Mreow
HAIIIIII THIS IS A DEVLOG YIPEEEEEEEEEEEEEEEEE

• Unread channels highlight
• Channels that you got pinged in highlight

This isn’t much of new stuff just really cleaning up the codebase and adding types

• Split up stuff into seperate modules on the rust side so compile time is faster as it doesnt have to rebuild whole of lib.rs
• Slack.ts file now uses types so its harder to make dumb mitakes
• Fixed new msgs not getting name fetched if they arent already saved

ITS A DEVLOG AFTER 3 MONTHS OMG WHAT?!

• Sections
• Channel priority followed as API gives it
• Usernames now work properly for channel history still broken on incoming

Changelog

• feat: semi decent message parser and sum other stuff
• feat: change how i fixed tauri callback id issue
• feat: live messages, post messages, my shit attempt at preventing dumb tauri issue

Woah is this that time already? It’s the last time too before the new ship I believe. So here are the changes chat this journey was great most of the stuff after this probably wont be in a ship

• Removed old db migrations from different dbs
• Fixed drizzle not following strict typescript
• Fixed huge n+1 with project map hitting db on every user doing opposite of optimising it
• Included manifest from slack, update readme to show bugsink stuff

Changelog

• fix: drizzle never got updated to strict type, chore: drop old db
• fix: huge n+1 with projectMap hitting db on every user doing opposite
• feat: include manifest, update readme to match new bugsink stuff

Guess who’s back. Back again. ITS ME SILLYYYYYYYYYYYYYYYY
HAIIIIIIIIIIIIIIIIII I DO ANOTHER DEVLOG

• 10x improved the projects command as it now has a better query system using –limit and –query for options and no – for just search and it no longer uses slack’s table block which was broken it is actually usable over 10 projects!
• Updated zod types for FT API
• User command now shows achievements
  OKAY BAIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

Woah another devlog? This is so figma balls

• Ratelimit on API so no stinky people can ruin it
• Updates README so it makes sense for new version
• Bugsink error cmd now so you can see if any errors logged
  OKAY BAIII CUTE PEOPLE

It’s about time I meow at you cuties :3c:

• Fixed cookie predictions as i made it user base not project base causing predictions to be wrong i also made it store in db with user current cookie amount + predicted contributing to issue
• Feature and Bug Report commands added :yay:

Woah a qucik update? This is just bugfixing mostly but

• Fixed docs using Goals request body instead of Multiplier request body
• Multiplier was written stupidly by me I made it user specific not project specific so I had to change API and DB to store in projects by making projectId requirement now in API requests

New devlog who dis? ITS ME YOU SILLY GOOSE! ANYWAY UPDATE TIME

• Sabio begged me for new api so they can give me project multiplier idk if ft utils would implement api but sabio said they would add it to spicetown
• Oopsie lets not add region again instead lets overwrite so it changes not causing issues
• Added new api to DOCS
• Added tags to docs to make it clearer then it all being under default
• Move api routes to seperate files as index file was getting fat
• Set vikunja stuff optional so self hosted instances dont need vikunja to run

3 hour devlog? YIPEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE
I spent ages just making 2 command cuz node-vikunja kinda sucks to use

• FEATURE REQUEST COMMAND SO YOU CAN REQUEST FEATURES
• Report command to report when logpheus is being a meanie pants and refusing to work properly >:(

Not much of a devlog but still worth a post so yes

• Goals are now on the shop command!
• Decrease shop item limit on shop command cuz slack sucks have to wait for cyteons pr to get ?limit= to create a pagination fix for this

Woah a new devlog?! YIPEEEEEEEEEEEE

• FIxed shop command not grabbing region correctly from meta
• Goals command to see your goals remove and add

Yknow this isn’t even really a devlog we went back in progress

• Goals on devlog post cuz yes thats sigma
• PUT request bug fix as it only showed goals given to api not updated db
• Drop embed support for videos as it never worked and prob will never work because well thats just say I have never seen an api more idiotic then what slack has made I seriously hope slack rebuilds their api from the ground up because whoever made it in the first place failed miserablely at supporting video

