[libvirt] [PATCHv3] build: fix virBufferVasprintf on mingw

[Eric Blake]

Gnulib documents that mingw vsnprintf is broken (it returns -1
on out-of-space, instead of the count of what would have been
printed); but while we were using the snprintf wrapper, we had
not yet been using the vsnprintf wrapper.

Meanwhile, mingw (but not mingw64) has a replacement snprintf
that fixes return values, but still lacks %1$s support; so in
that case, gnulib didn't replace snprintf, but libintl then
went ahead and installed a version that supported %1$s but not
return values.  Gnulib has since been fixed to guarantee that
the snprintf module will always guarantee the constraints needed
by libintl.

Also, we want to guarantee that strdup sets errno on failure.

* .gnulib: Update to latest, for vsnprintf fix.
* bootstrap.conf (gnulib_modules): Add vsnprintf, strdup-posix.
Reported by Matthias Bolte.
---

v3: more gnulib updates, also fix strdup
v2: see https://www.redhat.com/archives/libvir-list/2011-July/msg00042.html

* .gnulib 7269b35...8db4963 (23):
  > sys_select: define sigset_t more portably
  > * m4/sys_select_h.m4 (gl_HEADER_SYS_SELECT): Poison pselect.
  > pselect: Use pthread_sigmask, not sigprocmask.
  > modules/pselect: Add Bruno and Jim to maintainers.
  > * modules/pselect: Add select, LIBSOCKET.
  > * lib/pselect.c (pselect): Use plain name, without "rpl_".
  > pselect: document better
  > pthread_sigmask: new module
  > test-pselect: new module
  > * tests/test-sys_select.c (sigset_t): Test for it, too.
  > snprintf: guarantee %1$d, for libintl
  > maint: exempt stdio-read.c and stdio-write.c from the cppi check
  > maint: indent with spaces, not TABs, and add a rule to check this
  > maint.mk: correct omissions in prohibit_argmatch_without_use check
  > Comments about EINTR.
  > autoupdate
  > xnanosleep: Rewrite to use new dtotimespec module.
  > timespec-add, timespec-sub: new modules
  > dtotimespec: new module
  > * lib/timespec.h (timespectod): New inline function.
  > * lib/timespec.h (timespec_sign): New inline function.
  > pselect: new module
  > sys_select: don't depend on sys_socket

I'm also investigating how easy it would be to use gnulib's new
pthread_sigmask.

 .gnulib        |    2 +-
 bootstrap.conf |    2 ++
 2 files changed, 3 insertions(+), 1 deletions(-)

diff --git a/.gnulib b/.gnulib
index 7269b35..8db4963 160000
--- a/.gnulib
+++ b/.gnulib
@@ -1 +1 @@
-Subproject commit 7269b35c8d9be1a6f97906b9e29b8c422b92fc31
+Subproject commit 8db49630deab35ffbce5c113aed093359a75091d
diff --git a/bootstrap.conf b/bootstrap.conf
index 581d60b..d3ff150 100644
--- a/bootstrap.conf
+++ b/bootstrap.conf
@@ -80,6 +80,7 @@ socket
 stdarg
 stpcpy
 strchrnul
+strdup-posix
 strndup
 strerror
 strerror_r-posix
@@ -97,6 +98,7 @@ usleep
 vasprintf
 verify
 vc-list-files
+vsnprintf
 waitpid
 warnings
 '
-- 
1.7.4.4