v0.13: o Can now optionally store encoded article segments in memory, instead of disk. This can improve performance and CPU usage in some cases. Set CACHE_LIMIT to a non zero value to enable. Hellanzb.CACHE_LIMIT = -1 # for an unlimited sized cache, or Hellanzb.CACHE_LIMIT = '100MB' # for a certain size o Added SSL NNTP support: defineServer ssl=True. (patch submitted by Oscar H) o Fixed the PostProcessor on Windows, which would produce the error: exceptions.ValueError: The usePTY parameter is not supported on Windows. o Fixed download problems on Windows, which would repeatedly produce errors such as: Archive segment: 2 does not have a valid CRC/yend line! (thanks Ryan van Barneveld) o Fixed total archive time reported at the end of PostProcessing (downloadTime was tallied as 0) o Fixed the Hellanzb.CATEGORIZE_DEST option: it was impossible to disable in v0.12 o Fixed PostProcessor handling of filenames containing double quotes o Fixed the error message: exceptions.TypeError: errBack() takes exactly 2 arguments (1 given) occurring during failed newzbin download attempts (thanks mke & JJ2) o Fixed the following error triggered after failed yenc CRC checks: File "hellanzb-0.12/Hellanzb/NZBLeecher/ArticleDecoder.py", line 685, in assembleNZBFile decodedSegmentFile = open(nzbSegment.getDestination(), 'rb') exceptions.IOError: [Errno 2] No such file or directory: This issue was introduced in v0.11. v0.12: o Fixed v3.newzbin.com's DirectNZB support which was broken as of 2007-02-18 o Improved downloader performance (CPU used and max download speed), especially for faster connections (such as over 40mb/s) o Added support for libnotify notifications (patch submitted by Ruben Fonseca) o No longer hiding exceptions; many were previously only shown in the debug log file o Now printing the 'Transferred MB in Time at Speed (Archive)' message when the NZB is finished downloading, as opposed to when the archive has finished decoding. The difference could be significant on slower CPUs/ and or faster connections v0.11: o Fixed the error: ImportError: cannot import name daemonize with Twisted 2.5 o Added the Hellanzb.CATEGORIZE_DEST option. Enabling it saves archives into a sub directory of DEST_DIR named after their newzbin.com category (when queued using the enqueuenewzbin XMLRPC call); e.g. Apps, Movies, Music o enqueuenewzbin now downloads gzipped newzbin.com NZBs via the new v3 DirectNZB interface (Thanks DBordello and Thomas Hurst) o Added support for fillservers (by priority) o Now requeueing segments to other servers when they fail yDecode CRC checks o Added optional support for converting MacBinary files o Now properly parsing v3.newzbin.com NZB names o Fixed the enqueuenewzbin error: exceptions.AttributeError: 'module' object has no attribute 'b64encode' on Python 2.3 o No longer overwriting pre-existing files during rar extraction o Fixed handling of NZB files with invalid segment number and bytes attributes o Fixed .zip files in the QUEUE_DIR constantly being rescanned for NZBs o Fixed a couple issues when antiIdle is disabled, including reconnection failures after connections are lost o antiIdle disabled connections now avoid connecting at startup, and disconnect immediately when there is nothing left to download. i.e. they only are connected to the server when necessary. antiIdle disabled fill servers only connect when they have something to download o Windows compatibility changes -- hellanzb has been reported to work on Windows as of the 0.11 changes (thanks hax) v0.10: o Multiple servers are now represented with different colors in the scroller output o Added the ability to run an external handler script on an archive after post processing o Now including duplicate files during par2 repair o All filename paths in the config file now expand the "~" character to the user's home directory o Added support for compressed (.zip) nzbs (thanks Freddie) o Fixed the following IOError occurring on malformed NZBs that include multiple s of the same segment number for a particular (particularly those NZBs created by newzleech.com): IOError: [Errno 2] No such file or directory: '/nzb/daemon.temp/download-tmp/hellanzb-tmp-archvie.file0172.segment0001_ENC' o Added the ability to disconnect from the usenet server when idle (set antiIdle = 0) o Fixed defineServer's bindTo option o Added the enqueueurl RPC call (downloads the NZB at the specified URL) o Better handling of slow NZB copies to the queue directory (won't queue the NZB until it's fully copied) o Rars within rars are no longer deleted by the PostProcessor in some cases o A few misc. lockup/stability related fixes v0.9: NZBLeecher: o New feature: smartpar. hellanzb will now skip downloading extra par2/par1 files. If the post processor finds the archive requires more par data for repair, hellanzb will automatically force the download of a minimal amount of extra pars immediately Daemon: o New remote call: 'setrarpass' (From hellanzb.py --help) setrarpass nzbid pass Set the rarPassword for the NZB with the specified ID o Rar passwords are now saved in between CTRL-Cs. o No longer leaving enqueuednewzbin/dequeued/cancelled NZBs in the TEMP_DIR o Added the Hellanzb.UMASK option to force umask in daemon mode v0.8: Daemon: o Fixed a nasty bug with the 'clear' RPC call -- it was deleting all files in the QUEUE_DIR (not just queued .nzb files). This bug prevented hellanzb from working nicely with shared QUEUE_DIRs (such as your web browser's download dir. This technique was even recommended in hellanzb's OS X install document), in fact, it could cause unwanted data loss (thanks yak) o Quicker and more robust newzbin.com downloads (enqueuenewzbin call) NZBLeecher: o Added duplicate file detection/handling. Previously, NZB files containing duplicate files could cause the downloader to stop (at the end of an archive) under certain situations o Fixed a decoder bug (causing corrupt downloads) affecting some uuencoded posts (and potentially some rare yEncoded posts) o Fixed a yDecode specific decoder bug (causing corrupt downloads) o Fixed a couple nasty bugs in the new downloader code (submitted by