Woah woah woah did someone say a new… a new devlog? WOA AWOW THIS IS SO SIGMA
Yall get a new devlog you cutie pies

• Help command so you can see command params and command descriptions easily as slack kind of sucks for slash commands!
• Devlog command so you can see a devlog’s content in the slack
• Zod infer types for ft api functions as not being able to see what needed to provided was annoying meeeeeow
  OKAY BAI CUTIES TILL NEXT DEVLOG YOU CUTIES

This devlog is all just bugfixing and optimization so it works better with a lot of users

• Make the user check for automated posts more specific to not include regular users
• Decrease N+1 DB queries with prefetching projects and users and search using maps
• Fixed logic error where 1 bad handler could break all handlers
• Lock mechanism to prevent reruns of handlers when a handler is still running
• Dropped old 2s wait used for ratelimits as new ratelimit handler
• Keep parseMarkdown regex in mem already loaded so slight improvement on bursts from automated devlog posts
• Use utcFormatter which is already in mem as slight benefit on resources for each devlog post
• Drop clients if none in userRows
• API now reuses ftClient if it already exists in clients
• Fixed videos not working because misisng title and thumbnail
• Don’t hit db for data we already have
  In docker it didnt drop mem usage by much only around 16mb which is still nice and it did drop idle cpu usage a bit but still bursts during activity it mostly just improves speed for user for posts

Only 45 minutes? For all this commits? crazy i know right… ANYWAY HAI CUTIES IM DEVLOGGING Again
This is mostly bug fixing based it has new stuff but thats part of fixing issues for users

• Fixed region text to be more specific on format so users know how it needs to be to get it to work
• Reactivate command so you can get undisabled in db once you fixed an issue as config command makes u update something to reactiavte and logpheus not being in channel just requires being added to channel to fix
• Use more prefix based text instead of “logpheus” so it makes more sense on self hosted instances
• Disable users from automated posting when logpheus not in the channel
• Log on API so it shows in sentry and is useful then not making sense to debug
• use ctxLogger on timer based handlers because without it wont know what file is causing it
• Removed debug console logs i left in qwp

Whoops didnt mean to go this long its a 1 commit update! giant commit tho

• Swagger API Documentation for /api/v1 on logpheus
• Goals API so FT Utils/Spicetown can give me your shop goal data
  A lot of this was making goals api and also trying to get swagger to work with boltjs custom routes as slack sucks

Ts a basic ahh devlog but i mean hey i need the HOURS GRRRRRRRRRRRRRRRRRRRRR

• Fixed bug in data command causing double “:” to appear
• Add command now accepts a param for project id to make it less inconvinent adding it
• log better in prod to see why unexpected project response error cuz idk why that happens
• Fixed remove command not working for set id and removed remove all option as thats just revoke and it doesnt delete user from db now as revoke does that
• Data command now includes metadata

Quick bit of coding before I catch the bus

• Hopefully fix the issue where it says unexpected error from api as no status code
• Revoke command to remove yourself from logpheus with ease if u want to sadly

I did some stuff

• Some how THE CHANNEL CREATOER REQUIREMENET STAYED EVEN THO I KNOW I REMOVED IT GRRRRR
• Spelling mistake in region text in config view
• Zod error in shop for item id as it sent string not number
• Hours of project based on devlogs ;3c
• Devlogs function checking params for query throwing zod error
• Predicted cookie amount (based on avg of 10 cookies per hour not previous ship cookies)
• Ping groups are in wip

This devlog really isnt that much ngl I just spent the past hour and a min cleaning up old code to make the codebase less complicated so

