first of all I would like to thank all the Devs and everybody who helped for that very well working and looking piece of Software. You don't see such a quality that often!
I have 0.7.20 running very well on Ubuntu Server 14.04.2 LTS 64-Bit. I did NOT installed the Ubuntu package but downloaded the source and made sure all required modules are installed (for these I used Ubuntu Packages since these are nicely up-to-date). The SABnzbd console (python SABnzbd.py) shows that all modules are present. Also downloading, extracting, moving, etc. works without issues. I also use the chuchusoft.com multi-threaded par2 (par2cmdline-0.4-tbb-20141125-lin64.tar.gz).
I have enabled Categories and Movie and TV Sorting. Temporary Download Folder is "/media/raid/incomplete" and Completed Download Folder is "/media/raid/complete". Category "movies" appends "Filme" to the path and category "shows" appends "Serien". Sorting just renames the job output but doesn't create directories. TV Sorting is in Format %sn/%s. Staffel/%sn.S%0sE%0e.%ext thus it adds directories, eg. /media/raid/complete/Serien/Barbapappa/1. Staffel/Barbapappa.S01E15.mkv
Lets start with the Post-Processing Script problem because the bugs/suggestions belong to this part.
1) Suggestion more variables for post-processing scripts:
SAB doesn't submit the jobs output name to the script and neither nzb-name ($2) nor nice name ($3) help because the final name is changed by Category/Sorting. This makes it impossible to move the correct file out of the directory. In case of Movies the path submitted ($1) is correct (/media/raid/complete/Filme) but in case of TV Shows it is not (/media/raid/complete/Serien/Barbapappa/1. Staffel). In case of TV Shows it should be /media/raid/complete/Serien. Now and with an additional submitted variable containing the final name (eg. "Movie.2013.mkv" for Movies or "Barbapappa/1. Staffel/Barbapappa.S01E15.mkv" for TV Shows) the Post-Processing script knows exactly what to process.
The new variable should contain the name of the file or directory that contains the whole job. If the job just had one file and that file is not sorted into subdirectories, the variable contains "filename.ext". In case the job contains a directory with one or more files the variable should contain the directory's name. If the job containes two or more files/directories which are not in their unique directory the variable should be an array containing all the file/directory names.
2) Bug / Suggesstion Post-Processing Script returned text:
When a Post-Processing script succeeded its last ouput is shown in green color (CSS class "hvMessageScript") in History. That's perfect. But when such a script fails (Exit Code != 0) not only its last output is shown but it's prefixed with "Exit(n)". Additionally it is still shown in green color (same CSS class). It would be perfect if for a failed script just its last output it's shown but in red color (CSS class "hvMessageFail"). That way one can have just "Succeess" / "Failed" shown or - if he wants - additionally the exit code. Additionally removing the "(More)" link would make it more stylish
since there is the script icon linking to the same destination.3) Bug Job icons in History are swapped:
When a job is being verified the icon order is: down-arrow, clock, wrench. That's perfect. When a job is being extracted, the icon order is: down-arrow, clock, dot-dot-dot, wrench. That's wrong since verifying (wrench) happens before extracting (dot-dot-dot). The wrong order stays the same after extracting is done. The script icon is correctly shown at last.
4) Suggestion Missing articles causes high amount of Connects/Disconnects when using Backup Server:
Establishing SSL connections needs much more ressources than keeping such connections alive. Currently, when an article is missing on primary server one connection to that server is closed and a new connection to the backup server is established. If that article is successfully downloaded or also missing on the backup server the connection is closed again and a new connection to the primary server for the next article is established. There should be an option (or even make it default behavior) for keeping such connections alive to decrease server load. This primarily helps usenet server providers.
