* initial commit with horrible code dupcliation haha lol
Set up elapsedtimer to generate wait# stuff between packets recorded
* implement groundwork for internal demo server
* add core playback functionality
* make it work kinda by including SC packet in demo
* Add a file dialog for loading a demo file instead of a hardcoded path
* Change /play to > in OOC to begin playback or skip to next element
Pop up file dialog box *before* establishing the connection, not after.
TODO:
* Fix having to load the same file *twice* to be able to connect to the demo server for some reason
* Fix the segfault caused by calling the playback() function when there's no remaining data, it's almost like demo_data.isEmpty() is ignored for whatever reason???
* Clear demo data when loading a demo file to prevent stacking demos
* Properly disconnect the client when sending the request for file browser fails to obtain a good demo file
* Fix append_to_file newlining even if file didn't exist prior to calling this function
* Add a very scuffed exception to not log or demo record anything that happens in the Demo playback local server
* Reduce invalid file spam by checking for non-logging server better
Use file_exists() to be more consistent in append_to_file
* Fix the client crashing when receiving loading packets etc. at runtime such as SC, SM, CI etc. (TODO: parity???)
Remove useless debug stuff
* Preserve newlines for demo packets such as "CT", "MS" etc.
* Implement /max_wait, /min_wait for adjusting the maximum and minimum wait in milliseconds between wait packets
Add /pause or | shorthand to pause playback
Re-add /play and keep > as a shorthand
Remove clientside restrictions from sending empty OOC messages and sending OOC messages without a name - these should be serverside.
* Empty music list
Default the character to Spectator char no matter the selection in CSS
* Allow -1 character ID or character ID that does not fit into the local Character Select Screen list to still be parsed corectly, using the character folder in the MS packet as reference.
Allow servers with no selectable characters to still be properly loaded
Bypass the Character Select Screen when joining a server with no character select screen and automatically become a spectator
* Properly handle demo files without SC packet to dictate which chars exist
Add a /load command letting you load a demo file without rejoining the demo server (the CS packet will not be properly handled but I cannot think of a single tangible problem this causes lol)
Make sure all DEMO CT messages are colored properly
* Fix logs bleeding into each other if you disabled logging or you joined a demo server after leaving another
* Prevent logging even if log_filename is defined because a user might disable auto logging when running the game
* Fix custom shownames not appearing in the IC logs
* Set up new logic for max_wait, meaning that dead air being skipped is prioritized with MS (IC chat) packets being used as the anchor.
* Better logic for min_wait to only affect important packets (IC Chat)
* Fix encoding not being performed on packets that are saved to the .demo file, resulting in characters having something like "#1" in their message breaking that specific message
* Fix a weird setup in courtroom.cpp that happened out of the merge and bugged the code
Fix aopacket being busted up as well
* add missing feature flags to the demoserver
* use random port
* move writing to the demo file to a function
* only listen on localhost
Co-authored-by: oldmud0 <oldmud0@users.noreply.github.com>
* remove copypasta
* add a help to the demo server
* fix empty demo disconnecting the server
* tell the user how to begin
Co-authored-by: scatterflower <marisaposs@gameboyprinter.moe>
Co-authored-by: stonedDiscord <10584181+stonedDiscord@users.noreply.github.com>
Co-authored-by: stonedDiscord <stoned@derpymail.org>
Co-authored-by: oldmud0 <oldmud0@users.noreply.github.com>
* Begin trying to understand handle_chatmessage monstrosity
Rearrange some functionality to be more consistent with what the particular handle_chatmessage is *supposed* to be responsible for
Single responsibility principle doesn't apply not just to script files but even functions in this goddamned program apparently.
* Fix some oversights relating to "has presented evidence" logs
* Fix an edge case crash with negative evidence, idk why this didn't happen before.
* Completely refactor handle_chatmessage hell and split it up into single-responsibility-principle functions like a sane and organized coder would.
This will break any PRs that touch handle_chatmessage and will have to conform to the new standard, but it's very good in the long run.
Functions have also been meticulously commentated to make sure that everyone can understand at a glance what everything does.
This refactor is done to pave way for smooth viewport, making characters no longer randomly interrupt each other when speaking.
* Rename confusingly named "handle_chatbox" to "handle_ic_speaking"
Add a new function "handle_callwords"
* Parse blankposts in the packet unpacking phase
log_chatmessage is now an argument-based funciton instead of using m_chatmessage monstrosity
Commentate the heck out of log_chatmessage
Fix showname displaying logic being bonkers for log entries
Fix -1 char_id most likely crashing client, even tho they're supposed to be system messages
Fix [CHARID] being referenced instead of [OTHER_CHARID] in the pair display func
Fix chat message color being used for "has played a song" even though it shouldn't
* Move setting emote_mod from the objection thing to the packet unpacking thing
handle_emote_mod has become argumented
* Make display_pair_character take in only vars it needs
* Add functionality for message stacks (does not handle logging separately yet)
Currently segfaults at line 1823 if stacking behavior actually occurs. Plz help I'm poopie brain at this
* Make sense of pointers and dereference some pointers to prevent wacky errors
Replace _stack with _queue, and turn it into a QQueue
* Implement a minimum text delay queue timer so that each msg must stay on screen for a specified amount of time
Chat_arrow only shows up when the minimum text delay has expired
Check if target is muted in the enqueueing, as well as the validity of the character id, so we don't even bother adding it to the queue
Add a settings option to set text stay time (minimum amount of time to pass before displaying the next message).