• Dropped add command taking in api-key for new comers the register commands and its easier to maintain one command for adding a new user to logpheus then 2 ways
• Made views and commands more reliant on checkAPIKey function so less code in every file for doing the same check everytime
• Cleaned up register command to also use checkAPIKey but also easier maintability as changed way from spam of if statements by using a obj with type UserInsert to add to it and added ability to add region on register
• Both config and register command now has hyperlink leading straight to flavortown’s api key setting box to make it easier for user to get

I did a bit of updates :3c mainly to shop command but some other stuff too SO LETS DEVLOG

• Shop command now has region based showing (if you have a region set in /logpheus-config
• Config command now prompts for a region
• New meta column on users standing for metadata on users for stuff like regions
• CheckAPIKey returns db row too now so I can use it without new db call when data was already grabbed during that check
• Log Zod issues in bugsink now
  LETS HOPE WE DONT BREAK PROD WITH THE CHANGES

ps: matei if your reading this you are a cutie pants

Woah more changes IM LOCKED IN CHAT LITERALLY LOCKED IN AAAAAAAAAAAAAA HELP ME THEY LOCKED ME IN HERE IM TRAPPED IM STUCK GET ME OUT PLEASE LET ME OUT

• Fully changed to Zod for requests so thats sigma it points out my dumb mistakes more
• Finally handles http codes set out by docs mostly some stuff still no i dont think
• Changed shop command to not use disgusting slack table and also push it to prod

I mainly just worked on the code conversion in this update so it all uses ZOD NOW
The zod types autogenerate off of the openapi spec the use of scalar makes which flavortown now uses so MY BAD CUSTOM TYPES ARE NO LONGER IN USE WE USE ZOD
I also made the basic api data check better as it was kinda hard to understand and messy before

I should devlog right about meow so HAIIIIIIII x2

• Added da shop command so you can run /logpheus-shop [id] to view shop items and the id to look at a specific shop item just not pushed to prod yet (This is still a WIP i will be changing away from using slack tables because they don’t actually work properly so you can’t view whole table)
• Config command was still channel restricted which was meant to be dropped a while ago so I dropped it just need to push change to prod
• About to wip a /logpheus-help command to view all commands because I know slack can kind of suck to show all logpheus commands

Wow this is a short update so I ASKED FOR A PR AND IT GOT MADE BY CYTEON WE LOVE CYTEON CHAT YAY THEN IT GOT ACCEPTED AFTER A WEEK BY NEON WE LOVE NEON
so now

• Attachments on devlog posts
• Fixed checkAPIKey failing when trying to change api key in config command due to it not being able to find you as it used the new api key you are providing to do the check not old one

Rawr not much changes just bug fixes

• Fixed code in parseMarkdown thrown type errors as I didn’t updat to strict
• Shut up loadHandlers as it failed dupe check every min spamming console as already registered
• Fixed throwing error on empty optional env
• Removed sum debug stuff
• Fixed prod explosion cuz commands didnt work as bug

I did some things

• It now redacts sensitive info in logs as it didn’t before
• Improved FT API Handler so that I can easily modify it without it being annoying
• Made it throw an error on dupe handler names as I keep accidentally not changing the name
• Changed to use lib of checkAPIKey instead of manual implementation for each handler
• Enforces strict typescript so that its easier to see issues in files

Since I forgor to devlog my changes yesturday/day before I’ll devlog it today for you cuties

• Fixed the disabled message not getting posted which caused people not to get alerted when bot broke for them as unauthorized
• Wip of Projects Command mostly working but slack hates long tables idk how to fix that and still needs to fix pagination on it
• System information on the stats command now

58 minutes for this? Damn peak! ANYWAY UPDATE

• I added a drop command to make dev testing easier allowing me to drop data in db ONLY ON DEVPHEUS and only people with MY SLACK IDS so can’t happen on prod
• Dropped sentryEnabled mostly everywhere as logtape is now used so I do the check if sentryEnabled in the sink so if its not enabled it uses console and if it is enabled it uses sentry

