rsync.git
17 years agoMust use STRUCT_STAT not "struct stat" to be compatible with other
Martin Pool [Mon, 18 Feb 2002 22:49:00 +0000 (22:49 +0000)]
Must use STRUCT_STAT not "struct stat" to be compatible with other
rsync functions.

17 years agoCope with BSD systems on which mkdir() will not accept a trailing
Martin Pool [Mon, 18 Feb 2002 22:44:23 +0000 (22:44 +0000)]
Cope with BSD systems on which mkdir() will not accept a trailing
slash.

<http://www.opensource.apple.com/bugs/X/BSD%20Kernel/2734739.html>

17 years agoIf the daemon is unable to fork a child to accept a connection, print
Martin Pool [Mon, 18 Feb 2002 22:38:03 +0000 (22:38 +0000)]
If the daemon is unable to fork a child to accept a connection, print
an error message.  (Colin Walters)

17 years agoAdd -vvv trace statement to set_modtime to help with Debian bug
Martin Pool [Mon, 18 Feb 2002 22:25:55 +0000 (22:25 +0000)]
Add -vvv trace statement to set_modtime to help with Debian bug
#100295.

17 years agoThis test must specify --times because it compares listings that
Martin Pool [Mon, 18 Feb 2002 22:14:46 +0000 (22:14 +0000)]
This test must specify --times because it compares listings that
include mtimes.

17 years agoFix bug that made tls.o not be removed by 'make clean'.
Martin Pool [Mon, 18 Feb 2002 22:07:44 +0000 (22:07 +0000)]
Fix bug that made tls.o not be removed by 'make clean'.

17 years agoDoc:
Martin Pool [Mon, 18 Feb 2002 21:46:49 +0000 (21:46 +0000)]
Doc:

  #defiine lchown chown

could be bad on systems which have no lchown and where chown
follows symbollic links.  On such systems it might be better not to
try to chown symlinks at all.

17 years agoIgnore SIGPIPE and allow EPIPE to get through to the program so that
Martin Pool [Mon, 18 Feb 2002 20:06:57 +0000 (20:06 +0000)]
Ignore SIGPIPE and allow EPIPE to get through to the program so that
we don't get stuck in a recursive loop trying to report a broken pipe
across that same broken pipe.  Debian bug #128632 (Colin Walters)

17 years agoFix for rsync server processes hanging around after the client
Martin Pool [Mon, 18 Feb 2002 19:54:00 +0000 (19:54 +0000)]
Fix for rsync server processes hanging around after the client
unexpectedly disconnects.  (Colin Walters) (Debian bug #128632)

17 years agorwrite: Doc.
Martin Pool [Mon, 18 Feb 2002 19:51:12 +0000 (19:51 +0000)]
rwrite: Doc.

17 years agoDocument multiplex stuff.
Martin Pool [Mon, 18 Feb 2002 19:44:04 +0000 (19:44 +0000)]
Document multiplex stuff.

17 years agoDoc.
Martin Pool [Mon, 18 Feb 2002 19:10:28 +0000 (19:10 +0000)]
Doc.

17 years agoIf a daemon prints an error message of @ERROR, have the client treat the
David Dykstra [Mon, 18 Feb 2002 18:29:48 +0000 (18:29 +0000)]
If a daemon prints an error message of @ERROR, have the client treat the
message as an FERROR rather than an FINFO.

17 years agoAdd item about fixing "out of memory in flist_expand" on Sunos4.
David Dykstra [Thu, 14 Feb 2002 15:27:55 +0000 (15:27 +0000)]
Add item about fixing "out of memory in flist_expand" on Sunos4.

17 years agoAdded the two most important bug fixes to NEWS to make sure they're not
David Dykstra [Wed, 13 Feb 2002 18:57:06 +0000 (18:57 +0000)]
Added the two most important bug fixes to NEWS to make sure they're not
forgotten for the next release.

17 years agoPatch from Jos Backus <josb@cncdsl.com> to use HAVE_SOCKADDR_LEN rather
David Dykstra [Wed, 13 Feb 2002 18:45:17 +0000 (18:45 +0000)]
Patch from Jos Backus <josb@cncdsl.com> to use HAVE_SOCKADDR_LEN rather
than HAVE_SOCK_SIN_LEN around use of sin_len.  Correct usage was already
in place in clientname.c.

