X-Git-Url: http://dxcluster.net/gitweb/gitweb.cgi?a=blobdiff_plain;f=Changes;h=c490cbfa3e680e6b900ff91a3f2cfe8222e5e07a;hb=refs%2Fheads%2Fnewusers;hp=e921da8dadd143aca27e8a8b391a505fa28f96e4;hpb=4a47ee8996cd52566856a06476eb74aea02d68ac;p=spider.git diff --git a/Changes b/Changes index e921da8d..c490cbfa 100644 --- a/Changes +++ b/Changes @@ -1,5 +1,58 @@ +20May20======================================================================= +1. Strip out conversion from users.v2 and v3 to new json format file into a + new program called convert-users-v3-to-v4.pl. In theory, this program + *could* be run at any time and is backported to mojo and master branches. +2. Replace Storable in dxqsl/QSL.pm and rename everything (except QSL.pm). +3. Add CTY-3010 prefixes. +19May20======================================================================= +1. Convert all remaining commands and areas within the program that used the + DB_File/Storable interface to DXUsers.pm to use the (hopefully) more stable + JSON text file + caching interface. Here's hoping. +2. Add branch name to startup and sh/version. +16May20======================================================================= +1. Crontab can now spawn_cmd('export_users') - very safely. +15May20======================================================================= +1, Changed the underlying storage engine from Storable to JSON. Seems to run + much faster! Exporting the user file with 181000 records takes ~5.1secs + with Storable and ~1.5secs. No more thaw() version mismatches! +2. Completed conversion on initialisation + export +10May20======================================================================= +1. Added basic changes so that users *could* have multiple connections to the + same node if it is allowed. This is work in progress and is there to see + if it deals with some networking problems encountered on very high volume + sites which can see more than 1000 users (although the problems can occur + with many hundreds or by the practices of certain ISPs). More information + will be forthcoming if I get to the bottom of what's REALLY going on and + whether this is (or maybe just one part of) the solution. It won't be + ready for general use until then. +2. Fixed the script import error reported by Joaquin (EA3CV?), by the simple + expedient of restoring the version from the master branch. +09May20======================================================================= +1. Show the route by which this PCxx came in progress debugging reports. In + some other words: the spot/wwv/wcy/ann message arrived first from this + connection. +2. Improve progress WWV & WCY messages. +3. Stop random node isolations (at least in one place) +08May20======================================================================= +1. Tidy up routing table +2. Add new argument to show/version (ALL or list of regexes) that allow you + to see the version and build nos of all nodes on the system. +07May20======================================================================= +1. Revert changes made since 25Apr concerning IP address reconciliation. +25Apr20======================================================================= +1. Add maximum no of users on node to show/cluster. +2. Add ability to show last n lines of debugging ring buffer. +3. Remove redundant wpxloc.dat file. +24Apr20======================================================================= +1. Add 'progress' debugging for showing that stuff is happening in nologchan + a.k.a ringbuffer only mode. +2. Fix grepdbg so that no regex argument simply lists the file. +3. Add CTY-3008 prefixes. 22Apr20======================================================================= 1. Fix the module search path in update_sysop.pl +2. Add latest prefixes CTY-3007 +3. Fix unset/startup command. This will only work for users' start up scripts + things like user_default or startup must be edited or removed by hand. 21Apr20======================================================================= 1. Finally fix the "actually tranmit" any output (like from 'logout' files) to users before disconnecting. This also means that disconnecting nodes