bobmckenzie, thanks!) o Fixed "not1not2" KeyError exception in the multiple server pool automatic failover code (submitted by bobmckenzie, thanks again!) o Fixed a bug causing the downloader to stop and the scroller to scroll excessive lines, after a certain situation where hellanzb was paused twice o Correctly handling NZB files with out of order segments, previously this could cause corruption (thanks spinez) o Fixed a bug causing the downloader to stop (at the end of an archive) after pausing/continuing with bandwidth throttling enabled (MAX_RATE set to a non zero value) v0.7: Daemon: o Added the -D option to run the hellanzb queue daemon as an actual unix daemon process (forks and exits) o Changed most of the RPC calls to return the output of a list or status call o Added the enqueuenewzbin RPC call, downloads the NZB with the specified NZB ID from www.newzin.com, as long as you've included your newzbin login information in hellanzb.conf o Changed the RPC list call to show NZB ids by default o RPC status call now shows queued MB, ETA, NZB ids and the value of Hellanzb.MAX_RATE (if MAX_RATE is greater than 0) o Invalid NZB files weren't being removed from the list of current downloads seen in the RPC status call (fixed) o Fixed an odd case where a segment with no articleData could prevent the queue from continuing onto the next NZB. Also fixed a couple other possible bugs that might have caused hellanzb to deadlock in between NZBs NZBLeecher: o Optimized the downloader. Using approximately 2-4% less CPU on an Athlon 2600+, FreeBSD 4.10 o Added automatic failover -- multiple server pools (defineServer lines) will now automatically requeue articles totally missing from their respective servers for redownloading on other server pools (Thanks very much Rubberneck!) PostProcessor: o Fixed the defunct process bug affecting some Linux distributions (thanks georgek) o No longer using /bin/file for identifying rars. Running /bin/file via twisted seems to have triggered a twisted bug where it wouldn't allow the /bin/file process to exit cleanly (on some if not all platforms) o Added assembly support for split .ts files, recognizes patterns such as: hello.001.TS hi.0001.ts hello.002.TS hi.0002.ts hi.0003.ts ArticleDecoder: o Optimized the yDecoder. Benchmarks show it using approximately 6-8% less CPU on an Athlon XP 2600+, FreeBSD 4.10 o Added support for optionally yDecoding via the yenc python module (wraps a yDecoder written in C). This provides better performance than the stock yDecoder. Requires yenc-0.3, available from: http://www.hellanzb.com/hellanzb-content/yenc-0.3.tar.gz v0.6: Daemon: o Added XMLRPC calls to control an already running hellanzb, similar to nzbq and related programs. hellanzb, when invoked with a 'remote-call', will send xmlrpc calls to the main hellanzb process: (From hellanzb.py --help) remote-calls (via XML-RPC): cancel Cancel the current download and move the current NZB to Hellanzb.TEMP_DIR clear Clear the current nzb queue. Specify True as the second argument to clear anything currently downloading as well (like the cancel call) continue Continue downloading after being paused dequeue nzbid Remove the NZB with specified ID from the queue down nzbid [shift] Move the NZB with the specified ID down in the queue. The optional second argument specifys the number of spaces to shift by (Default: 1) enqueue nzbfile Add the specified NZB file to the end of the queue force nzbid Force hellanzb to begin downloading the NZB with the specified ID immediately, interrupting the current download last nzbid Move the NZB with the specified ID to the end of the queue list [showids] List the current queue. Specify True as the second argument to include the NZB ID in the listing maxrate [newrate] Return the Hellanzb.MAX_RATE (maximum download rate) value. Specify a second argument to change the value -- a value of zero denotes no maximum rate move nzbid index Move the NZB with the specified ID to the specified index in the queue next nzbid Move the NZB with the specified ID to the beginning of the queue pause Pause downloading process archivedir Post process the specified directory. The -p option is preferable -- it will do this for you, or use the current process if this xml rpc call fails shutdown Shutdown hellanzb. Will quietly kill any post processing threads that may exist status Return hellanzb's current status text up nzbid [shift] Move the NZB with the specified ID up in the queue. The optional second argument specifys the number of spaces to shift by (Default: 1) PostProcessor: o Multiple par2 file sets are now handled correctly (fixed) o Added config option DELETE_PROCESSED to delete the processed sub directories after successful post processing o Post processing now also occurs on sub directories (and walks down the tree of sub directories) o Post processing will now be automatically resumed upon restart of hellanzb, if previously interrupted (such as via CTRL-C) o Timing unrar/par2 commands o par2 left-over repair ('.1') files now moved to PROCESSED_SUBDIR NZBLeecher: o Fixed bandwidth throttling to be a global value applying to all defineServer lines in the config file. Set the Hellanzb.MAX_RATE value to enable o Supporting Usenet servers not requiring authorization o Better handling of invalid groups (failed nntp GROUP commands) (thanks No1UKnow) o Now overwrites (re-downloads) 0 byte files o Fixed connect timeouts permanently killing connections (blank download status lines, requiring CTRL-C & restart to un-blank) o Quicker automatic reconnections when they are lost ArticleDecoder: o Fixed the 0 CRC problem occurring with some files/servers (e.g.: 'CRC mismatch 00000000 != 3B0B8959') (Thanks drak0 and Weavus) Misc: o Fixed darwin port unrar dependency o Other small bug fixes/changes v0.5: o NZB downloader rewritten from scratch (and some pyNewsleecher code by Freddie ) to use the twisted framework. With new NZB downloader status ticker o Automatic reconnection and anti idle of NNTP server connections o Bandwidth throttling o Can resume interrupted (CTRL-C, etc) downloads, at the segment level o Improved logging