Changelog

• feat: drop command for dev instance and only attached to my user ids so I can be lazy and not go into db just to test, chore: drop sentryEnabled from most stuff and instead use sentryEnabled on logger sink to see if enabled use sentry else console

Quack another devlog colon three
Update not really that big but still I need the hours

• Increased deployment speed on builds from new commits by making apk (alpine linux package manager) cache and use the right bun cache location
• Added MP (Multi-Platform) Dockerfile as I think my changes are really platform specific to alpine as I used apk cache folder on host too as my server runs alpine linux so to ensure other instances don’t break they can use MP
• Updated the README to reflect new changes for docker with the MP but also to reflect the change in code for chosing how the prefix is chosen
• I made the markdown parser more maintainable by swapping from a bunch of if’s to a switch statement and not using a bunch of numbers with match[] to find it and instead using match.groups! for group.,emoji etc so it actually makes sense
• Fixed FT Utils emoji handler on FT messing up Logpheus’s emojis as it uses it was links so I had to make a parser to look for emoji.slack-edge.com and parse that for emojiName to use emojiName for emoji in slack

I LIED THIS WAS A BIG CHANGE BOZO GET LIED TO

Yeah I should probably devlog
HAIIIII FLAVORTOWN VIEWERS AND QUACKERS
so this time was literally spent on literally debugging

• I bwoke prod of logpheus so I needed to debug and a lot of the debugging was kinda vague so I swapped over to logtape so I can get better drizzle logging in sentry and also allow me to log easy with sentry
• I added a lot of debug stuff to prod to see why it was down and found out IT WASNT MY CODE YAY it was a psql misconfiguration causing it to not allow wireguard connection
• I moved prod back to server with db as db reqs was slow but now there is more chance of slack timing out requests but ig cope i want fast db requests not slow ones

HEYYY ITS ME FINALLY AGAIN ima just say the changes i made today as i cant remember all of the stuff i fixed a hwhile ago JESUS 17 DAYS SINCE LAST DEVLOG DAMN

ANYWAY

• The logs was pretty vague especially for drizzle so I swapped to logtape to log all drizzle stuff to sentry enabled breadcrumbs to help debugging even more
• Swapped the contexts to use logtape so its even better for contexts
• Broke prod logpheus because slack broke saying i didnt own bot meaning i couldnt work on it adding new commands n stuff so i had to kick it from the workspace remake new account
• DMed all private channel users about the kick so if they want to use it they just have to add it back and for public channel users I added it back for them

Woah an update? IN THIS FLAVORTOWN?! UNACCEPTABLE
Too bad accept it
ANYWAY UPDATEEEEEEEEEEEEEEEEEEEE:

• License page just so i can link to it in footer
• 88x31 buttons in footer
  Thats about it small change but bleh

Changelog

• minor: update 88x31 to link to my img, clean up license page bleh, link to license
• feat: license and other stuff
• fix: just saw their pr and it used name as estrogen
• minor: add alyx to 88x31
• fix: 88x31 buttons not leading to sites
• feat: 88x31 buttons :3c

Bleh haiii again
I spent the past hour adding hc webring to my site (would of been easier using the provided embed but that doesnt fit the theme so why would i do that?)
I remade the embed that provided the webring but with govuk design to match better it doesnt look like much but it was pain to get it to be like that

Another day another update BLEH
I mostly did fixes but also some nice to have changes

• Vertically align emojis from devlogs with the text
• Make checkboxes more visible but also uglier because i couldnt set a custom bg for them
• Drop solid-icons for unplugin icons to shutup a babel error which also slowed down load in dev
• Change with from solidjs router as i heard its better for solidjs idk
• Throws a 404 now when project doesn’t exist instead of erroring
• Shut up solid hydration because of using not for 88x31
• Drop kaboom a tag which was hidden from viewers but not needed anyway as i went with different fix for 404
  Thats basically it also I GET UNBANNED FROM THE SLACK TODAY YAY

HAIIIII IM BACK AGAIN ;3c
SOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO i made sum changes if you didnt notice dumb ahh

• Modified FTToMDX to not just use minutes but hours and if you need days too that but like that prob wont be hit for days
• Swapped da order from oldest -> newest to newest -> so it looks better
• Better markdown support with h1 support and other stuff like checkboxes just not existing
• Improved detection of lang in markdown code block but if it cant detect it just say unknown then nothing
• Copy block no longer copies its own text
  I got the 4167 devlog id?

