setuid/setgid bits

2005-07-06 Thread Gordon Lack
version: rsync v2.6.1 (+ a minor, unrelated patch). I'm rsyncing files (not as root) and am happy (indeed, for what I want, delighted) that the files at the target side end up owned by the account doing the rsync. However, I've found that if I have a setuid/setgid file on the source side, t

Hang on remote FS full (2.6.1)

2004-12-17 Thread Gordon Lack
Bug in rsync (version 2.6.1) over rsh. Remote end (destination) gets a Disk Full error when writing the (temporary) updated file. remote ends aborts. Client end sits there forever waiting for a reply. -- To unsubscribe or change options: https://lists.samba.org/mailman/listinfo/rsy

Bug: in 2.6.1

2004-09-17 Thread Gordon Lack
directory). But using --delete on a job by default isn't a good idea! Is there some way to get symlinks to replace non-empty directories without using --delete? -- Gordon Lack Application Environmentalist -- To unsubscribe or change options: http://lists.samba.org/mailman/listinfo/rsync B

large file support in 2.6.1

2004-04-30 Thread Gordon Lack
The large file support in configure tests for working file-locking. This may fail if you are building over NFS and the NFS locking isn't working. rsync doesn't actually use file-locking, so this test causes it to remove large-file support unnecessarily. A workaround might be to bui

rsync 2.6.1 fails tests on OSF1

2004-04-30 Thread Gordon Lack
At least it does for me on v5.1 It is a result of a bug in sh on OSF1. "$var"? doesn't do file-globbing if $var contains a / (it does if it doesn't). This causes a touch command in teh merge test to fail. WORKAROUND: Set BIN_SH=xpg4 in your environment while running the test. Thi

Bug in rsync 2.6.1

2004-04-28 Thread Gordon Lack
It fails to compile on Irix (compiler 7.2.1). Whereas rsync.h has a #define for __attribute__(x) to remove it for non GNUC and APPLE, this isn't used when compiling popt/popt.c, so it produces an error as it doesn't like __attribute__((__unused__)) showing up in the parameter list for pop

Bug in 2.6.1

2004-04-28 Thread Gordon Lack
I'm not sure what the code is trying to do, but the SGI Irix compiler spotted this usage of an undefined variable. In uidlist.c: === static int map_uid(int id, char *name) { uid_t uid; if (uid != 0 && name_to_uid(name, &uid)) return uid; return