forums" variable, and excluded those forums in the 2 SQL statements.

3) The way this is written is S-L-O-W on a large forum. The forum Im using has 184,000 threads and it took 38 minutes to create the sitemap. The box is a bad-ass machine too, with 32 gb of ram and 4 quad processors. Its not a 386. Its because of the "select *" and "mysql_result()" grabbing alot of data, and then picking each piece from the results one at a time. So yeah, SLOW.

4) Having 184k urls, my sitemap is too big for google to eat.

So here's what I had to do to make it work:

1) I replaced the I/O with a database library I use all the time. Just this one change took the processing of this program from 38 minutes down to 11 seconds. This is because instead of grabbing the data piecemeal from the database, i only selected the fields i was using and grabbed the data in-bulk from the db and dropped it into an array for processing. So yeah, 11 seconds.

2) The file was still big, so i changed the program to write the file to disc instead of echoing out the output. This way I can generate it via cron, whenever i want.

3) I also added a new variable called $urlsperfile, and created a sitemap_#.xml with that many urls, cycling to a new file when i got more than $urlsperfile.

4) Added a call to gzip to compress all of the files i made.

5) created an uncompressed sitemap.xml file, which is a sitemap index file pointing to all of the sitemap_#.xml.gz files i just processed.

6) Finally, made some very MINOR changes to the xml to meet the sitemap.orgs requirements, found at: http://www.sitemaps.org/protocol.php

The final outcome? a cronjob that runs in less than 15 seconds, generates 10 sitemap files with 20k urls per, and one sitemap file linking them all.

Oh, and this is on a 3.8.7 forum.