Changelog

• minor: change from just mins to days, hours, mins
• feat: improved detection of language by parsing child, fix for copy button coping the copy button too not just the code
• fix: lacking rmark-gfm causing some markdown to not be supported, h1 missing so it didnt match
• feat: ratelimit handler + fix for it not moving stuff for new devlogs i forgor about that case
• fix: swap order from oldest to newest to newest to oldest

An hour? Damn not that long to be fair, ANYWAY hai chat I did sum fixes

• 404 continued to explode on github so I fixed that finally by making a static 404.html not solidjs one,
• 404, Details and Home page is less fucked on mobile now
• Improved birthday count down text to do less typedjs typing

KABOOM WE DO UPDATE YIPEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE

• 404 fixed i think on prod it works in dev
• DETAILS PAGE ABOUT ME
• Sum other stuff i forgor

Changelog

• fix: 404 exploding bleh
• feat: fix 404 on github, 88x31 for this site, github follower count and star count i stole from cyteon, details page

Okay i apparentally forgot to hit post so let me write this again. HAIIIIIIIIIII
i DID SOME THINGS

• Prod setup exists on github pages?! https://gizzy.gay
• Swap replacing stuff to prevent execution with just using .md instead of mdx
• Updated breadcrump to show current page too
• Fix ftToMDX to stop spamming new lines and changing tags to tag
• 404 Page now
• Search for projects or tags
• Fixed postsPlugin to not explode when md
• Page navigation section
• Some other things i cant rememebber
• No longer shows posts with tag projects on blog posts page
• Made tags and blog posts page more sexy

bleh i had to fix script and aciton that was painfal its 1 am im eepy but update

• github action runs update every 10m now correctly
• it prevents xss now from mdx bleh
• IM TIRIED

BLEHHH IM BACK AGAIN since it has a blog now i gotta do the natural thing of being lazy! I made a github action to generate an mdx from these FT devlogs and projects and add it to the website mdx i also did sum other stuff but thats not really important compared to that

HAIIII ITS MEEEE AGAIN ;3c
This wasnt much but yipeee

• Uses giscus for comments (it uses GitHub discussions for comments)
• Default blog page to list all blogs instead of it 404ing on /blog
• Breadcrumbs instead of bunches of back buttons i have to add to every single page as its top level and dynamic

Changelog

• feat: default blogs page to list blogs, breadcrumbs at top level so not spamming back buttons
• feat: giscus for comments, gov giz logo redirect to /

BLEHHHHH ANOTHER UPDATE SO SOON? YES YES I UPDATE RIGHT MEOW
ANYWAY

• Tag system exists meow
• a bit of changes to blog.tsx i forgor what they was

Changelog

• feat: tags
• fix: add copy button

Wow an hour just spent on this? uk gov being fucking annoying again! The UK Gov Design System lacks a code block which a github issue was opened for it in 2018. so i spent past hour literally combining other components and jankiness to make one using prismjs for highlighting it looks pretty similar to the real one used on design system docs (which EXISTS BUT IN THE NPM ITSELF DOESNT????).

