When libvirt calls virInitialize it creates a thread local
for the virErrorPtr storage, and registers a callback to
cleanup memory when a thread exits. When libvirt is dlclose()d
or otherwise made non-resident, the callback function is
removed from memory, but the thread local may still exist
and if a thread later exists, it will invoke the callback
and SEGV. There may also be other thread locals with callbacks
pointing to libvirt code, so it is in general never safe to
unload libvirt.so from memory once initialized.

To allow dlclose() to succeed, but keep libvirt.so resident
in memory, link with '-z nodelete'. This issue was first
found with the libvirt CIM provider, but can potentially
hit many of the dynamic language bindings which all ultimately
involve dlopen() in some way, either on libvirt.so itself,
or on the glue code for the binding which in turns links
to libvirt

* configure.ac, src/Makefile.am: Ensure libvirt.so is linked
  with -z nodelete
* cfg.mk, .gitignore, tests/Makefile.am, tests/shunloadhelper.c,
  tests/shunloadtest.c: A test case to unload libvirt while
  a thread is still running.

* .gitignore: Exempt recently added generated files.

Add a generator script to generate the structs and serialization
information for OpenWSMAN.

openwsman.h collects workarounds for problems in OpenWSMAN <= 2.2.6.
There are also disabled sections that would use ws_serializer_free_mem
but can't because it's broken in OpenWSMAN <= 2.2.6. Patches to fix
this have been posted upstream.

I did 'git add .' while in the middle of 'make syntax-check', and
it picked up a temporary file that should not be committed.

* .gitignore: Ignore sc_* from syntax check.

* .gitignore: Ignore qemu-sanlock.conf.

Gettext annoyingly modifies CPPFLAGS in-place, putting
-I/usr/local/include into the search patch if libintl headers
must be used from that location.  But since we must support
automake 1.9.6 which lacks AM_CPPFLAGS, and since CPPFLAGS is used
prior to INCLUDES, this means that the build picks up the _old_
installed libvirt.h in priority to the in-tree version, leading
to all sorts of weird build failures on FreeBSD.

Fix this by teaching configure to undo gettext's actions, but
to keep any changes required by gettext at the end of INCLUDES
after all in-tree locations are used first.  Also requires
adding a wrapper Makefile.am and making gnulib-tool create
just gnulib.mk files during the bootstrap process.
Signed-off-by: NEric Blake <eblake@redhat.com>

Add virtkey lib for usage-improvment and keycode translating.
Add 4 internal API for the aim

const char *virKeycodeSetTypeToString(int codeset);
int virKeycodeSetTypeFromString(const char *name);
int virKeycodeValueFromString(virKeycodeSet codeset, const char *keyname);
int virKeycodeValueTranslate(virKeycodeSet from_codeset,
                             virKeycodeSet to_offset,
                             int key_value);

* include/libvirt/libvirt.h.in: extend virKeycodeSet enum
* src/Makefile.am: add new virtkeycode module and rule to generate
  virkeymaps.h
* src/util/virkeycode.c src/util/virkeycode.h: new module
* src/util/virkeycode-mapgen.py: python generator for virkeymaps.h
  out of keymaps.csv
* src/libvirt_private.syms: extend private symbols for new module
* .gitignore: add generated virkeymaps.h

Commit 8665f855 introduced a slight regression in doc generation,
since make only quits a rule on the first failed command ending
with a newline rather than a semicolon.

* docs/Makefile.am (html/index.html): Don't use xmllint unless
xsltproc succeeded.
* .gitignore: Ignore recently updated stamp file name.

* .gitignore: Exempt jsontest binary.

This guts the libvirtd daemon, removing all its networking and
RPC handling code. Instead it calls out to the new virServerPtr
APIs for all its RPC & networking work

As a fallout all libvirtd daemon error reporting now takes place
via the normal internal error reporting APIs. There is no need
to call separate error reporting APIs in RPC code, nor should
code use VIR_WARN/VIR_ERROR for reporting fatal problems anymore.

* daemon/qemu_dispatch_*.h, daemon/remote_dispatch_*.h: Remove
  old generated dispatcher code
* daemon/qemu_dispatch.h, daemon/remote_dispatch.h: New dispatch
  code
* daemon/dispatch.c, daemon/dispatch.h: Remove obsoleted code
* daemon/remote.c, daemon/remote.h: Rewrite for new dispatch
  APIs
* daemon/libvirtd.c, daemon/libvirtd.h: Remove all networking
  code
* daemon/stream.c, daemon/stream.h: Update for new APIs
* daemon/Makefile.am: Link to libvirt-net-rpc-server.la

'make syntax-check' regression introduced in commit 60b9c693.