TODO: Handle queued interjections in a cooler way
* Adjust default option for stay_time to be 200ms
Make the stay_time setting of 0 disable the queue system entirely
* Make interjections force-clear the message queue
Use interjections to instantly process those messages - objections are usually extremely important and must have everyone's immediate attention.
TODO: Add a settings option for interjection behavior
* Clear message queue when bg display packet is received
* Add an instant objection setting which, when enabled, will interrupt and skip the message queue on objections (effectively bypassing text_delay system)
* Stop using AOPacket for the queue and use QStringList instead (packet contents, we don't care for the header)
Implement objection threshold which works in a queue if instant objections are disabled (not exposed to the settings yet)
Modify packet contents directly instead of waiting for m_chatmessage shenanigans (we should deprecate this soon)
Rename isinstant_objection_enabled options check to be more consistent
* Make it so objections display as soon as the message is over if instant_objection is disabled
* Remove emote_mod = "1" thing that had no real purpose
Implement desynchronized logs settings option, which is false by default, meaning users' messages and display IC logs are going to be in sync.
Split up display and logging behavior, meaning IC logs will be saved with accurate timings on the .logs file, while display can be slower.
Remove dumb "System" showname thing cuz it was dumb
* Bring back minimum packet size safety check
* Make sure shouts are properly logged. I am very sorry for this horrible amount of code duplication though.
* fix compile errors
* fix html parsing error
* Rearrange the order of logging so IC messages go like this:
* shouts go first
* evidence goes second
* chat message goes last
* Fix No Interrupt coming back, go away you foolish beast
* fix borked options dialog
* Crank down "press enter delay" from 600ms to 200ms so it's not as harsh
Co-authored-by: in1tiate <radwoodward@vikings.grayson.edu>
Co-authored-by: oldmud0 <oldmud0@users.noreply.github.com>
Implement scalable maximum clock count, right now it's at 5 clocks a theme can have max
Theme "clock_" starts from 1 instead of 0 since users don't know when stuff starts at index 0
TODO: testing lol
Theoretically fixes#313
The "FA" packet from the server invalidates the client's previous area
list so it makes sense for the arup information tied to that list to
be invalidated as well (especially seeing that the area and arup share
the same index). As we repopulate the client's area list, we do the
same for the arup list.
Implement clock setting w/o starting or stopping
Both of these should make it possible for the server to start/stop/pause/resume the clock with perfect synchronization to the true time.
Display ping in the application window title.
keepalive timer now fires every second instead of every minute
Remove meme clock starting on set_widgets()
implement get_ping() on w_courtroom
* Fix colors in ic logs not parsing HTML properly when log_goes_downwards=false and colorlog=true
* Restore KFO branding
* Reset the settings values when receiving the 'FL' packet. This allows enabling/disabling features at runtime, such as ARUP system.
* Revert "Restore KFO branding"
This reverts commit 7f19b6257a0d75ec1d392b9b8c02a384c8dba43d.
Apparently this had never been staged, and while everyone was having
problems with the connect button, everything was working just fine for
me.
This enables the connect button a bit earlier in the handshake process
than in 2.6, since there are some older servers out there that don't
send FL (and so the connect button would never get enabled). This might
lead to race conditions, and sometimes the progress bar does hang,
but you can just cancel and reconnect - I haven't encountered a crash
so far.
Due to a countless number of changes made to the core that were not
fully understood, tested, or documented, it was decided to roll
everything back to the last known stable version (2.6.2).
Changes dropped include:
- Witness needed
- Shake
- Frame SFX
- Multiple custom objections
- Multithreaded thumbnail generation
- Looping
- Various translation additions
- "Mirror IC"
- Color in IC log
- An invocation of clang-format
Next time, work together and split your big fork into independently
testable feature branches.
{ BasedOnStyle: LLVM, UseTab: Never, IndentWidth: 4, TabWidth: 4, BreakBeforeBraces: Stroustrup, AllowShortIfStatementsOnASingleLine: false, IndentCaseLabels: false, ColumnLimit: 0, AccessModifierOffset: -4, NamespaceIndentation: All }
(this is the Visual Studio preset with only "BreakBeforeBraces" changed from Allman to Stroustrup)
Indentation fixed to 2 spaces per tab. Braces set to Stroustrup style.
Lines reflow at 80 characters. One-line method bodies are on the same
line as the signature. Space always after `//`. No indentation
on preprocessor macros. Includes are sorted lexicographically.
If you don't want to see this commit on blames, use the hidden
whitespace option on GitHub, or use `-w` in git-blame.
CR likely stands for "CentsRaidensnake." Like the Case Cafe mega-merge
before it, this was not a clean merge, and it had to be split up into
two parts: the actual changes, and the attempt it made to reformat the
entire code via clang-format.
This branch had a complicated set of changes that would be difficult to
describe in this commit message. It would be better described in a
proper changelog.
