X-Git-Url: http://dxcluster.org/gitweb/gitweb.cgi?a=blobdiff_plain;f=Changes;h=90b26c80d3c53dae979a456aae78672511b4094e;hb=6e8ea6c3fb2e0702e616a6553629905ca6ae22c1;hp=b77d74c05a280ccf116805c0bf49b382635285b6;hpb=2b58ccdf81685a1167a43c38705a0d84b9d8d661;p=spider.git diff --git a/Changes b/Changes index b77d74c0..90b26c80 100644 --- a/Changes +++ b/Changes @@ -1,3 +1,163 @@ +09Oct01======================================================================= +1. change Aliases to allow shutdown as a command again. If you have changed +/spider/cmd/Aliases and you update from CVS you may find it better to remove +this file and then do another 'cvs update'. Otherwise you may get conflicts +and the Aliase file will not load properly. The cvs update will recreate +the (now) missing file. +2. Remove the 'please don't use XXXX on here' messages to prevent people +from exploring the 'bad words space'. Three bad words will still disconnect +without warning. +3. Sending dx spots/announces with bad words will succeed - but only back to +the sender. These two mods are designed to annoy the perpitrators of the +sort of filth as much as possible whilst not inconveniencing 'normal' users. +4. Add expunge feature to kill. Deleting messages now marks a msg for +deletion two days hence. To have the old behaviour, do a 'kill exp' or use +the standard alias "expunge". This will cause the message to be deleted at +the next queue message cycle. +5. Put in prototype echo handling for direct connections using ExtMsg and +Listeners.pm. +6. fix msgs to g1tlh-2 (as opposed to g1tlh) +04Oct01======================================================================= +1. allow a residence time of (default) 2 days before actually deleting a msg. +2. reduce the number R: lines on a bull from a bbs to 2 (the origin and the +last port of call). +3. ignore deleted messages for queuing purposes +4. store deleted status across restarts! +5. make callsign checking more rigorous +6. dup check PC49 (kill full) +7. improved the regex matching of badwords (more efficient, better coverage) +8. added default badword and badw_regex tables (as .issue files) which will +activate unless there is one there already. It will NOT overwrite an existing +file - so if you change it your changes will remain intact. +03Oct01======================================================================= +1. don't allow @WWW to become a 'TO' field... +2. handle @gb7tlh.#35.eu type addresses as well +02Oct01======================================================================= +1. fixed the 'ever lengthening msg' syndrome and probably made the whole +thing more reliable at the same time. +2. do some restructuring of the msg system and upissue the version no. +3. fix @xxxx syntax for bbs input +01Oct01======================================================================= +1. made login info not the default +2. change the looping protoection for PC16/17/19/21 and also be more rigorous +in who we accept which PC16/17/21's from. +3. fix MAJOR disconnect bug introduced accidentally in the NP work. +4. Try each badword with an S on the end as well and also check for 'f.u c' +type things as well (this only works for a few wellknown english ones). +5. Go back to the old way of caching commands (but without the silly warnings +because I know better what I am doing now). This allows symbolic debugging +of commands again without coredumps. Hurrah! +6. Add badword handling for messages. Users will get rude messages back if +they try to send naughty words in msgs. Incoming messages with badwords will +be dropped on receipt and their contents logged. +7. and if you swear 3 or more times you are unceremoniously logged out. +30Sep01======================================================================= +1. made some small bug fixes in rspf checking and also messages. +23Sep01======================================================================= +1. I have incorporated Steve's (K9AN) lastest sh/sun and sh/moon stuff which +copes much better with some of the end effects that occur in these calc- +ulations. You can now also see up to a year into the past or future. The +results compare favourably with the US Naval Obs (unlike certain other +software's output which "I would not want to use their results to aim my +moon rocket..."). It also fixes some ambiguity about which day the results +are actually for. +21Sep01======================================================================= +1. fix set/lockout so that it is possible to lock out all SSIDs except those +specifically unlocked and so that you don't need to lock the non-SSID call in +order to lock an SSID call. So set/lock g1tlh will lock out all instances of +g1tlh, g1tlh-1, g1tlh-15 etc except (for instance) unset/lock g1tlh-9. +2. show/lock allows partial callsign matching so sh/lock gb7 will only show +GB7* calls that are locked. +3. Had a grand shift around for the start of NP. +19Sep01======================================================================= +1. put in some rudimentory rsfp checking for various things +2. tried to do some fixes on console.pl - YOU WILL REQUIRE Curses 1.06 from +CPAN (perl -MCPAN -eshell, then at cpan> install Curses). +17Sep01======================================================================= +1. Fixed help and apropos commands so that they default to english if +the language is not english and help in the 'not english' language isn't +available. +2. added some help text for sh/newc. +15Sep01======================================================================= +1. released 1.48 +2. found 