[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
reserved requests, cord message names
Erik Ostrom writes:
> FWIW, I sort of like John's idea of a reserved request name that
> stands for "not an MCP message". My only qualm is that the MOO
> implementation is a little bit tricky. Assuming somebody actually
> USES the #$#text-line message, the MCP dispatcher has to use
> force_input() to get the regular server parser to deal with the line.
> I guess that's not so bad, it just feels vaguely like voodoo.
>
ok; there seems to be some agreement this is a good idea. if so, i'm
willing to put it in... i'm not comfortable with the name, however.
the message should either be a special format (like #$#*) or be in the
mcp- namespace, obviously. my vote is for something like either
#$#mcp-inband 21409321 msg: ...
or
#$#> 3210953 msg: ...
the reason for the latter is mostly 'cause it seems like for something
like this you want behavior similar to the behavior on multiline value
lines; that is, i think you want the bulk of the message to be opaque
to the parser to avoid having to worry about silly quoting issues just
to get a line of text across.
as for parser voodoo, hrm. i don't have a good answer for this. yep,
i think this'll be a pain to deal with in certain cases, but i can see
where it'd be a good thing.
another, smaller issue: it's been a while; does anyone seriously
object to my prefixing the #$#cord* messages with mcp- ?
another one, that i never properly responded to: there were complaints
a while back that the null message (without a dash, like '#$#cord')
should go away. i really like these messages, but i'm willing to get
rid of them; they DO make parsers somewhat trickier. votes?
dk