BLEHHHHHH 2 HOURS ALREADY DAMN BLEHHH
ANYWAY
UPDATEEEEEEEEEEEEEE AAAAAAAAAAAAAAAAAA

• I did sum code theft someone made a blog in solidjs with mdx before me so i stole sum code from it there blog on it is here so i made use of that code for mine but modifying it so it fits uk gov design bleh
• I did other stuff but forgor

Changelog

• feat: basic ahh blog i ripped off from someones site but trying to make fit this design

Im eepy its 1 am but gotta devlog so yipeeeeeeeeeeeeeeeeee

• Uses hackatime now to get my hours coded today
• Shows my local time which also live updates using typedjs to update it
• Small desc about me uses typedjs to change values bleh

Blehhhh so new project first devlog i spent most of time fighitng to get it to work with solidjs but it works AND LOOKS SEXY i also wiped up custom logo in figma balls

Chat, are we codemaxxing rn? no stupid we are bugfixing dumb ahh

• Fixed my channel being spammed in dev cuz i dropped project table by adding fallback to checker to prevent that
• Fixed sentry context not being included in commands/views sending sentry errors
• I realised slack modals have private metadata and also has body so instead of manually parsing the channel and user id through view i used that to hide it from user
• Idiotproofed the bot more because people somehow keep using other self hosted instances so now titles of modals use prefix as title which is different on every self hosted instance and for offical prod its just logpheus so if you STILL use wrong instance you are dumb nothing to say more about that
• DATA COMMADN TO SEE WHAT DATA LOGPHEUS HAS ON YOU YIPEEEEEEEEEEE
• Config command not channel specific now it uses user id to update api key
• Tells user that prod exploded instead of ghostting them when it does

Almost 3 hours again GOD i kinda forgor was gonna do one this morning but devlogs was done ANYWAY

• Register command and project command to register to use bot as general purpose just for its commands, user to get a persons profile and project to look at project by id
• Bug fixes such as shutting up the error logged when FT did a restart or anything like that
• Drops disabled clients from saved clients arr because they are disabled
• More bug fixes as an FT api route changed without me knowing breaking the bot so had to fix that and fallback so if it happens again it logs to sentry as sus
• Did a update to readme to match new stuff

oop 3h already past jesus anyway
i fixed sum bugs
user command! get someones ft profile from sometimes slack account
opt outs column for something soon owo

Another time for a devlog!!!!!!!!!!!!!!!!!!!!!!! BLEH :hyper-dino-wave:
I spent most of the time improving the error handling of logpheus so it makes use of MY BUGSINK MOREEEE and its more specific with errors so now if it isnt 401 it doesnt default to no project exists at id it only does that at 404 :3c with MORE CONTEXT TOO so better easier debugging

uhm pushed to prod the changes i madeto table etc fixed sum stuff took down prod for like half and hour as u do so fixed that because of dumb migration issues bleh but it does make use of new tables now for data not all data in new tables tho like last devlog timestamp or the store of user ids so i need to manually add uids in for users previous to the update but also change it so it stores and uses last devlog timestamp as main check and fallback to id check if need be

bleh more time for me! I did sum cleaning moving devlog check to sept file and no longer rely on loadApiKeys which is super overkill for 1 function using it since func was legacy code used a lot before changing to drizzle etc

AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
I work on moving schema :3c and also cleaning up the codebase so its more readable as ts confusing also improved schema migration by learning how to actually use drizzle frfr

Well uhm this was a crazy one, a self hosted instance caused offical to die because of how slack handles slash commands its extremely secure setup allows other bots to overwrite the slash commnads you have to their own and dw it does the exact opposite of telling user this it still shows that the command you are about to run is from the offical bot so i had to modify it so self hosted instances now use user id and bot user as prefix to fix it also updating schema to work more efficiently then and clean up code

Mreowwwww! I’m literally just improving logpheus handling of issues to use bugsink to post errors so I can read but also handle 401 errors better so I’m not being spammed with errors because you guys being annoying with your api keys

Bleh not devlogged this in a while but i got time so might aswell
Fixed bugs!

