"Fossies" - the Fresh Open Source Software Archive

Member "sendpage-1.001001/TODO" (3 Jan 2008, 4339 Bytes) of package /linux/privat/old/sendpage-1.001001.tar.gz:


As a special service "Fossies" has tried to format the requested text file into HTML format (style: standard) with prefixed line numbers. Alternatively you can here view or download the uninterpreted source code file.

    1 # $Id: TODO 193 2005-03-11 23:12:18Z keescook $
    2 
    3 Things that need to be done...
    4 
    5 Next
    6 ----
    7 *Code
    8 - have PCs shutdown on signal if active, but not sending a page.
    9 - clean up stale/running test during BecomeDaemon
   10 - option to not bounce "user not found" errors (to stop domain scanning)
   11 - have a "quiet" startup mode
   12 *Docs
   13 - finish sendpage-manual
   14 - add info about "snpp-acl"
   15 - find counting error for tempfail
   16 - document sendpage-db
   17 - document DBI schema
   18 
   19 Next Next
   20 ---------
   21 - implement a list of addresses for SNPP server to listen on (for multihomed)
   22 - do something so that queues that are waiting on a modem will run
   23   as soon as the modem is available.  How can I do this when there are
   24   multiple modems available?  Weird multi semaphore notifications queues
   25   based on modems?  Gah.  Strange.  Only the PCs know about modem use,
   26   not the queue manager.  Perhaps another pipe for modem reservations?
   27 - implement "dialretries" (attempt dialing again for PC)
   28 - implement character-based translation tables
   29 - create more fallback PC config options (temp-fail, etc)
   30 
   31 
   32 Would Be Nice
   33 -------------
   34 *Code
   35 - implement SNPP "HOLD" function
   36 - add warnings when seeing {0x00} chars in modem buffers.  Possible incorrect
   37   comm settings 8N1 vs 7E1, etc.
   38 - do something about the 500 functions packed into "sendpage" instead of
   39   cleanly implemented in modules elsewhere.
   40 - implement email notification fall-back for recips
   41 - have all the conf files get handled by Makefile.PL  --  /etc dir?
   42 - change PagingCentral module to deal with modems differently so we don't
   43   have to use a KeesConf object.
   44 - change Recipient module to deal with macros differently so we don't
   45   have to use a KeesConf object.
   46 - figure out how to do filelocking under Windows
   47 - figure out what to replace my Syslog calls with under Windows
   48 - add "include" verb to sendpage.cf for sucking in other files?  This allows
   49   for separate files for PCs and recip lists.
   50 - implement "fieldsplits" (fields cannot be split across blocks for PC)
   51 - handle calculating parity as described by Shaun Dawson (is this needed
   52   after the parity improvements in Device::SerialPort?)
   53 - TCP modems (open a socket instead of a file) by Jim Gottlieb
   54 - generalize the "xmit message" code to deal with whatever type of sending
   55   we're doing, TAP, UCP, SNPP, etc. (from Jon Meek)
   56 - figure out what the cryptic message "ESC EOT hangups" meant, and fix it
   57 *Docs
   58 - update API doc to reflect changes made for PageQueue, Page, and Recipient
   59 - more general in-code commenting & documentation
   60 - more API documentation for all the modules via POD
   61 - write a data flow graph
   62 
   63 
   64 
   65 Stuff I Thought Of, But Will Most Likely Never Need
   66 ---------------------------------------------------
   67 - implement multiple TAP protocols? (are there real differences?  PG1, PG3, etc)
   68   I don't think I'll ever need this due to the "pc:fields", and "pc:proto"
   69   variables.
   70 
   71 
   72 I'm Probably Never Going To Implement These Crazy Ideas
   73 -------------------------------------------------------
   74 - if you have A LOT of pages, you'd want multiple modems running for
   75   a single PC.  Since that would REALLY be a lot of pages, I'm going to
   76   leave out the idea of parallel tasking modems for the same PC.  That
   77   seems like insane overkill, but it's something that would be pretty
   78   cool for HIGH volume paging centers.  Some day, perhaps.
   79 
   80 
   81 Detailed Notes
   82 --------------
   83 Parity - Shaun Dawson
   84   After a long struggle, I think I've figured out what's going on here, and
   85   have fixed it.  Even though the serial port is being set to even parity, I
   86   don't believe that the parity calculations are being done by the system.  I
   87   believe that both the incoming and outgoing parity calculations will need to
   88   be done inside sendpage.  I added a function to AND out the 8th bit in the
   89   incoming byte stream, and I stopped getting garbage from any of the paging
   90   terminals, but some subset of the ones I had problems with didn't see the
   91   0x0D byte I sent, and so didn't return 'ID='.  This led me to conclude that
   92   some of the paging terminals set to even parity just did the same thing that
   93   I did, which was to AND out the parity bit (and those worked fine), but some
   94   others discarded bytes with invalid parity.  Since 0x0D has an odd parity,
   95   that byte was discarded, and sendpage was never able to negotiate with those
   96   terminals.
   97 
   98