home *** CD-ROM | disk | FTP | other *** search
- Allow for account details after the password if response 332 is given.
-
- Make ftp.pl just check the first digit of response codes to be RFC 1123
- conformant.
-
- The xxx_to_go totals still ain't 100% correct, since name_mapping's aren't considered. (pi).
-
-
- Cache the work list and just restart that if less than a given age.
- Allow local dir scan to be done before connection is made.
- Cache the local/remote dir listing and use those if less than a given age.
-
- limit the amount a mirrored area can grow to.
- don't mirror unless free space (Joachim)
-
- allow both flock and fcntl locks (see plexus)
-
- Allow per-filestore defaults:
- default_fs_settings=unix{
- recurse_hard=false
- ....
- }
-
- Handle hard-links:
- In theory add 'i' to the ls flags, keep the inode number and spot
- files that are really the same by having a reverse assoc lookup...
- Also check file size, type, .... just in case the remote area is
- actually several disks. On recreateing the link have to check that
- everything is on the same disks too.
-
- Track down severe memory leak on processing ftp'dir_open -> get_remote_dir..
- that doesn't seem to be there if local_file -> get_remote_dir..
-
-
- If I have a file here, foo.tar, which is over split_min, then when the
- mirror next runs, it will get a new foo.tar and split it to
- foo.tar-split/, but not remove the old foo.tar. (tale@uunet)
-
-
- Also, should there be an option as to whether people would want fetch
- both FILE.Z and FILE.gz? Mirror already doesn't give any sort of
- option between FILE and FILE.{Z or gz} --- it just takes the
- compresssed one. If we're interested in saving space, why do we take
- two compressed copies? (tale@uunet)
-
-
- Requests pending:
-
- Masanobu UMEDA <umerin@orchid.mse.kyutech.ac.jp>
- uncompress on pulling files.
-
-
- mm additions from x920031@rubb.rz.ruhr-uni-bochum.de:
- } a) at which time of day can it be mirrored
- } b) retry times for different conditions (no contact, transfers failed,
- } mirror was ok, etc.)
- } c) a measure of the load this particular mirror will produce
- } (different mirrors may have weights, and I like to limit the total
- } weight)
-
-
- Pre-size the remote_* arrays based on the size of the local ones.
-