X-Git-Url: http://dxcluster.net/gitweb/gitweb.cgi?a=blobdiff_plain;f=Changes;h=167d28b43ecc242414fcdcf14f8b2c97b9db4417;hb=refs%2Fheads%2Fnewpc92timings;hp=ded4c90ee1004ef0c7a2e294fc5a823a634b0d7e;hpb=ac14c6983861c08b8a2842af46e67407600f8065;p=spider.git diff --git a/Changes b/Changes index ded4c90e..167d28b4 100644 --- a/Changes +++ b/Changes @@ -1,8 +1,90 @@ -24Jun06======================================================================= +15Sep07======================================================================= +1. Add *optional* INET6 connectivity. In order to use this you must load +IO::Socket::INET6 and its dependencies from CPAN or get the distro's packaged +versions. I used 'sudo aptitude install libio-socket-inet6-perl' for ubuntu +(which should work for debian as well), for fedora/rpm based systems +'sudo yum install perl-IO-Socket-INET6' or near equiv should work. This has +been done quite simplistically, but it seems to work. +10Sep07======================================================================= +1. Remove warnings for EINPROGRESS etc for Windows perl 5.8. +2. Try to see if using just one lastid is viable in all situations. If it is +then this will cure the talk problem. +20Aug07======================================================================= +1. add cty.dat + wpxloc.raw changes for CTY-1706. +31Jul07======================================================================= +1. extend 60m band slighty and add ssb section. +2. add a new cty.dat (CTY-1705) from Jim AD1C. +27Jul07======================================================================= +1. Added Angel's changes to Messages for Spanish (finally). +25Jul07======================================================================= +1. don't clear out routing tables on shutdown. This should speed that up. +24Jul07======================================================================= +1. Try to target PC93 (the combined talk/ann/chat sentence) better. Prevent +PC12 chat going to non-dxspider and non-ak1a nodes. None of the others seem to +handle it correctly (Lee VE7CC: do you?). +2. a few other tidyups, +see: http://www.dxcluster.org/cgi-bin/gitweb.cgi?p=spider.git +for details. +18Jul07======================================================================= +1. Change the meaning of set/isolate. Set/isolate now works in both directions +in that an isolated node will only have its local config remembered and will +only be sent this nodes local config. This is exactly equivalent to both +ends setting set/isolate. If that is already the case then there should be +no noticeable change. Any extra that the far node sends, over and above its +local config will now be ignored. +17Jul07======================================================================= +1. remove dupefile on startup and on ending the node. Add clear/dupefile +command to allow people to do it at runtime as well. +12Jul07======================================================================= +1. Change disconnection code so that nodes that are no longer routable are +(all) cleared out. +2. Add help for sh/band and also allow query of individual band or regions. +11Jul07======================================================================= +1. improve speed of sh/c/n +2. remove all $Id$ strings from cmd tree +09Jul07======================================================================= +1. remove check for PC93 from legacy nodes. This will allow PC10s to be +propagated correctly. Probably. +08Jul07======================================================================= +1. put in a temporary fix to the crashing on incoming PC10 talks, whilst I +try to work out how it is happening. +06Jul07======================================================================= +1. re-arrange id checking so that it is more logical and check that config +records, that can be an hour coming, are not thoughtlessly ignored, thus +causing nodes to time out. +2. Make sure that the 'over midnight' distance calculation actually does what +is required! +3. Tart up version number reporting in sh/node. +4. Pass the correct number of parameters in pc93 -> pc12 conversions which +should stop returning PC12s coming out as SYSOP announces. +05Jul07======================================================================= +1. Try to make sure that the node config broadcasts are actually sent... +03Jul07======================================================================= +1. fix unwanted dupe notifications if a PC12 comes back in on a loop. +02Jul07======================================================================= +1. Add a new command to show the contents of the cmd_cache. This allows those +people whom develop their own commands to see which version they are using. +It is called show/cmd_cache (or sh/cmd for short). +2. Add some extra info to the links command. Show whether a link is isolated +and also whether it has filters (if applicable) and whether these are +personalised ('Y') or the default node_filter ('D'). +29Jun07======================================================================= +1. If a node is set/isolated then make sure that a) pc9x is not advertised +and b) pc9x is ignored. +2. Add (un)set/wantpc9x command (but please don't use them without talking +to me first), here be dragons!! +25Jun07======================================================================= +1. make sure that a C record is sent for node call every update period. +2. make announces work again (probably). +3. Fix long standing possibility of a crash with a new user. +4. Add continuous chatting (like talk with no message eg 'chat #9000' just on +its own with no message). +24Jun07======================================================================= 1. Fix the routing algorithms to allow route selection in the face of incomplete trees of nodes (trees as produced by sh/newc). -2. Fix the obscout issues which mean that many nodes get timed out after 3hrs. -23Jun06======================================================================= +2. Fix the obscount issues which mean that many nodes get timed out after 3hrs. +3. Extend 4m to 70631 to cover CT. +23Jun07======================================================================= 1. fix basic deduping algorithm 2. rearrange node startup protocol to avoid issuing unnecessary PC92 and also to spread the PC92Cs out for dependent nodes. @@ -12,24 +94,24 @@ to spread the PC92Cs out for dependent nodes. old style protocol and carry on, rather than disconnecting (protocol facist that I was). 6. Fix isolated pc9x nodes so that they remain in oldstyle protocol. -22Jun06======================================================================= +22Jun07======================================================================= 1. Fix Alias for sh/mydx (so it isn't the same as sh/myfdx). 2. Fix sh/c/n uninitialised variable message 3. Fix (and improve) sh/newc gb7 problem 4. Fix broken set/sys_qra command -21Jun06======================================================================= +21Jun07======================================================================= 1. merge back SIMPLEROUTE branch to issue as official 1.54. -18Jun06======================================================================= +18Jun07======================================================================= 1. add a optional dependency on Encode (included in 5.8.x) to encode strings to latin1 for deduping purposes, hopefully getting rid of some dupes. 2. add a default INPUT filter for a node (by_dxcc ) if no specific or default INPUT node exists. -13Jun06======================================================================= -1. start using git. -2. change all the version / build numbering. 14Jun07======================================================================= 1. prepare for git repository and moving of anon cvs repository to scm.tobit.co.uk. +13Jun07======================================================================= +1. start using git. +2. change all the version / build numbering. 11Jun07======================================================================= 1. Change the frequency normalisation for DX Spot dupe checks so that any decimal part is thrown away (in other words: truncate the freq to integer khz)