Discussion
Studying Papyrix's support for rendering Markdown ebooks...
Its Markdown Provider subclass opens such files by closing any existing file if any, constructs a new Markdown object & remembers it, defers to its load() method, clears its metadata properties, copies the title over, zeroes out the author, copies the cached path over, copies the cover path over, sets initial page data.
To compute a page count it estimates that the screen can fit 2KiB of text, same as for plaintext.
1/2 for today?
To find a cover image it extracts the file's directory & defers to a GfxRenderer helper. To generate the cover bitmap it with results & failures cached on the SD card it locates the cover image, invokes that cover-image helper or errors out, creates the cache directory if necessary, & has the graphics library convert the image to a bitmap.
Then there's similar methods for thumbnails.
There's a method to read the file entirely into RAM.
3/4!
To extract the title with caching on the SD card, it allocates a buffer & reads the file entirely into RAM, iterates over non-blank lines, validates results from that loop, truncates the title to 127 chars, & remembers it. For each line it checks for a preceding "#" counting up subsequent hashes, validates there's at most 6 & that there's trailing text on the line, & strips surrounding spaces & hashes to hopefully get the title.
4/4 Fin for today!
A support library, among other things, provides a Markdown class whose constructor concatenates together a cache's directory path, & extracts the file's basename to be its title.
The load method carefully opens the file from the SD card, retrieves file's size, closes the file, & invokes a helper to extract the file's title.
To clear its cache it removes that directory if it exists. And there's a corresponding method to create the directory if it doesn't. A 3rd returns cover.bmp from the cache.