• You can now add multiple projects to one api key as fixed bug preventing this
• you can no longer be a stinky and try add the same project twice under a api key

I implemented more of the formatting :3c also added names back as i broke that

Better parsing of rich text blocks :3c not completely dont yet tho

A bit better message parsing :3c it now has semi good image support setup image is kinda large gotta fix that but it also unescapes html encode stuff now so yipee

Names! Yes this could of been done in like 30 minutes but im a twat i made a stupid mistake causing me to not match the user id to the gotten ids for the channel because i used channel: channelId in the body not channels: [channelId] causing slack to throw workspace users not channel users. I also moved chat to its own file instead of just in the index.tsx file

Yeah my previous fix to shutup couldnt find callback id annoyed me so i spent this whole devlog time just trying to get my shit rust code to work so it prevents default reload and instead causes a window relaunch clearing out any callback ids so no console spam because it relaunches on reload

I forgor to make a devlog qwp. ANYWAY 10 days past and i did some stuff :3 I tried to get tauri to shutup about “couldnn’t find callback id” because it sucks at cleaning up its own resources with its plugins that took a while until i just made console.warn filter it out of console to shut it up instead it aint my job to fix tauris shit, I also made a bit of a ui for it and real time messages from the websocket and ability to post messages! (dont mind the random person on the right i stole the ui from flowbite :3c)

Nvm i hate code! 16m of more bug fixing to get it to stop spamming qwp

I’m logging this cuz free time! Logpheus felt like going on a spamming spree because of a bug in my code. Every minute a check ran and my dumbass made it update db to the value of the check so every 2 minutes it would change value back to one that causes shipped message to post causing spam hopefully this should be fixed tho

WOOOOOOO just took down prod for 10 minutes because im dumb! I spent this time fixing my dumb mistakes, adding support to use full postgres in prod so I only need pglite in dev making my life easier in both cases and adding validation of api keys when added!

This whole devlog time was spent on another migration because I found out about pglite (postgres in wasm) and drizzle supported it so I migrated over to that instead because drizzle makes my life way more convienent especially working on logpheus it just took so long to setup as it was my first time using drizzle and i had to also create migration from sqlite and json to pglite to ensure no data loss.

Migration!

Using json was really annoying especially for every project using their own cache json file for devlogs so instead i swapped to SQL which is using bun:sqlite for storing api keys, projects etc taking advantage of buns implementation which is pretty fast but also its less resource usage slightly because I’m not using json for everything I need parsing it stringifying it pushing it to file a lot. All previous json files will be migrated on start to the new db.

OKAY REAL DEVLOG THIS TIME, markdown support is now fully added i think yes

Trying to implement support for markdown in logpheus
meow

i like potatos

Haiiiiiiiiiiii

I NEED MARKDOWN TO TEST LOGPHEUS THIS ISNT REALLY A DEVLOG JUST TESTING SOMETHING DEVLOG SOON THO

REA
DK
YOUR BALD

• QUACK
• A

1. A
2. a

• quacked at people today
• meowed at people today
  sniffs you

arf arf

quacker

sadgagagdag

UH MORE HOURS FOR ME YIPEEEEEE so basically i in this time explode prod a couple times because of the new dockerfile but fixed it added improved error handling so that it doesnt spam my logs when something like a 401 happens i added ability to change api key in a channel using /logpheus-config, i converted timestamps to utc and now use the good format not the us format for dates so DD/MM/YYYY and it is now clickable timestamp so users can click it and convert it to their timezone if they want to and i moved the ext header which is useless to axios.create instead of spammed in every func call

Might aswell make a devlog for the stuff i did even though not really important. During this time I improved upon the add command to post more detail of the project when being added, I made changes to the Dockerfile to improve its build times using docker’s cache mounts and removing unnecessary stages (i stole the build file from buns website and its very over complicated for this) and I made pending review after ship message more clear and once reviewed message more clear.

Clean up of code to be more easier to develop in by seperating views/commands out into seperate files, fixed cache files that was done before latest update using differnet cache setup not causing any issues but with this update would of so it now migrates old caches to new format and now includes ship status!!!!!!!

