John Baldwin <[EMAIL PROTECTED]> writes: > Ran into this at work. Suppose I am fetching a file over ftp for a user foo > whose home dir is /home/foo' and I want to retrieve the 'bar/baz' file out of > his home directory. [...]
This is tricky because libfetch caches FTP connections, so you can't always assume that CWD is ~ at the beginning of the transfer if you've already transferred something from the same host/user/password tuple in the same process. See (ironically) RFC 1738 section 3.2.5. Connection caching is an essential feature, because it has a huge impact on the speed of a network install. Although sysinstall still uses libftpio + cpio, one may hope that it will one day in the not- too-distant future switch to libfetch + libarchive. > I verified this by doing 'fetch -vv' and seeing that it does one big CWD > (instead of the multiple CWD's the RFC says should happen) and that it > includes the leading '/' when it should not. I am currently testing a patch for this. > Also, it seems that fetch ignores the ';type=X' optional part of the url-path > and always uses TYPE I. Correct, that was never implemented. At the very least, the typecode should be stripped from the URL. DES -- Dag-Erling Smørgrav - [EMAIL PROTECTED] _______________________________________________ freebsd-net@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "[EMAIL PROTECTED]"
