SanitizeString: Allow hypen char

[maflcko]

Maybe someone wants to put their-domain.org in the UA?

Ref: #4983.

[laanwj]

Maybe we could reverse the approach, and allow everything allowed by the BIP14. Otherwise we can keep creating 'also allow XX' pulls (yes, I'm guilty of this one) .

I think the conceptual issue is that SanitizeString is used for two purposes:

• to avoid formatting strings that are dangerous to pass to the shell, terminal or debug log, and
• to filter characters for user agent

Splitting of these two use-cases with a parameter was a great idea. Now we can make SanitizeString(SAFE_CHARS_UA_COMMENT) pass through all characters allowed per BIP14.

Unfortunately that BIP isn't clear on that. It says that / : ( and ) are reserved, but it doesn't make any statement about e.g. which character set is used, or even about control characters...

[maflcko]

Assuming ASCII and forgetting about all non printable ASCII chars, leaves us with 33 chars (alphanum is fine) to make a decision about.

.,;_/:?@() is already in, so what is left:

 !"#$%&'*+-<=>[]\^`{|}~

IIrc, UAs get dumped to debug.log as well, so is it save to allow any char?

[jgarzik]

ut ACK

[laanwj]

IIrc, UAs get dumped to debug.log as well, so is it save to allow any char?

Yes, but only processed through SanitizeString(SAFE_CHARS_DEFAULT) - in principle, just the fact that we don't log some characters doesn't mean that they shouldn't be allowed in uacomments.

[laanwj]

Meh, I don't think it's worth it to take this to BIP level - although the BIP should have mentioned valid/invalid characters! Good to check for in next proposal if anything relating strings.
utACK