THE backlog

[SEAPUNK]

this is the big list of things thats on my radar. if you submit an issue and it ends up in here, that means that i intend to address it one way or another (whenever i decide to work on this again)

docs

• record demo
• booru browser, not booru viewer
• write README
• write post explaining it and my vision for it (working on this now)

deploys

• figure out how i want to dockerize it
• github actions dockerization + publishing

features

• adapter selection #5
• desktop client #4
• improved mouse browsing (a control for browsing and saving that doesnt require a lot of movement)
• resizing sidebar, fitting more thumbnails in (maybe i have a min and max thumbnail size and dynamically scale based on the width)
• can we use fetch to get download progress on something?
  • aand can we put this into an image/video? maybe we can use this to no longer need to deal with the whole lagfest that firefox does when wemake lots of requests
  • i can use fetch's streaming response to figure out the data count as it comes in,
    • fetching will be cached aggressively,
  • i can turn the response into a Blob, which then becomes a File with the filename
  • and set it to the image
  • unloading it with revokeObjectUrl, and rlying on fetch to have it cached whe ni need it again
  • how does saving work when i do this?
  • get rid of that image flash when image gets loaded in (pass down the image preview data? fetch all images with the image loader and pass blobs around?)
  • caching image blobs, accessing them refreshes their timeout
  • ^^^ build a POC
    • fetch gives download progress
    • we can create the File and put it into object url
    • and attach it to img
    • and see how it saves
    • ensure mime type is correct
    • then unload, and see how things change if we have the blob in a new tab or something
• search results into address bar (adding to history)
  • search query gets saved into address bar, along with the post ID
  • on navigate, use the address bar query
    • if there's a post ID, load the specific post in the viewer as standalone (and set browse cursor to none)
    • is there a better way to do this?
• tag autocomplete
• tag coloring/grouping
• post translations
  being able to hover over some text in the viewer and see the translation show up if specified

bugs

• safebooru adapter returns no results randomly #2
• browse: filter out dupes on page append (this is why keys break btw)
• PostView: fix control hiding sticking sometimes (something about useDebouncedHide or something idk)
• search: loading state (indicate when we're searching for something, or loading next page)
• search: pass down AbortSignal into fetch (i want to move off rtk query since it doesnt fit my usecase)
• search: dont cache results

improvements

• CI friendlier build scripts #3
• move off elysia to hono(?) (and document what i did to move off of it for my scaffolding project)

misc

• gather TODOs in code, put em here
• animation: scrolling to image?

---

misc scaffolding stuff you can ignore this its for my other project
notion doc: atsugami: open source-ing has my scaffolding stuff as well so i can come back to that later

• eslint: move to flat config
• tsconfig: use esnext(?)
• tsconfig: see how typescript wants us to set shit up
  • look at what bun does for inspiration
• see what other things we can do with the setup
• uhhhh recreate the thing step by step, in a batch script or something (batch script should use latest everything!!!!)