Search options (Page 23 of 77)

That error message and the defined path in that code above would appear to be contradictory.

May be useful in regards to debugging:

http://archives.postgresql.org/pgsql-ge … g01385.php

That helps too.

Don't guess. Disable them one by one until the problem disappears. You then know which one is causing the problem.

If you're going to use sleep to pause it, it might be worthwhile chunking them then pausing, i.e: sending 25 then pausing, sending another 25 and pausing, etc. If not, it'll take a coons age on a large userbase, and most likely get timed out.

Just out of curiosity, with regards to this in 1.2*, where is the sig caching done? Just grepped for it, and all I found were these three lines:

viewtopic.php:303:              if (isset($signature_cache[$cur_post['poster_id']]))
viewtopic.php:304:                      $signature = $signature_cache[$cur_post['poster_id']];
viewtopic.php:308:                      $signature_cache[$cur_post['poster_id']] = $signature;

Work fine on K-Meleon too.

Personally, not fussed on it. I'm not a fan of these minimalistic layouts. I like tables to look like tables, with each part perfectly sectioned from the next.

http://punbb.org/forums/viewtopic.php?id=10518

Well, decided to call it a day with the html2pdf, and switched over to TCPDF. Output is nowhere near as pretty, (doesn't have the CSS support), but by christ, it's a damned sight less flaky than the previous software/parser. When it decided to work smoothly, it could easily ramp up stupid amounts of memory whilst parsing/processing. Set the PHP mem limit upto 256MB at one point, and it still ditched. So, thought sod that for a lark and switched. This one, (http://www.tecnick.com/public/code/cp_d … p_dp=tcpdf), is far less intensive, but still has a fairly straight forward API that doesn't require cartloads of cryptic pointers.

Same topic, slightly different question. I've implemented caching on the pdf feed now. Any chance someone could double check it's feeding okay, not corrupting, etc? Working fine from this end, but ne'er hurts to double check. First click on the pdf icon now creates the cache and feeds from it, and the second click just feeds directly from the cached copy unless the thread/article has been updated.