Reworked the emote parsing to properly handle "wrapped" emotes. A
wrapped emote is one that is wrapped in colons or square braces (eg,
:Kappa: or [Kappa]). This allows emotes to be input without a space
between them, like is required with non-wrapped emotes.
A new configuration setting has been added to only allow parsing of
wrapped emotes: "WrappedEmotesOnly". This defaults to False as to not
break current configurations.
The emote autocompletion will only insert non-wrapped emotes. Setting
"WrappedEmotesOnly" configuration value to True will not change this.
Lastly, made the bare-word emote parsing a bit cleaner by removing a for
loop in favor of a map lookup with a check. This should in theory be
more efficient.
This change is in response to #111. The issue should not be considered
resolved until the autocompletion handles the new setting correctly.
It's ugly, but it get's the job done. This list will need to be
improved a lot (ie, make it searchable and maybe even clickable
emotes?). This is for #99, but should not be considered fixed yet.
Don't reload emotes directly into the global common.Emotes variable.
Instead, load them into a new variable and write that to common.Emotes
after the search for new emotes completes. Emotes should no longer
duplicate for each run of `/reloademotes`.
Also tweaked `/addemotes` to automatically reload emotes if nothing went
wrong downloading a new set.
Fix name highlighting when emotes are involved. When the emotes were
moved to subfolders the channel name was put into the URL of the emote
image. If a user with that name joined the chat, all of the emotes in
that folder highlighted the name in the URL, breaking the <img> tag.
This change removes the regex used to identify and replace the user's
name and instead works on whole words delimited by spaces.
This reworks how emotes are cached in relation to their physical
location on disk. Functionally, they should be the same from the user
perspective but it sets up some stuff that will make it easier to add
emotes from various sources.
Added a few more stats to keep track of:
- Max users in chat
- Stream start
Also added a `/stats` command to list the user count (current and max),
as well as uptime (sever and stream, separately).
Resolves#85
This is for #32.
This adds a starting framework for loading, caching, and executing
templates on the server and the client. Currently only the server uses
this functionality.
Change the function signature for commands to return a string as well as
an error. Currently, errors are handled no differently than normal
response messages.
Also removed the redundant `/setpin` command.
This command launches a goroutine that does the actual downloading of
the emotes so the client doesn't hang until the process is finished.
subscribers.json needs to be manually downloaded and added next to the
main executable for now. Eventually, it will download and cache this
file (for 24 hours, at a minimum).
Issue #10 should not be considered finished yet. This needs to be
tested a bit more with multiple channels. Ideally, some that exist and
some that don't to verify that it fails gracefully (ie, adds the emotes
it was able to download).
Functions added:
- LogErrorf()
- LogErrorln()
- LogChatf()
- LogChatln()
- LogInfof()
- LogInfoln()
- LogDebugf()
- LogDebugln()
- LogDevf()
- LogDevln()
New settings configure the logging: LogLevel and LogFile. LogLevel can
be set to one of: error, chat, info, or debug and will default to
error. LogFile is an optional file to write to. Providing a file will
not prevent output in the console.
LogDevf() and LogDevln() only compile when the "dev" flag passed to
build. This will cause Travis-CI to fail the build if it finds any
calls to either function.
Names can be passed to commands with @ prefixed to them.
When forcing a color change on another user the @ explicitly defines
the name, instead of simply being stripped off. This will allow the
a user named "red" to have their color changed: `/color @red blue`
A bunch of the color command code has changed with this to make things
less ambiguous and to add some other checks. `/color red #FF0000` and
`/color @red red` will change the color of the user "red" to the color
red while `/color red red`will return with an error. Note that the
color an name arguments can be in any order.
Resolves#64
All emotes are set to have a height of 28px. This fixes the
auto-scroll on message when the message has emotes, but it also forces
all emotes to be 28px high.
So far only PIN and Open modes are implemented. It uses a session
cookie to store the validity of the pin/password.
The "Enter pin" page has some unreadable messages on it right now, but
it kinda works.
The processing of ChatData was changed so ChatData has a DataInterface object, which will be converted to a
ChatDataJSON that will have the DataInterface field converted to a json.RawMessage field. This change allows
greater customizability in the chat messages, since the Data field can still be read and modified.
Moved to common/utils.go:
- IsValidName() (function replaces direct calls to regexp.MatchString())
- RandomColor()
Moved to common/emotes.go:
- LoadEmotes()
Using this command will change the current user's nickname in chat.
It has an alias of /name.
An admin can force a name change by supplying the current name and a new
name. If a name has been changed by an admin, that user can no longer
change their name (similar to forced color change).
Resolves#42
Move the help html to a template and populate the lists dynamically.
This also splits out the base HTML stuff (eg, the <html> and <body>
tags) to a base template file.
