parameter multiplicity of LIST

[progval]

the spec for LIST writes:

  Parameters: [<channel>{,<channel>}] [<elistcond>{,<elistcond>}]

But this seems to be wrong in practice, in several ways.

In the following tests, #chan1 has 1 user, #chan2 has 2, and #chan3 has 1:

LIST #chan1,#chan2:

• returns nothing on hybrid/plexus4, inspircd
• returns both #chan1 and #chan2 on: irc2, ircu2/nefarious, ngircd, unrealircd (<- current specced behavior)
• n/a on charybdis/solanum and ergo, because LIST:1 is in TARGMAX

LIST *an1,*an2:

• returns both #chan1 and #chan2 on: unrealircd (<- current specced behavior)
• n/a on others (either because LIST:1 is in TARGMAX or because they don't have M in ELIST)

LIST #chan1 <1:

• returns #chan1 on: charybdis/solanum, hybrid/plexus4, unrealircd
• nothing on: ergo, inspircd (<- current specced behavior)
• n/a on bahamut, irc2, ircu2/nefarious, ngircd (because they don't have U in ELIST)

LIST #chan1,#chan2 >0:

• not tested on charybdis/solanum, hybrid/plexus4, unrealircd (considering the previous result)
• nothing on inspircd
• spec behavior would be to return both #chan1 and #chan2
• n/a on bahamut, irc2, ircu2/nefarious, ngircd (because they don't have U in ELIST)
• n/a on charybdis/solanum and ergo, because LIST:1 is in TARGMAX

(full logs: https://pr-193-6559bb34c0--radiant-beignet-ceec1b.netlify.app/irctest.server_tests.list.xhtml ; examples above are respectively from testListTwoChannels, testListTwoMasks, testListTwoParams, and testListTwoParams again)

[ghost]

Hello, I commented already on the channel:

At least on solanum (libera), where I tested this,

Parameters: [<channel>{,<channel>}] [<elistcond>{,<elistcond>}]

channel (or mask) and elistcond are not two whitespace-separated params to LIST, but have to be passed as one single parameter, separated by a comma.

So this

LIST *mask* >10

doesnt work and the second parameter is ignored, but this

LIST <10,*mask*,>10

does work, because all three conditions are only one parameter.

That also seems to be the case on unreal but I only have tested it on solanum.

[progval]

@McParen that's because *mask* is not a channel, it's an elist cond (ELIST=M, specifically)

[progval]

and LIST #channel,>10 works too, because masks are a superset of valid channel names