1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
|
# $Id: TODO 316 2008-01-03 20:21:19Z keescook $
Things that need to be done...
Next
----
*Code
- have PCs shutdown on signal if active, but not sending a page.
- clean up stale/running test during BecomeDaemon
- option to not bounce "user not found" errors (to stop domain scanning)
- have a "quiet" startup mode
*Docs
- finish sendpage-manual
- add info about "snpp-acl"
- find counting error for tempfail
- document sendpage-db
- document DBI schema
Next Next
---------
- implement a list of addresses for SNPP server to listen on (for multihomed)
- do something so that queues that are waiting on a modem will run
as soon as the modem is available. How can I do this when there are
multiple modems available? Weird multi semaphore notifications queues
based on modems? Gah. Strange.
- implement "dialretries" (attempt dialing again for PC)
- implement character-based translation tables
- create more fallback PC config options (temp-fail, etc)
Would Be Nice
-------------
*Code
- implement SNPP "HOLD" function
- add warnings when seeing {0x00} chars in modem buffers. Possible incorrect
comm settings 8N1 vs 7E1, etc.
- do something about the 500 functions packed into "sendpage" instead of
cleanly implemented in modules elsewhere.
- implement email notification fall-back for recips
- have all the conf files get handled by Makefile.PL -- /etc dir?
- change PagingCentral module to deal with modems differently so we don't
have to use a KeesConf object.
- change Recipient module to deal with macros differently so we don't
have to use a KeesConf object.
- figure out how to do filelocking under Windows
- figure out what to replace my Syslog calls with under Windows
- add "include" verb to sendpage.cf for sucking in other files? This allows
for separate files for PCs and recip lists.
- implement "fieldsplits" (fields cannot be split across blocks for PC)
- handle calculating parity as described by Shaun Dawson (is this needed
after the parity improvements in Device::SerialPort?)
- TCP modems (open a socket instead of a file) by Jim Gottlieb
- generalize the "xmit message" code to deal with whatever type of sending
we're doing, TAP, UCP, SNPP, etc. (from Jon Meek)
- figure out what the cryptic message "ESC EOT hangups" meant, and fix it
*Docs
- update API doc to reflect changes made for PageQueue, Page, and Recipient
- more general in-code commenting & documentation
- more API documentation for all the modules via POD
- write a data flow graph
Stuff I Thought Of, But Will Most Likely Never Need
---------------------------------------------------
- implement multiple TAP protocols? (are there real differences? PG1, PG3, etc)
I don't think I'll ever need this due to the "pc:fields", and "pc:proto"
variables.
I'm Probably Never Going To Implement These Crazy Ideas
-------------------------------------------------------
- if you have A LOT of pages, you'd want multiple modems running for
a single PC. Since that would REALLY be a lot of pages, I'm going to
leave out the idea of parallel tasking modems for the same PC. That
seems like insane overkill, but it's something that would be pretty
cool for HIGH volume paging centers. Some day, perhaps.
Detailed Notes
--------------
Parity - Shaun Dawson
After a long struggle, I think I've figured out what's going on here, and
have fixed it. Even though the serial port is being set to even parity, I
don't believe that the parity calculations are being done by the system. I
believe that both the incoming and outgoing parity calculations will need to
be done inside sendpage. I added a function to AND out the 8th bit in the
incoming byte stream, and I stopped getting garbage from any of the paging
terminals, but some subset of the ones I had problems with didn't see the
0x0D byte I sent, and so didn't return 'ID='. This led me to conclude that
some of the paging terminals set to even parity just did the same thing that
I did, which was to AND out the parity bit (and those worked fine), but some
others discarded bytes with invalid parity. Since 0x0D has an odd parity,
that byte was discarded, and sendpage was never able to negotiate with those
terminals.
|
