source: server/lib/TODO

* This is a scrap todo file while the new IPP server based Gutenbach is
* being developed.  Once this branch is released and becomes master,
* items in this file will be merged into the Trac.

- finish implementing a bunch of the IPP handlers
     [X] print job
     [X] validate job
     [X] print uri
     [X] pause printer
     [X] resume printer
     [X] set printer attributes
     [X] send uri
     [ ] set job attributes
     [ ] restart job
     [ ] promote job
     [ ] cups get document

- add support in printer.py for:
     [x] print job
     [x] validate job
     [x] pause printer
     [x] resume printer
     [\] set printer attributes
     [\] send uri
     [\] set job attributes
     [X] restart job
     [\] promote job

- add support in job.py for:
     [X] restart job
     [x] resume job

- server stuff
     [ ] write convenience client-side API
     [ ] setuptools entry points for notification system (e.g. zephyr)
     [ ] deal with queue management/reordering
     [ ] add support for volume management
     [ ] support authentication/security

- client-ish stuff
     [ ] rewrite gbr/gbq/etc. scripts to use client API
     [ ] rewrite server-side queue display to use API

- misc
     [ ] make sure all CUPS commands are compatible
     [ ] deal with all the code marked 'XXX'

- documentation and testing
     test cases for printer.py
          [X] TestEmptyGutenbachPrinter
          [X] TestBadEmptyGutenbachPrinter
          [ ] testPrintJob
          [ ] testValidateJob
          [ ] testGetJobs
          [ ] testPrintUri
          [X] testCreateJob
          [X] testPausePrinter
          [X] testResumePrinter
          [ ] testGetPrinterAttributes
          [ ] testSetPrinterAttributes
          [ ] testCancelJob
          [ ] testSendDocument
          [ ] testSendUri
          [ ] testGetJobAttributes
          [ ] testSetJobAttributes
          [ ] testRestartJob
          [ ] testPromoteJob

     test cases for requests.py
          [ ] testPrintJob
          [ ] testValidateJob
          [ ] testGetJobs
          [ ] testPrintUri
          [ ] testCreateJob
          [ ] testPausePrinter
          [ ] testResumePrinter
          [ ] testGetPrinterAttributes
          [ ] testSetPrinterAttributes
          [ ] testCancelJob
          [ ] testSendDocument
          [ ] testSendUri
          [ ] testGetJobAttributes
          [ ] testSetJobAttributes
          [ ] testRestartJob
          [ ] testPromoteJob
          [ ] testCupsGetDocument
          [ ] testCupsGetDefault
          [ ] testCupsGetPrinters
          [ ] testCupsGetClasses

     [ ] write test cases for future client API
     [ ] all the documentation
     [ ] check test case coverage ( http://nedbatchelder.com/code/coverage/ )
     [ ] why doesn't the bug tracker work?! relatedly, figure out a
         solution that makes it easier for non-MIT developers to contribute

Ponies
   - gapless playback
   - display video
   - youtube
   - pandora
   - last.fm
   - local streaming
   - generic streaming
   - zephyr control (e.g. youtube link)

Tickets that should be fixed by this new version:
   [ ] http://gutenbach.mit.edu/ticket/21
   [ ] http://gutenbach.mit.edu/ticket/2
   [ ] http://gutenbach.mit.edu/ticket/7
   [ ] http://gutenbach.mit.edu/ticket/39
   [ ] http://gutenbach.mit.edu/ticket/47
   [ ] http://gutenbach.mit.edu/ticket/48
   [ ] http://gutenbach.mit.edu/ticket/8
   [ ] http://gutenbach.mit.edu/ticket/34
   [x] http://gutenbach.mit.edu/ticket/38
   [ ] http://gutenbach.mit.edu/ticket/41
   [ ] http://gutenbach.mit.edu/ticket/18

 === CONSIDERATIONS FOR MPD ===

We are considering using MPD (the Music Player Daemon) for playback, instead
of mplayer.  This gives us a pony (gapless playback), and makes some other
things, like keeping track of the queue, much easier.  There is a lot of work
needed to make this work, though.

THIS ADDS A DEPENDENCY:  python-mpd

We will have to re-implement all the functions in jobs.py, and completely
re-structure player.py

If we want video playback eventually (and we do), there will be some hackery
involved.  We will pause mpd playback, do the video playback completely
separately, and then resume mpd playback.

Some almost-pseudocode:

startup:
  client = MPDClient()
  client.connect(**{'host':'/var/run/mpd/socket', 'port':'6600'})

adding a song to the queue:
  Receive job
  Put it into a file                    [gutenbach/FILENAME]
  Tell MPD to add it to the queue       [client.addid('gutenbach/FILENAME')]
  Find out what the id is
  Store all the data [including the temporary filename]
  IF WE ARE SUPPOSED TO BE PLAYING, MAKE SURE WE *ARE* PLAYING
    In particular, if the queue was empty, start playback   [gutenbach.play()]

when a job completes playing:
  Remove the file                       [rm gutenbach/FILENAME]

getting the queue:
  get playlist, parse return            [client.playlistid()]

removing a job:
  dequeue it                  [client.deleteid('NUMBER')]
  remove the file
  tell them what you did

restart song:
  client.seek(0,1)   [0% through song 1...]