[DGD] Kotaka v0.49

[Raymond Jennings]

Whoa!

I just screwed up big time and left a critical inheritable's source
code out of the repository.

Version 0.49 should be completely disregarded in favor of verison 0.49.1

On Sun, Dec 17, 2017 at 10:18 AM, Raymond Jennings <shentino at gmail.com> wrote:
> One important caveat for the future.  Due to a recent discussion on
> the DGD list, it has come to my attention that I made a bad design
> decision when I added constructors and destructors.
>
> For starters, an object is free to rewrite its inheritance tree any
> time it recompiles and inheritables (and for that matter, any daemons
> they interface wiith) cannot rely on staying in the tree, especially
> if they are added after an object has already been created, and thus,
> the constructor doesn't get called.
>
> So in a moment of self reflection, they are going to be deprecated as
> well.  I never should have added them.
>
> Even doing a call_touch bull-rush of sorts doesn't guarantee
> consistency if the object in question completley rewrites the
> inheritance tree.
>
> On Sun, Dec 17, 2017 at 9:55 AM, Raymond Jennings <shentino at gmail.com> wrote:
>> Finally got around to a new release.
>>
>> Major points:
>>
>> I simplified the monitor script
>> some debugging code to help with MSP and mudclients, to help out
>> gnome-mud as well
>> Sitebanned IPs are now hard-forbidden for connection attempts
>> Siteban messages can now be delegated to the connectio manager
>> A full rebuild feature that nukes all inheritables AND orphaned
>> objects without source, and recompiles everything else
>>
>> Caveats:
>>
>> Please read the UPGRADE notes
>> Objects having multiple archetypes is now officially deprecated and
>> will be banned in a future release
>>
>> git://github.com/shentino/kotaka
>>
>> Shortlog:
>>
>>       Forgot to update safe versions list
>>       Version 0.48.1
>>       Simplify the monitor script, all of the script's own signals are
>> exiters anyway.
>>       Add fill handler for clouds
>>       The walk command requires us to know if an object is solid or not
>>       Update copyright dates, muck out old years
>>       Reorganize README
>>       Small update to qsort in case we care about future stable sorting
>>       Add test for sorting library
>>       Add notes about next upgrade nuking and compiling saveload
>>       Destruct saveload during next upgrade
>>       Rewrite saveload
>>       Small bugfix: Handle case in saveload where we attempt to purge
>> an objdir that is already gone
>>       fixup test: Forgot to include a header
>>       Make sure monitor script redirects stderr in the correct directory
>>       When flushing logs, rotate the log BEFORE it hits 1GiB
>>       Remove a global variable from logd
>>       Move cd before exec in hotboot script
>>       More prominent highlight of an aborted hotboot
>>       Remove obsolete command in monitor script
>>       Multiple archetypes are officially deprecated
>>       Clarify what an api change actually means
>>       Elaborate on the deprecation process for multiple archetypes
>>       Rewrite dumpd to be more simple
>>       Set new dumpd to make a dump every hour, with a full one every day
>>       Add symlinks to documentation
>>       Fix includes in archetype deprecation
>>       Don't forget to tell IntermudD to actually listen when we enable
>> I3 on a channel
>>       Have IntermudD listen to channels on bootstrap when a newly
>> discovered channel is registered as I3 on ChannelD
>>       Small cleanup of loop in IntermudD
>>       Remove spurious variable
>>       Have chlist command distinguish I3 channels
>>       Add file utility library
>>       Load TLSD before ObjectD
>>       Load ErrorD in System core
>>       Rename boot_error to log_boot_error in InitD for greater clarity
>>       Disable MSP if told to do so by the client
>>       Add debugging code for MSP
>>       Add debugging code in mudclient for unknown telnet options in general
>>       Add variable for general debug
>>       ...forgot to reenable actually sending the telnet options when I
>> added debugging...
>>       Refine I/O logging a bit
>>       Increase Intermud ticks limit
>>       Reject attempts to connect to a malformed IP address
>>       Forbid attempts to connect to a sitebanned IP address
>>       Rewrite StatusD
>>       Add a full rebuild in System's SubD
>>       /misc/grid.c was an LWO, move it to /misc/lwo/grid.c to restore
>> create function argument list
>>       Don't print a report (or continue the callouts) if StatusD has
>> been disconnected
>>       Don't forget to include account.h, if outbound connections check
>> the siteban list they need access to BAND :P
>>       When doing a full rebuild, only report interesting things, like
>> destruction of orphans or compilations of new non inheritables
>>       Tighten up gaps when reporting sitebans and overloads...and make
>> sure that blacklisted IPs leave plenty of room for good people
>>       Refine logging of inbound and outbound user messages, also
>> remove spammy InitD message
>>       Add ~Text/LogD, to handle the forwarding to LogD
>>       Add fullrebuild verb
>>       Refactor logging code
>>       Save sitebans AFTER ejecting banned connections
>>       When having System::UserD close a sitebanned connection, give
>> the connection manager a chance to handle it
>>       Have Text send a siteban message to a sitebanned connection.
>> Also, sitebans look cooler with bright red text
>>       Restore relation system but stash it in /misc
>>       Simplify Text::InitD
>>       Access control for Text::UserD->siteban_notify
>>       Rewrite call_other interceptor to provide richer debug information
>>       When doing a full rebuild, (re)compile the initds first
>>       Remove old versions from safe_upgrade_versions
>>       Version 0.49