17 years agoSuggested patch from Jim Ogilvie <jogilvie@us.ibm.com> to print out the
David Dykstra [Wed, 13 Feb 2002 18:42:20 +0000 (18:42 +0000)]
Suggested patch from Jim Ogilvie <jogilvie@us.ibm.com> to print out the
system error message when mkstemp fails.

17 years agoReversing the order of maybe_emit_filelist_progress() and
David Dykstra [Wed, 13 Feb 2002 18:30:27 +0000 (18:30 +0000)]
Reversing the order of maybe_emit_filelist_progress() and
emit_filelist_progress() makes the native compilers on systems
including Solaris and Irix happier.

17 years agoSome systems, notably Sunos4, do not support realloc(NULL, n), so if
David Dykstra [Wed, 13 Feb 2002 18:06:36 +0000 (18:06 +0000)]
Some systems, notably Sunos4, do not support realloc(NULL, n), so if
nothing has yet been malloced in flist_expand(), call malloc instead of
realloc.  Problem introduced in revision 1.106 of flist.c on January 25.

17 years agoAt least change INO64_T and DEV64_T back to just 'int64', not
Martin Pool [Wed, 13 Feb 2002 02:57:55 +0000 (02:57 +0000)]
At least change INO64_T and DEV64_T back to just 'int64', not
'unsigned int64'.  This should fix some compile problems on machines
where int64 is not a simple integer type, but I'm not convinced it is
the ideal fix.

17 years agoFactor out code for filelist progress. Copy&paste considered harmful.
Martin Pool [Wed, 13 Feb 2002 02:44:31 +0000 (02:44 +0000)]
Factor out code for filelist progress.  Copy&paste considered harmful.

Add a little doc about potential optimization of stat() calls.

17 years agoAdd a test that --owner correctly propagates ownership of files for a
Martin Pool [Sat, 9 Feb 2002 07:43:13 +0000 (07:43 +0000)]
Add a test that --owner correctly propagates ownership of files for a
local transfer as root.

17 years agoAdd a test that --group correctly propagates groups of which the local
Martin Pool [Sat, 9 Feb 2002 07:42:37 +0000 (07:42 +0000)]
Add a test that --group correctly propagates groups of which the local
user is a member.

17 years agoAdd $preserve_scratch and $always_log so if you want to see details
Martin Pool [Sat, 9 Feb 2002 03:36:33 +0000 (03:36 +0000)]
Add $preserve_scratch and $always_log so if you want to see details
about successful tests, you can.

17 years agoOops -- have to call setgroups() before giving up root.
Martin Pool [Sat, 9 Feb 2002 03:30:22 +0000 (03:30 +0000)]
Oops -- have to call setgroups() before giving up root.

Doc some of the peculiarities about starting rsyncd as root vs
non-root.

17 years agoMake sure we call setgroups() after setuid(). (Ethan Benson)
Martin Pool [Sat, 9 Feb 2002 02:18:42 +0000 (02:18 +0000)]
Make sure we call setgroups() after setuid().  (Ethan Benson)

17 years agoRemove the "rsync:" prefixes on FINFO messages. Return the "building file
David Dykstra [Thu, 7 Feb 2002 16:36:12 +0000 (16:36 +0000)]
Remove the "rsync:" prefixes on FINFO messages.  Return the "building file
list ... done" to the way it was in 2.5.1 and before when not using -P.
Apply the file list progress messages when receiving files in addition to
sending files.

17 years agoMake batch mode actually work and add man page documentation. From Jos Backus.
David Dykstra [Wed, 6 Feb 2002 21:20:48 +0000 (21:20 +0000)]
Make batch mode actually work and add man page documentation.  From Jos Backus.

17 years agoMerge modified --with-rsh patch: we now determine the default
Martin Pool [Wed, 6 Feb 2002 04:37:09 +0000 (04:37 +0000)]
Merge modified --with-rsh patch: we now determine the default
remote-execution command as follows:

 1) if --with-rsh is specified, use that.