More improvements to the code! The bot message has improved to be more detailed with the devlog and also improved the code by setting types to the functions used so when called you can know what data is returned easily from typescript and your ide and also cleaned up some of the code.

Okay bot now makes use of devlog_ids to check so less resources during hosting and should also mean less api hits :3c also now /logpheus-stats to see how many users it has and every req has header that makes bot show up on the extensions page

I need a devlog to test but the bot’s handling system of new devlogs is being changed to do less work by making us of the new devlog_id in /api/v1/projects/:id making the bot and probably the api do less work thats just assuming but atleast the bot will do less work so peak!!!

An extra 21 minutes for me because docker sucks! I literally spent 21 minutes just trying to get apiKeys to be persistant in the docker containuer using docker volumes i hate docker NOW ITS PRODTATO READY

Prodtato. The bot is in prod and seems stable working in my channel! The docker is complete and also finished the readme showing how to use it but also how to self host it if you dont trust me with your api key which is BORING TRUST MEEEEEE its basically done i ship i think i did other stuff too but i forgor

Woah fast fix too i swapped it to use api key instead sas the prefix only allowing 1 api key to be used for a channel but a channe4l can have multiple projects subscribed so yipeeeeeeeee

Woah this was a fast devlog i made /logpheus-remove so you can now remove your project from a channel! and i also realised im an idiot! i now gotta modify the code to not use project id as a prefix and instead use channel id so someone cant block you from adding your proj to your channel because they added it to theres!

The past hour I have been debugging and making the bot worked implementing /logpheus-add to add projects to the list and it works i just need to add /logpheus-remove now :3c and also docker support

I need a devlog to test so uhm past hour and half i well made lib file for ft its a class file so i can use it in other projects if i want to too then using that i made it pull projects then pull devlogs if no cache file exists it adds all devlogs of project id which is in apiKeys.json which stores project id as key then under it the api key and channel to post it to

More modifications to auth! This time not because slack sucks but because I wanted auto login. It now saves xoxd and xoxc in a json file with name scheme “slack-${user.name}.json” user,name being active system user so different system users dont got same file and don’t worry it’s encrypted with a key set in system keyring under that users account. The user account name is grabbed using a tauri command with using the rust crate “users’ to get the active system account.

Well this was semi painful but WEBSOCKET mhm yes we got slack websocket in it now and because slack sends so much data even on channels your not currently looking at you could multiple channels at once on it :uuh: mhm yes you can see people typing, that someone joined someone sent a message while not being in that channel very sigma must say but this was pain to get to work

30 minutes later and it works mostly now! The webview window now closes once finished and you can view channels and message history again :3c

Well this one is a long one. I literally spent 4 hours and the feature isn’t done yet. I can confirm that I hate slack, I have gained more hate for tauri and more love for tauri. I tried first to make the request for data i need manually which for some god known reason refused to work and kept 404ing aka i hate slack, i then spent my time trying to grab data from session storage before slack wiped it which that didnt work out then I realised it moves the data to localStorage which is well weird and I wrote rust code to grab that then I spent more time just trying to get that data from webview to frontend code and FINALLY I GOT THE DATA login still not done tho now take a picturee of my shit code

I spent most of the 2h 33m trying and debugging just to get a single d cookie that slack auth uses it was pain in the ass since webkit2gtk kept crashing (because it sucks ass) when i tried to intercept responses to look through there headers under I realised tauri supports grabbing webview windows cookies so I swapped to on_navigation to look for path /checkcookie which slack auth uses to check for cookie is valid so we know the webview will have the cookie at that time then we extract and print it for now. I still gotta grab the second token although that should be 10x easier since slack sends that one over a network request only requiring the d cookie to get it although it does require some parsing since the response is js script? thats really stupid

Yes I am 1h 48m in to this and I have made barely any progress. Blame slack for that I spent all this time just trying to get convo history to work :hs: