Discussion
Studying how Papyrix renders XTC ebooks... Which is a format specifically designed for this device.
To open such a file it closes anything already open, defers to a support lib for parsing it, clears metadata, retrieves the title to copy over to that metadata falling back to using the filename, retrieves the author falling back to an empty string, creates a cache directory, assembles (via helper) for the cover BMP, retrieves the parser's page count, & zeroes out some progress properties.
1/?
Generating thumbnail not only checks if the file already exists but also a ".thumb.failed" file to cache errors, & synthesizes the cover 1st before converting to a thumbnail.
As for that parser... To do its part in opening the file it closes the previous one if any, opens the new file, reads & validates a header struct via a helper, propagates any errors, reads & validates title (via a helper) from a header specified offset, does similar for author, ...
3/5?
... does similar for the array of pages propagating any errors (with another helper in turn parsing the 1st page to get dimensions), carefully reads & validates a has-chapters flag & an offset from offset hex-30 (via a helper) to know where to read & parse all chapters (plural!) from, propagates any errors, flags as open, logs results, & returns success.
The close method closes the file & clears various properties.
Skipping those helpers next is...
4/5?
To load a particular page after validation Papyrix's XTC parser reads the page entry at a computed offset copying the data out, reads a page-header from where that said propagating errors, validates its magic number, computes & validates the bitmap size, reads that much data in propagating errors, & returns success.
There's a variant of this method feeds each chunk of the bitmap to a callback.
There's a public API around that page entry.
There's a validation method, & some getters.
5/5Fin!
The close method defers to the parser & clears that captured metadata.
There's getters for the properties & those of the parser.
To get a ToC entry it validates with the parser that it has a table-of-contents & that the index is in range, retrieves the chapter array from the parser & the given index into it, & copies the data out.
There's a couple utils for assembling filepaths.
Another method defers to `xtc::generateCoverBmpFromParser` if the file doesn't already exist.
2/5?