 2) otherwise if remsh is in the path, use that.

 3) otherwise use rsh

If remsh is present, we always modify the order of parameters to suit
it.  This is a bit strange.

17 years agoDoc: Rusty's /*/* exclude hack produces spurious output with -vv.
Martin Pool [Wed, 6 Feb 2002 04:34:40 +0000 (04:34 +0000)]
Doc: Rusty's /*/* exclude hack produces spurious output with -vv.

17 years agoAdd --no-whole-file and --no-blocking-io options
David Dykstra [Tue, 5 Feb 2002 23:05:31 +0000 (23:05 +0000)]
Add --no-whole-file and --no-blocking-io options

17 years agoOnly print the command used to open connections with -vv, not just -v.
Martin Pool [Tue, 5 Feb 2002 00:37:53 +0000 (00:37 +0000)]
Only print the command used to open connections with -vv, not just -v.

17 years agoRoll over NEWS
Martin Pool [Tue, 5 Feb 2002 00:35:31 +0000 (00:35 +0000)]
Roll over NEWS

17 years agoOnly print the command used to open connections with -vv, not just -v.
Martin Pool [Tue, 5 Feb 2002 00:34:03 +0000 (00:34 +0000)]
Only print the command used to open connections with -vv, not just -v.

<Pine.LNX.4.33L2.0201301015260.11155-100000@phong.blorf.net>

17 years agoPatch from wayned so that add_exclude_list produces clearer debugging
Martin Pool [Tue, 5 Feb 2002 00:25:52 +0000 (00:25 +0000)]
Patch from wayned so that add_exclude_list produces clearer debugging
output with -vvv.

17 years agothe signed/unsigned change seems to have caused a logic bug on some
Andrew Tridgell [Sun, 3 Feb 2002 01:38:39 +0000 (01:38 +0000)]
the signed/unsigned change seems to have caused a logic bug on some
systems (only those without large file support perhaps?)

this fixes it

17 years agoBetter explanation of --force. It is applicable whenever --delete is
David Dykstra [Tue, 29 Jan 2002 21:52:57 +0000 (21:52 +0000)]
Better explanation of --force.  It is applicable whenever --delete is
not in effect.

17 years agoA more accurate description of --force as I know it.
David Dykstra [Mon, 28 Jan 2002 21:09:03 +0000 (21:09 +0000)]
A more accurate description of --force as I know it.

17 years agoClarify the --force entry in the rsync man page.
David Dykstra [Mon, 28 Jan 2002 17:06:04 +0000 (17:06 +0000)]
Clarify the --force entry in the rsync man page.

17 years agoOops, version should be just 2.5.2. v2.5.2
Martin Pool [Fri, 25 Jan 2002 23:19:21 +0000 (23:19 +0000)]
Oops, version should be just 2.5.2.

17 years agoBump version to 2.5.3.
Martin Pool [Fri, 25 Jan 2002 23:16:18 +0000 (23:16 +0000)]
Bump version to 2.5.3.

17 years agoMore signedness fixes; should be harmless.
Martin Pool [Fri, 25 Jan 2002 23:07:33 +0000 (23:07 +0000)]
More signedness fixes; should be harmless.

17 years agoio_end_buffering doesn't need (or use) it's fd parameter: there's only
Martin Pool [Fri, 25 Jan 2002 23:01:50 +0000 (23:01 +0000)]
io_end_buffering doesn't need (or use) it's fd parameter: there's only
one multiplexed stream.

17 years agoDEV64_t and INO64_T should probably be unsigned
Martin Pool [Fri, 25 Jan 2002 23:00:21 +0000 (23:00 +0000)]
DEV64_t and INO64_T should probably be unsigned

17 years agocheck_name doesn't need a socklen_t, because it knows what is inside
Martin Pool [Fri, 25 Jan 2002 22:59:37 +0000 (22:59 +0000)]
check_name doesn't need a socklen_t, because it knows what is inside
each sockaddr type.

