[WIP] Adds support for vCard 2.1

[caugner]

Fixes #187.

• [1] Support QUOTED-PRINTABLE inline encoding [cf. Section 2.1.2].
• [2] Do not interpret "\n" or "\N" as a CRLF(but instead literally as "\", followed by "n" or "N").
• [3] Do not interpret COMMA as a multi-value separator [all values are single values].
• [3] Note: Despite what vCard 3.0 states, a SEMICOLON must also be backslash escaped in vCard 2.1. "in a component of a compound property" [cf. Section 2.1.3]
• [4] Require VERSION to be "2.1" [cf. Section 2.6.6].
• [5] [6] n/a
• [7] Support CHARSET type parameter [and default to ASCII, cf. 2.1.6].
• [8] Support non-significant WSP characters [i.e. HTAB, cf. 2.9].
• [9] Don't require parameter name, if the value is unambiguous [e.g. for TYPE= and ENCODING=, cf. 2.1.2].
• [10] Require TYPE of LOGO [cf. 2.5.3.1], PHOTO [cf. 2.2.3.1] and SOUND [cf. 2.6.3.1] as defined by the spec [e.g. GIF].
• [11] Parse inline binary content according to ENCODING parameter [and default to 7-bit, cf. 2.1.5].
• [12] Do not support the PCS type parameter value for TEL [from vCard 3.0].
• [13] Do not support CATEGORIES, CLASS, NICKNAME, PRODID and SORT-STRING [from vCard 3.0].
• [14] Do not require N and FN [Note: N is only required for writers, not for readers, cf. 2.2.2].
• [14] Note: Despite what vCard 3.0 states, VERSION is also required in vCard 2.1 [cf. [4] and 2.6.6].
• [...] Use COMMA to separate GEO values [cf. 2.4.6 versus SEMICOLON in vCard 3.0]
• [MR] Add tests.
• [MR] Fix test coverage.

[skjnldsv]

Awesome!! @caugner
Let me know if I can be any help!

[coveralls]

Coverage decreased (-0.2%) to 96.764% when pulling 23a1d4a on caugner:187-support-vcard-2.1 into 2b4e50d on mozilla-comm:master.

[tasn]

Glad to see you are working on it, life saver!

I tried testing it, but am getting "unsupported vCard version" with a vCard 4.0. Taking a brief look at the code, it look like your switch statement only handles 3.0 and 2.1 (though it could be intentional in that specific spot).

[caugner]

@tasn Can you share the vCard 4.0 which triggered the error? It sounds as if the vCard combines properties of vCard 3 and 4, because this switch statement should only be evaluated for vCards other than 4.0: !(name === 'version' && value === '4.0')

[caugner]

Suggested change

var debug = line.includes("37.386013");

[tasn]

Apologies for the noise!
I just checked following your comment and it seems that the failing one is a purposefully malformed entry I created. Interesting that none of the other clients (ez-vcard, python-vobject) fail when they get to it.

I guess you can ignore my comment. The entry for reference:

BEGIN:VCARD
PRODID;VALUE=TEXT:-//iCal.js EteSync Web
VERSION:4.0
UID:0561483d-251d-44e7-bb63-0cfd6285ffe3
FN:This contact checks broken vcard version
REV;VALUE=DATE-TIME:20190503T091907
IMPP;TYPE=jabber:jabber:
END:VCARD

[stefan123t]

regarding [1] there are two old issues at nextcloud contacts & server with some example ENCODING=QUOTED-PRINTABLE characters of danish and german users in case you want to extend your test case vcard21.vcf

nextcloud/contacts#584
nextcloud/server#10653

[tasn]

I guess this was accidentally closed because of the deletion of master?

[NickCrews]

I am trying to revive this in #958 if anyone wants to help. Still quite rough and not ready for review, I will remove the WIP mark when it is ready.