cite-reader / codepoint-dump Goto Github PK

Consider colorizing the following classes of characters:

• Non-BMP (each plane ought to get its own color)
• Private-use
• Obsolete
• Reserved

Broken software has barfed on these things often enough that having the immediate visual indicator ought to improve debugging turnaround.

Currently, being fed a byte sequence that can't be decoded into a codepoint results in the following (fairly unhelpful) error message:

codepoint-dump: : hGetContents: invalid argument (invalid byte sequence)

Getting a byte index at the least would be useful.

Output is absolutely at proof-of-concept levels right now.

It works fine for now, but stuffing absolutely everything on one line doesn't exactly play nicely with tools that, say, use line-buffering. Granted at this point the only thing I've wanted to do with the output is read it, but that could easily change.

Replicating something like codepoint-dump | fold -s is a good enough place to start.

I know what all the various bits of gumbo in this thing do, but Six Months From Now Alex might not. So, uh, how does Haddock work?

At some point create a long-form introduction to the tool.

Automated tests are good and I should have them.