17 years agoAdd test case for device nodes. This test will be skipped unless you
Martin Pool [Fri, 25 Jan 2002 10:56:43 +0000 (10:56 +0000)]
Add test case for device nodes.  This test will be skipped unless you
run "make check" as root.

17 years agoWhen comparing directories, use `find .` to call diff, rather than
Martin Pool [Fri, 25 Jan 2002 10:55:59 +0000 (10:55 +0000)]
When comparing directories, use `find .` to call diff, rather than
diff -r.  Two reasons: diff -r might not work everywhere, and it also
might complain about nonregular files.

17 years agoAdd test_skipped function.
Martin Pool [Fri, 25 Jan 2002 10:47:47 +0000 (10:47 +0000)]
Add test_skipped function.

17 years agoNew --ignore-existing option, patch previously distributed with
Martin Pool [Fri, 25 Jan 2002 10:42:23 +0000 (10:42 +0000)]
New --ignore-existing option, patch previously distributed with
Vipul's Razor.  (Debian #124286)

17 years agoFix for device nodes. (dann frazier) (Debian #129135)
Martin Pool [Fri, 25 Jan 2002 10:39:08 +0000 (10:39 +0000)]
Fix for device nodes.  (dann frazier)   (Debian #129135)

17 years agoFix for device nodes. (dann frazier) (Debian #129135)
Martin Pool [Fri, 25 Jan 2002 10:28:13 +0000 (10:28 +0000)]
Fix for device nodes.  (dann frazier)   (Debian #129135)

17 years agoAnother DEV64_T change.
Martin Pool [Fri, 25 Jan 2002 10:16:11 +0000 (10:16 +0000)]
Another DEV64_T change.

17 years agoWith -vv, when the file list grows, show a message.
Martin Pool [Fri, 25 Jan 2002 10:12:36 +0000 (10:12 +0000)]
With -vv, when the file list grows, show a message.

17 years agoWith -vv, when the file list grows, show a message.
Martin Pool [Fri, 25 Jan 2002 10:12:02 +0000 (10:12 +0000)]
With -vv, when the file list grows, show a message.

17 years agoUndo overzealous deletion.
Martin Pool [Fri, 25 Jan 2002 10:09:00 +0000 (10:09 +0000)]
Undo overzealous deletion.

17 years agoAdd dummy show_flist_stats().
Martin Pool [Fri, 25 Jan 2002 10:06:36 +0000 (10:06 +0000)]
Add dummy show_flist_stats().

17 years agoFactor out code to grow the file list into a common location.
Martin Pool [Fri, 25 Jan 2002 10:05:49 +0000 (10:05 +0000)]
Factor out code to grow the file list into a common location.

17 years agoindent -kr -i8
Martin Pool [Fri, 25 Jan 2002 09:59:00 +0000 (09:59 +0000)]
indent -kr -i8

17 years agoAdd link to the message that introduced string_area.
Martin Pool [Fri, 25 Jan 2002 09:54:21 +0000 (09:54 +0000)]
Add link to the message that introduced string_area.

17 years agomallinfo is implemented.
Martin Pool [Fri, 25 Jan 2002 09:45:45 +0000 (09:45 +0000)]
mallinfo is implemented.

17 years agoNotes about flist.
Martin Pool [Fri, 25 Jan 2002 09:44:17 +0000 (09:44 +0000)]
Notes about flist.

17 years agoIgnore autoconf fluff.
Martin Pool [Fri, 25 Jan 2002 02:53:48 +0000 (02:53 +0000)]
Ignore autoconf fluff.

17 years agoAdd code to compare sin6_addrs. v2.5.2pre3
Martin Pool [Fri, 25 Jan 2002 02:45:09 +0000 (02:45 +0000)]
Add code to compare sin6_addrs.

17 years agoBack out last change -- to see whether an address is spoofed, we don't
Martin Pool [Fri, 25 Jan 2002 02:43:35 +0000 (02:43 +0000)]
Back out last change -- to see whether an address is spoofed, we don't
want to look at the whole sockaddr, but rather just at the sin_addr
that it contains.

Also fix silly bug where ai_flags was set incorrectly for getaddrinfo.

