Hello Kees!
Thursday January 01 2015 11:58, Kees van Eeten wrote to Andrew Leary:
AL>> Good. I have not been able to reproduce the circumstances that
AL>> led to the corrupted segment file. However, I am looking at the
AL>> logic for handling update files, to ensure that the segment in
AL>> the master directory is not overwritten until the update is
AL>> validated as correct.
KvE> Although I applaud that the update is validated, be aware that that
KvE> requires that the update is generated by makenl. Some RC's collate
KvE> their segement by other means and do not provide a valid CRC in the
KvE> first line.
By validated as correct, I mean processed without error. ie: no duplicate
numbers, invalid baud rates, phone numbers with too few parts, etc. MakeNL
does not require segments to have CRC entries in the first line. I am
considering adding FSC-0055 (password protected updates) support to MakeNL, and
while I'm in there it's probably worth checking the CRC value if present, and
rejecting the file if either the CRC or password doesn't match.
Andrew
---
* Origin: Bits & Bytes BBS * V.Everything! * 860/535-4284 (1:320/119)
|