__dead before the type

kamiftp: accept an optional 9p:// prefix in the connstring

kamiftp: add note to drop argv[1] handling after 0.3

kamiftp: for bell mode print \a to stderr

kamiftp: deprecate -c it's useless as -C already implies it and -C must be use when connecting with TLS. To be definitely removed after 0.3.

kamiftp: add -o to chose the path where to save the named file

kamiftp: print diognistic messages to stderr excluding the one from cmd_*

kamiftp: automatic cd or fetch on the given path before the given path was used for Tattach. Turns out at least u9fs doesn't seem to use that field for the initial directory (not sure if it should). kamiftp now always issues a Tattach with aname="/" and then does a Twalk on the path: if it names a directory, it becomes the remote working directory like a `cd' was issued, if it names a file it is fetched and kamiftp quits.

kamiftp: always use "/" for attach

kamiftp: use the [user@]host[:port][/path] syntax instead of taking the path as a separate argument. For some time the old style will be supported.

refactor kamiftp internals use a FILE (constructed either via fdopen over a socket or funopen over libtls) for remote I/O

ftp: issue slightly smaller requests to please u9fs u9fs like to return "i/o count too large" on reads/writes that exceeds msize - 24. seems arbiratry, as in theory we should be able to issue a Tread/Twrite for msize - (HEADERSIZE + 4) = msize - 13. Don't know where the other eleven bits come from. To conclude the rant: even if a client issues a Tread/write too large, why don't just return a value smaller than requested? It's explicitly documented in the plan9 manpage for Tread.

ftp: allow user@host syntax; guard against empty user or port

do optimal (i.e. maximum) reads and writes

cmd_ls: fix read size don't use a constant: msize may be lower than that. Instead, use `msize - 4' which is guaranteed to be the maximum transferreable size. (it's not possible for msize to be lower than 4 since we reject ridiculously small msizes, so that difference can't underflow.)