1st bug, fixed unset/logininfo reseting itself after logout +14Sep01======================================================================= +1. added the start of script files on login/startup. You can now add +commands (for example to set variables) to a file called +/spider/scripts/startup. You can also put in scripts for a callsign which +will be executed when that callsign logs on. A sample script is in the +directory (you did remember to do cvs update -d?). User settable versions +of this will happen when the general mechanism for doing that is available. +2. Added set/unset/anntalk commands. If you unset/anntalk then you will +no longer see many of the of talk like announces. Unfortunately you may +miss the odd useful one as well (but since everything is repeated many times +you probably won't :-) +3. centralise the invalid callsign check so that g1tlh1 and similar calls +will not be allowed to login anymore. +4. Fiddle with the lockout mechanism so that set/login g1tlh also locks out +g1tlh-1 -> g1tlh-15 UNLESS one of these is specifically unlocked. +5. Make sh/log only show stuff that sh/ann, sh/talk, sh/rcmd doesn't. +6. Make the scripts look for user and node_default files if there isn't a +more specific one. +7. add the blank command which prints nn blank lines (default 1), this cmd +can also print lines of characters: see help for examples. +13Sep01======================================================================= +1. did some work on making talk more intelligent and fixed a>b problem. +2. fixed a nasty problem on input when being hit with full buffers of +data (eg at init time with large lists of node/users on fast links). +3. fixed realtime input filter changing. +4. added announce->talk conversion for routable calls when announces of the +form 'to g1tlh hello', 't g1tlh hello' or 'g1tlh hello' appear. +This also suppresses similar announces for users whose callsign is not the +one in the announce. +5. Change the badwords interface to be the same as baddx, badspotter et al. +added set/badword, unset/badword and show/badword. This routine will auto +convert (and delete afterwards) the old badwords file. +6. fix intermittant bullitin mail forwarding? +11Sep01======================================================================= +1. added IP address logging of connections +10Sep01======================================================================= +1. clear/* node_default and user_default now work. +2. in acc or rej/* you can now use (comma separated) prefixes as well as lists +of numbers for things like *_dxcc, *_itu and *_cq (eg: by_dxcc ve,w); just +remember that countries like ve and w have lots of zones and itus! +3. Added all the clear/* commands to help and added some comments about 2. +4. All filters (including *_default ones) are now reinstalled on channels in +realtime, so changing a filter does NOT require a disconnect and reconnect +cycle anymore. +09Sep01======================================================================= +1. Add show/dxstats command +2. Add show/hfstats command +3. Add show/vhfstats command +4. Add show/hftable command +5. Add show/vhftable command (BE WARNED, the first time you run one of these +commands it will take some considerable time to complete, thereafter it will +be a lot faster). +08Sep01======================================================================= +1. Update the sgml docs on filtering a bit. +06Sep01======================================================================= +1. prevent nodes appearing as users from incoming PC16s +2. Change ping averaging to be like TCP and become a smoothed RTT. +3. Make all the Prompts in stat/* like output appear in alphebetical order +4. Fix problem with 17m spots not appearing in sh/dx +5. pump up eph dups everytime they re-occur and increase life to 180 secs +04Sep01======================================================================= +1. add delete/user command +03Sep01======================================================================= +1. make spot dups look back 5 mins. +2. strip the top off each character in any text field of a Spot for duping. +3. Try to make the routing tables more accurate for users that login to +more than one node. +4. Make PC50s come out in one heap on all channels every 14 mins, instead of +on the 14th minute in the connection time for each channel. This should +reduce (slightly) the dups that are dumped. +5. Speed up input queue processing (a lot). +6. make set/isolate and acc/route mutually exclusive (and issue appropriate +error messages) and fix set/iso on PC18 input. +7. Improve stat/route_node and stat/route_user for diagnostics. +8. strip top bit of comments in ann/spots. +9. fix the announce filtering problem. +10. Fix the double announcement of node disconnections? 01Sep01======================================================================= 1. Change build number calc (hopefully for the last time) 27Aug01======================================================================= @@ -9,7 +169,8 @@ privilege is 1 or less and which isn't mentioned as an argument to the command on the command line. 3. make set/node, set/spider and their friends unlock a node as well as make them one. -4. Make sh/log et al more efficient / less memory hungry +4. Make sh/log et al more efficient / less memory hungry (and the display +the correct way round!) 24Aug01======================================================================= 1. Allow badmsg to reject on interface callsign ('I') 20Aug01=======================================================================