17 years agocompare_addrinfo_sockaddr: Add code to compare AF_INET6 addresses.
Martin Pool [Fri, 25 Jan 2002 02:37:20 +0000 (02:37 +0000)]
compare_addrinfo_sockaddr: Add code to compare AF_INET6 addresses.

17 years agoSplit out code to compare addrinfo and sockaddr into it's own
Martin Pool [Fri, 25 Jan 2002 02:29:53 +0000 (02:29 +0000)]
Split out code to compare addrinfo and sockaddr into it's own
function.  The comparison cannot be done just byte-by-byte, because
different parts of the sockaddr will be meaningful depending on the
protocol.  It looks like on some systems the library sets the unused
parts to 0, but this is not reliable.  IPv6 not implemented yet.

17 years agoDoc.
Martin Pool [Fri, 25 Jan 2002 02:15:58 +0000 (02:15 +0000)]
Doc.

17 years agoThe name resolution stuff is getting complicated -- split it out into
Martin Pool [Fri, 25 Jan 2002 02:13:04 +0000 (02:13 +0000)]
The name resolution stuff is getting complicated -- split it out into
its own file.

17 years agoVerbose messages for spoof check... doesn't work on old linux libc?
Martin Pool [Thu, 24 Jan 2002 09:42:52 +0000 (09:42 +0000)]
Verbose messages for spoof check... doesn't work on old linux libc?

17 years agoWhen doing a name->addr translation to check for spoofing, give the
Martin Pool [Thu, 24 Jan 2002 08:52:28 +0000 (08:52 +0000)]
When doing a name->addr translation to check for spoofing, give the
resolver the address family of the original address as a hint, so that
we're more likely to find the correct A or AAAA record.,

17 years agoMessage on successful configuration.
Martin Pool [Thu, 24 Jan 2002 08:50:35 +0000 (08:50 +0000)]
Message on successful configuration.

17 years agoFix comment.
Martin Pool [Thu, 24 Jan 2002 08:42:21 +0000 (08:42 +0000)]
Fix comment.

Bump version to 2.5.2pre3

17 years agoName resolution on machines supporting IPv6 is improved. v2.5.2pre2
Martin Pool [Thu, 24 Jan 2002 08:20:51 +0000 (08:20 +0000)]
Name resolution on machines supporting IPv6 is improved.

17 years agoignore gmon.out
Martin Pool [Thu, 24 Jan 2002 08:19:28 +0000 (08:19 +0000)]
ignore gmon.out

17 years agoFix getpeername call.
Martin Pool [Thu, 24 Jan 2002 08:16:27 +0000 (08:16 +0000)]
Fix getpeername call.

17 years agosize_t fix.
Martin Pool [Thu, 24 Jan 2002 08:09:46 +0000 (08:09 +0000)]
size_t fix.

17 years agowrite_batch_csums_file: Opaque IO buffers should be void*.
Martin Pool [Thu, 24 Jan 2002 08:08:56 +0000 (08:08 +0000)]
write_batch_csums_file: Opaque IO buffers should be void*.

17 years agoFix cast that was breaking HP/UX.
Martin Pool [Thu, 24 Jan 2002 08:07:35 +0000 (08:07 +0000)]
Fix cast that was breaking HP/UX.

17 years agoMove both calls to getpeername into a common wrapper function that
Martin Pool [Thu, 24 Jan 2002 08:05:22 +0000 (08:05 +0000)]
Move both calls to getpeername into a common wrapper function that
handles IPV4_MAPPED addresses.

17 years agoMust get declarations from addrinfo.h before prototypes.
Martin Pool [Thu, 24 Jan 2002 07:22:38 +0000 (07:22 +0000)]
Must get declarations from addrinfo.h before prototypes.

17 years agoAnother size_t warning
Martin Pool [Thu, 24 Jan 2002 05:57:49 +0000 (05:57 +0000)]
Another size_t warning

17 years agocheck_name: Print out name *before* clobbering it.
Martin Pool [Thu, 24 Jan 2002 05:57:22 +0000 (05:57 +0000)]
check_name: Print out name *before* clobbering it.