* tests/networkxml2argvdata/*.argv: Break long lines.

This patch defines the basics of a generic RPC protocol in XDR.
This is wire ABI compatible with the original remote_protocol.x.
It takes everything except for the RPC calls / events from that
protocol

 - The basic header virNetMessageHeader (aka remote_message_header)
 - The error object virNetMessageError  (aka remote_error)
 - Two dummy objects virNetMessageDomain & virNetMessageNetwork
   sadly needed to keep virNetMessageError ABI compatible with
   the old remote_error

The RPC protocol supports method calls, async events and
bidirectional data streams as before

* src/Makefile.am: Add rules for generating RPC code from
  protocol & define a new libvirt-net-rpc.la helper library
* src/rpc/virnetprotocol.x: New generic RPC protocol

* .gitignore: Ignore recently added file.

Stop storing the generated files for the remote protocol client
and server in source control. The generated files will still be
included in the result of 'make dist' to avoid end-users needing
to generate the files
Signed-off-by: NEric Blake <eblake@redhat.com>

Unfortunately, this means that the strings marked for translation
in generated files are not picked up by gnulib's syntax-check,
I'm working on fixing that in gnulib.

* .gitignore, cfg.mk, po/POTFILES.in: Reflect deletion.

* .gitignore: Add exemption for hashtest.

Make it so we don't have to 'git add -f' particular files like
po/POTFILES.in all the time (tested by fixing one of our
special-case files as part of the patch).

* .gnulib: Update to latest.
* bootstrap: Resync from coreutils.
* .gitignore: Sort whitelist entries correctly, including ignoring
files rather than directories.
* m4/virt-compile-warnings.m4: Convert tabs to space.

Split the bit acinclude.m4 file into smaller pieces named
as m4/virt-XXXXX.m4

* .gitignore: Ignore gettext related files
* acinclude.m4: Delete
* m4/virt-compile-warnings.m4: Checks for GCC compiler flags
* m4/virt-pkgconfig-back-compat.m4: Backcompat check for
  pkgconfig program

* .gitignore: Exclude libvirt_iohelper.

Try to avoid future regressions on the VPATH front.

* autobuild.sh: Uncomment VPATH use.
* .gitignore: Exclude build directory.

* configure.ac libvirt.spec.in docs/news.html.in: bump version and add
  documentation
* po/*po*: regenerate po and pot files

* .gitignore: Ignore recent built file, sort.

Allows bootstrap to work on FreeBSD, where gzip doesn't have a '.'
in its version; and silences false positives in the new
'make syntax-check' rule.

* .gnulib: Update to latest.
* bootstrap: Synchronize to upstream.
* .x-sc_bindtextdomain: New exemptions.
* Makefile.am (syntax_check_exceptions): Ship new file.
* .gitignore: Regenerate per latest bootstrap, anchor entries that
are only in the root directory, and consolidate entries from other
generated .gitignore files.
* build-aux/.gitignore, m4/.gitignore, po/.gitignore: Remove from
version control, since bootstrap generates them.

Suggested by danpb, as it's not up-to-date anymore and
lacks many functions that were added to libvirtd.

A missing shell was noisy, and the use of command to decipher a
shell's absolute path requires "" rather than ''.

* configure.ac (lv_cv_wrapper_shell): Fix logic errors if candidate
shell is not available.
* .gitignore: Ignore file created when /bin/sh is old dash.
Reported by Matthias Bolte.

'./autobuild.sh' with lcov installed discovered that our
coverage support has been bit-rotting for a while.  This
restores it back to a successful state, although I have
not yet spent any time looking through the resulting files to
look for low-hanging fruit in the unit test coverage front.

* configure.ac: Clear COMPILER_FLAGS at right place.
* Makefile.am (cov): Newer genhtml no longer likes plain -s.
* m4/compiler-flags.m4 (gl_COMPILER_FLAGS): Don't AC_SUBST
COMPILER_FLAGS; it is a shell variable for use in configure only.
* src/Makefile.am (AM_CFLAGS, AM_LDFLAGS): New variables, to make
it easier to provide global flag additions.  Use throughout, to
uniformly apply coverage flags.
* .gitignore: Globally ignore gcov output.
* daemon/.gitignore: Simplify.
* src/.gitignore: Likewise.
* tests/.gitignore: Likewise.

