rmgk/viscel

Read and archive webcomics. It‘s awesome. Probably.

viscel's Issues

write something to make debugging narrators easier

maybe some helper methods for the scala repl, so one can poke around a bit to see whats failing.

chaptered archives with next can skip images

happens if there is a new chapter before all images of the current chapter are fetched.
the current chapter is never looked at

metacores

they worked once. now i am too scared to test.

neo4j

works quite well all things considered,
but it is heavily overpowered for what it is used here.
viscel basically just stores a tree inside the graph, replaces subtrees while updating,
and does a preorder traversal when connecting the user interface.
maybe a simpler scheme of serialising the trees to disk would suffice.

narrator creation without recompiling

a narrator is what parses a page for content.
these are currently all defined inside scala code and need to recompile on change.
thats really annoying for fast test cycles you usually want when trying to find something that works.

documentation

it would be nice to have

• what is viscel good for
• how to start
  • sbt run
• how to find the interface
• what the heck is happening
• how to fix stuff that breaks. which it will.
• how to write new narrators (that probably depends on #10 as no one is going to understand the full versions the first time :)
• overview over the different parts

improved monitoring

currently logging is little more than writing to stdout.
investigate triggering events instead of unstructured logging.

make the updater sane

the updater received its fith or so rewrite today. it looks promising now, but i will not hold my breath.
this issue is to track how often i will fail from now on.

update the ui without manual refresh

currently you have to manually reload the page to see newly downloaded pages. that sucks.

add dockerfile

distribution independent installation method

reaction in the ui when bookmarking something

currently you click and never know if something happened

UI to manually mark parts of a collection as broken

collections break. sometimes spuriously. which is often notable while reading.
add some UI elements to directly redownload some images / collection archive

fix the export data format

the current plan:

• reuse .cbx format
  • each chapter a folder
  • each asset a file
• store an additional metadata json file containing a list of assets, and a list of chapters
  • investigate sane structure for that format

spray

spray may be overkill for our purposes.
keep an eye on the future of spray (akka/http) and maybe on http4s as another possibility.
it currently works so maybe its better not to touch it :)

simple narrators

related to #9 defining narrators is currently very powerful.
but a lot of comics can be parsed by just giving some hints how to find the image,
and guessing the next link.
that worked incredibly well in 4.x so we want it back :)

stand alone archive viewing

investigate the possibility to include the compiled js viewer in an exported archive,
to make it usable with just the archive and a web browser.

tests!

the code actually might be testable by now …
all thats left is writing some

areas where tests seem particularly useful:

• the database navigation stuff.
  • that can easily go wrong, without very visible consequences.
• the serialisation of stories to json and into the database.
  • if that breaks, the database might become corrupted.