17 years agoTry to fix error on Solaris:
Martin Pool [Thu, 24 Jan 2002 05:54:47 +0000 (05:54 +0000)]
Try to fix error on Solaris:

+ [ -f /export/home/build/build_farm/rsync/testtmp.symlink-ignore/to/referent ]
+ [ -d /export/home/build/build_farm/rsync/testtmp.symlink-ignore/to/from ]
+ [ -L /export/home/build/build_farm/rsync/testtmp.symlink-ignore/to/dangling ]
./testsuite/symlink-ignore.test: test: argument expected

Solaris says it supports -L, so I'm not sure what's happening...

17 years agoIf name lookup fails, then show the relevant IP address in the error message.
Martin Pool [Thu, 24 Jan 2002 05:41:46 +0000 (05:41 +0000)]
If name lookup fails, then show the relevant IP address in the error message.

17 years agoAt the connections that just get a list of modules are not logged,
Martin Pool [Thu, 24 Jan 2002 04:49:07 +0000 (04:49 +0000)]
At the connections that just get a list of modules are not logged,
but they should be.

17 years agorsync_module: If host-based access fails, show the exact name/address
Martin Pool [Thu, 24 Jan 2002 04:41:09 +0000 (04:41 +0000)]
rsync_module: If host-based access fails, show the exact name/address
used for the check in the error message.  (Just in case...)

17 years agoShow command used to start connection child.
Martin Pool [Thu, 24 Jan 2002 04:36:00 +0000 (04:36 +0000)]
Show command used to start connection child.

17 years agoindent -kr -i8
Martin Pool [Thu, 24 Jan 2002 04:26:55 +0000 (04:26 +0000)]
indent -kr -i8

17 years agoDoc.
Martin Pool [Thu, 24 Jan 2002 04:24:12 +0000 (04:24 +0000)]
Doc.

Oops, connection program message was in the wrong place.

17 years agoShow helper program with -v when opening connection.
Martin Pool [Thu, 24 Jan 2002 04:21:42 +0000 (04:21 +0000)]
Show helper program with -v when opening connection.

Doc.

17 years agoDoc.
Martin Pool [Thu, 24 Jan 2002 04:19:41 +0000 (04:19 +0000)]
Doc.

17 years agoFix const
Martin Pool [Thu, 24 Jan 2002 04:07:07 +0000 (04:07 +0000)]
Fix const

17 years agoUnbreak the old behavior of using UNKNOWN as a hostname if any of the
Martin Pool [Thu, 24 Jan 2002 04:03:06 +0000 (04:03 +0000)]
Unbreak the old behavior of using UNKNOWN as a hostname if any of the
addr->name->addr translations fail, because people might count on this
in "hosts deny" lines.

17 years agoFix inverted sense of error check.
Martin Pool [Thu, 24 Jan 2002 03:31:28 +0000 (03:31 +0000)]
Fix inverted sense of error check.

17 years agoRefactor client_name() into smaller functions.
Martin Pool [Thu, 24 Jan 2002 03:28:20 +0000 (03:28 +0000)]
Refactor client_name() into smaller functions.

Better messages for DNS failure.

If we can get a reverse name for an IP address, but not confirm that
it is correct using a forward lookup then we still proceed to use the
name, but also emit a warning.

17 years agoDoc.
Martin Pool [Thu, 24 Jan 2002 03:03:20 +0000 (03:03 +0000)]
Doc.

17 years agoBump version to 2.5.2pre2.
Martin Pool [Thu, 24 Jan 2002 03:01:42 +0000 (03:01 +0000)]
Bump version to 2.5.2pre2.

17 years agolog_formatted: Fill the log buffer with nuls to make sure we cannot
Martin Pool [Thu, 24 Jan 2002 02:41:38 +0000 (02:41 +0000)]
log_formatted: Fill the log buffer with nuls to make sure we cannot
accidentally leave the string unterminated.

17 years agoAnother harmless size_t warning.
Martin Pool [Thu, 24 Jan 2002 02:33:45 +0000 (02:33 +0000)]
Another harmless size_t warning.

17 years agoDoc. v2.5.2pre1
Martin Pool [Wed, 23 Jan 2002 08:08:48 +0000 (08:08 +0000)]
Doc.