izzy10 years ago12 commits to master since this release
adding the possibility of download limits to the downloader class. This allows you to restrict downloads per IP and program to a given number within a given time frame (see doc/limits.txt) - limiting those ones who download entire directory trees (i.e. all available versions of a given prog).
replaced the E_USER_ERROR by E_USER_WARNING in the download class, since the former would still stop the script processing (and thus the download to fail, just for statistical issues - not nice)
adding possibility to map filenames to prognames for the download class, to e.g. map two different branches to the same prog (‘mylib’ and ‘mylib2’ can both point to ‘mylib’ this way)
fix: histview::set_max_relnum was doing nothing - now it does something (hey, that’s an improvement, huh?)
if the optional second parameter to the constructor of histview was omitted, it now falls back to the “basename” of the first (i.e. truncating the path and file extension from it)
izzy10 years ago17 commits to master since this release
adding referer check. Options for noRefererAction (empty referer) include “pass”, “deny”, and “whois” (check for crawler/bot nets)
adding remote network check. Especially some MS crawlers fake their UAs. But since they always leave the referer empty, we can limit the network checks to those. Use the histview_crawlernets file to specify their network information (one spec per line, all lowercase). These will be checked against the whois information for ‘netname’, ‘OrgName’, ‘OrgNOCName’, and ‘descr’.
the histview.inc file has been renamed to class.histview.inc. If you’re afraid you may miss to update some of your scripts on an upgrade, you may either create a symlink (preferred method on Linux/Unix systems), or create a dummy histview.inc just including/requiring the real one.
izzy11 years ago23 commits to master since this release
adding the possibility to reject file requests based on user agent (or to ignore them for the DL counter). Useful to e.g. exclude bots
download class now uses a configuration file for generic settings (you still can override those lateron, as usual). To make future updates easier, it automatically includes a personal configuration file if that exists
histview class now also uses the configuration file(s)
statistic functions (database related stuff like download counter) can be disabled now for the download class
if a database update fails, script is no longer stopped. Instead, an error is thrown (E_USER_ERROR) and the script continues, ignoring the issue
izzy12 years ago46 commits to master since this release
added the download class, which provides scanning of directories for downloadable files and creating links for them - or send a given file to the visitor requesting it, while optionally recording the download in a MySQL database
modified the histview class to (optionally) use the download class
modified the img tags to conform w3c
added the possibility to chose a different URL argument separator