Fix fragment parameter case sensitivity

[DanGould]

Bech32 fragment parameters should not be case sensitive. This fixes that by converting params and bech32 HRPs they match against to uppercase.

Close #442

--

Depends on lint fixes from #447

[coveralls]

Pull Request Test Coverage Report for Build 12528347962

Details

• 6 of 6 (100.0%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.04%) to 61.844%

---

Totals
Change from base Build 12528252226:	0.04%
Covered Lines:	2898
Relevant Lines:	4686

---

💛 - Coveralls

[nothingmuch]

fragment.to_oppercase() instead of param? seems a bit more intuitive to me, and only needs to be done once

[DanGould]

I've addressed this and chosen a function chain approach rather than nested conditional statements

[nothingmuch]

utack, sorry for not doing this myself when it came up

[DanGould]

On second pass I realize this change constrains fragment parameters to case insensitivity. BIP 77 is going to need to specify that, and IMO must specify params are bech32 encoded to minimize dependency requirements and QR size

[nothingmuch]

and IMO must specify params are bech32 encoded to minimize dependency requirements and QR size

in other words uppercase is normative? FWIW that's what i was intending in #417. what motivated creating #442?

[DanGould]

what motivated creating #442?

bbmobile's bitcoin URI logic has bugs, and it lowercases the URI in display and for clipboard, so we end up with a lowercased URL to parse in the pj= parameter. Possible others will pass us lowercase URLs as well.

[nothingmuch]

we end up with a lowercased URL to parse in the pj= parameter. Possible others will pass us lowercase URLs as well.

ouch... that's incompatible with RFC 3986 FWIW.

RFC 3986 section 6.2.2.1 says case insensitive fields should be lowercased, e.g. scheme, host, but that's not a requirement, so QR alphanumerical encoding seems reasonable, and that % encoding should be uppercased, which they are. and:

The other generic syntax
components are assumed to be case-sensitive unless specifically
defined otherwise by the scheme (see Section 6.2.3).

[nothingmuch]

ouch... that's incompatible with RFC 3986 FWIW.

and would break if a directory has a path element, see #416

[DanGould]

We're set to resolve this by

• getting rid of set_path in Verify v2 client requesting directory hosted with intermediate path elements sets subdirectory ID correctly #416
• MUST uppercase fragment params in bip77
• [-] Fixing URI implementations downstream

I believe this issue itself is wontfix