Commit f30ccb24 was incomplete.

* .gitignore: Ignore intermediate directory during failed 'make
distcheck'.
* tests/Makefile.am (qemuhelpdata): Distribute new file.

Necessary on cygwin, where uid_t and gid_t are 4-byte long rather
than int, causing gcc -Wformat warnings.

* src/util/util.c (virFileOperationNoFork, virDirCreateNoFork)
(virFileOperation, virDirCreate, virGetUserEnt): Cast uid_t and
gid_t before passing to printf.
* .gitignore: Ignore Windows executables.

* .gitignore: Add exemption.
* cfg.mk (local-checks-to-skip): Ignore a test to silence a skip
warning.

* .x-sc_prohibit_gettext_noop: Add new exemption.
* .x-sc_prohibit_test_minus_ao: Likewise.
* Makefile.am (EXTRA_DIST): Distribute new files.
* .gitignore: Ignore built file.

Fallout from the new bootstrap.

* .gitignore: Commit sorting done by bootstrap.
* build-aux/.gitignore: Likewise.
* po/.gitignore: Likewise.

Removes some auto-generated files still under version control.
It also moves the rule for generating NEWS into the Makefile.am
that's in the same directory as the output file to avoid confusion

* docs/libvirt-api.xml, docs/libvirt-refs.xml, NEWS: Remove
  auto-generated files from source control
* Makefile.am: Add rule for generating NEWS file
* docs/Makefile.am: Remove rule for generating NEWS file

When "git pull" (or any other operation) brings in a new version of the
gnulib git submodule, you must rerun the autogen.sh script.  With this
change, "make" now fails and tells you to run ./autogen.sh, when needed.
* autogen.sh: Maintain a new file, .git-module-status, containing
the current submodule status.  If it doesn't exist or its content
is different from what "git submodule status" prints, then run
./bootstrap
* .gitignore: Add .git-module-status
* cfg.mk: Diagnose out of date submodule and fail.
* README-hacking: Update not to mention bootstrap.
* Makefile.am (MAINTAINERCLEANFILES): Add .git-module-status,
so that "make maintainerclean" will remove it.

* Makefile.am: List -I m4 first, in ACLOCAL_AMFLAGS
* .gitignore: ignore gnulib/, ChangeLog, *rej, *orig, *#*# (emacs
  temporary files)

* bootstrap (modules): Add gnumakefile and maintainer-makefile.
* GNUmakefile: Remove file, now provided by gnulib.
* Makefile.maint: Remove.  Replaced by maint.mk from gnulib.
.gitignore: Add GNUmakefile and maint.mk.
* cfg.mk (prev_version_file): Disable this feature.
Setting this to /dev/null avoids an otherwise harmless diagnostic.

This makes it so we record (via a git submodule)
a snapshot of whatever version of gnulib we're using,
and none of gnulib sources are in the libvirt repository.
The result is that we have as much reproducibility as when
we version-controlled imported copies of the gnulib sources,
but without the hassle of the manual process we used when
syncing with upstream.

Note that when you clone libvirt, you get only the libvirt
repository, but when you first run ./bootstrap, it clones
gnulib (at the SHA1 recorded via the submodule), creating
the .gnulib/ hierarchy.  Then, the bootstrap script runs
gnulib-tool to populate gnulib/ with the files that make
up the selected modules.

Put the following in your ~/.gitconfig file.
[alias]
  syncsub = submodule foreach git pull origin master

The update procedure is simple:
  git syncsub
  ...build & test...
  git commit -m 'gnulib: sync submodule to latest' .gnulib

* .gitmodules: New file.
* .gnulib: Initialize.
* bootstrap: Set up to use the new submodule.
Stop using --no-vc-files.
Don't remove .gitignore files.
Don't use or create .cvsignore.
Diagnose an invalid --gnulib-srcdir=DIR argument.
* build-aux/vc-list-files: Delete file, now pulled from gnulib.
* build-aux/useless-if-before-free: Likewise.
* po/POTFILES.in: Remove gnulib/lib/gai_strerror.c, since
it no longer contains translatable strings.
* gnulib/*: Remove gnulib/ hierarchy.

* configure.in libvirt.spec.in NEWS docs/* po/*: release of
  libvirt-0.6.5
* .gitignore: adding cscope files
Daniel