| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
| |
* src/estream.c (_gpgrt_fflush): Skip the read-only stream.
--
GnuPG-bug-id: 8518
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
| |
* src/estream.c (trylock_stream): Return 0 on success, -1 on error.
--
GnuPG-bug-id: 8519
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
| |
* src/gpg-error.h.in: Redefine GPGRT_HAVE_PRAGMA_GCC_PUSH
--
The old definition only covered push_option which is available since
gcc 4.4 (released 2009). The new definition also covers the
diagnostic push pragma which was introduced with 4.6 (released 2011).
|
| |
|
|
|
|
|
|
| |
* m4/Makefile.am (EXTRA_DIST): Add gnulib-common.m4.
--
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
| |
* m4/threadlib.m4: Detect gl_pthread_in_glibc on Hurd with GNU C
library >= 2.43.
--
GnuPG-bug-id: 8430
Co-authored-by: Samuel Thibault <[email protected]>
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
| |
* m4/threadlib.m4: Update.
* m4/gnulib-common.m4: New for gl_CHECK_FUNCS_ANDROID.
--
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
|
| |
* tests/t-printf.c (check_large_float): Use _ARCH_PPC to identify
Power architecture, both of 32-bit and 64-bit.
--
Fixes-commit: 9cc750a816554e46d3e1496533f927b17ef3fcef
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
|
| |
* tests/t-printf.c (check_large_float): Skip the test when
it's IBM double-double.
--
Reported-by: C. Neidahl
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
| |
--
|
| |
|
|
| |
Signed-off-by: Daniel Cerqueira <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
| |
* configure.ac (AC_TYPE_LONG_DOUBLE_WIDER): New.
* tests/t-printf.c (check_large_float): Fix a typo.
[!HAVE_LONG_DOUBLE_WIDER]: Skip the LDBL_MAX test.
--
GnuPG-bug-id: 8309
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* src/visibility.c (gpgrt_w32_set_errno): New.
* src/w32-add.h: Add function.
* src/gpg-error.def.in: Ditto.
* src/estream.c (_gpgrt_w32_set_errno): Return the windows error code.
* src/gpg-error.c (main): Add CMD_MAPW32ERR with name "mapw32err" and
implement it.
--
We use various mappings from Windows Error codes to ERRNO and thus to
gpg-error. it is better to eventuially have one central place for
such mappings.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
* configure.ac: For invocation of gen-lock-obj.sh, use NM instead of
OBJDUMP, use no AWK.
* src/gen-lock-obj.sh: Use NM with option -P and -t d for portable
output in decimal. No use of AWK, simply use cut.
--
Signed-off-by: Alexis Lothoré <[email protected]>
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
| |
--
GnuPG-bug-id: 8242
|
| |
|
|
| |
--
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
| |
* src/estream.c (_gpgrt_fflush): Do not use the in_atexit flag.
--
We need to research how we can better handle atexit handlers and
stream locking; in particilar for stdout and stderr. For now we keep
the code but disable it.
Fixes-commit: 50f0c56bdd36f7f709a62503942ac55399dafb9c
|
| |
|
|
|
|
|
|
|
|
|
| |
* src/estream.c (_gpgrt_fflush): Add arg in_atexit and skip the flush in
this mode if the lock can not be taken. Change all callers.
--
The problem should up with gpgsm encryption on the standard mode and a
no public key error. In this case the process got stuck in the atexit
handler after a FAILURE gpgsm-exit status line to stderr. It is
better not to have a hanging process than having flushed everything.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* configure.ac: Add test for snprintf.
* src/gpgrt-int.h: Do not re-map snprintf if KEEP_SYSTEM_SNPRINTF is
defined.
* src/estream-printf.c (KEEP_SYSTEM_SNPRINTF): Define.
(pr_float): Increase static buffer size. Use snprintf for long
double.
* tests/t-printf.c: Include float.h.
(check_large_float): New.
--
GnuPG-bug-id: 8240
Reported-by: Dirk Mueller <[email protected]> (AI)
|
| |
|
|
|
|
|
|
| |
* src/stringutils.c (_gpgrt_vfnameconcat): Use the correct value for
the trailing slash check.
--
Reported-by: Dirk Mueller <[email protected]> (AI)
|
| |
|
|
|
|
| |
--
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
* configure.ac: Add wasm*-*-emscripten target to generate
lock-obj-pub.native.h.
--
Reported-by: Clément Péron <[email protected]>
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
| |
--
|
| | |
|
| |
|
|
| |
--
|
| |
|
|
| |
--
|
| |
|
|
|
|
|
|
|
|
| |
* src/gen-lock-obj.sh: Don't specify section name.
--
It may be .bss.mtx depending to a toolchain.
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| | |
|
| |
|
|
|
|
| |
* src/spawn-w32.c (_gpgrt_process_spawn): Move the hd[32] array
declaration from inside the inner block to the function scope, so that
the array remains live when CreateProcessW is called.
|
| |
|
|
|
|
|
|
|
| |
* configure.ac: Don't use PKG_CHECK_MODULES.
* src/Makefile.am (mkheader): Remove EXTRA_LIBS_FOR_BUILD.
--
Signed-off-by: Sachin T <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
* m4/libtool.m4 (AC_LIB_PROG_LD): Define here.
--
This is local solution. It should be handled by the upstream, Gnulib
and libtool.
GnuPG-bug-id: 7677
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
|
| |
* m4/Makefile.am: Update.
--
Mention the upstreams.
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
| |
* m4/intlmacosx.m4: New.
* m4/Makefile.am: Update.
--
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
| |
--
This will help to implement
GnuPG-bug-id: 6644
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* src/estream.c (IS_INVALID_SOCKET): New.
(func_sock_read, func_sock_write, func_sock_ioctl, func_sock_destroy):
Use IS_INVALID_SOCKET.
--
It is possible for some architectures that the value of extended -1
and ~0 is different in C. Better to avoid the issue in advance.
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
| |
--
GnuPG-bug-id: 8066
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| | |
|
| |
|
|
| |
--
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
* src/spawn-w32.c (_gpgrt_process_spawn): Don't close
the handle when GPGRT_PROCESS_*_KEEP is specified.
--
GnuPG-bug-id: 8052
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
* src/spawn-w32.c (prepare_env_block): Copy environment
variables correctly by memmove.
--
With "export GNUPGHOME=something", t-spawn.exe failed before the fix.
GnuPG-bug-id: 8052
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
| |
--
For some projects it is required to use the --force options. The new
variable allows to force force using autogen.rc.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* src/spawn-posix.c (spawn_detached): Don't call _gpgrt_post_syscall
on the child process.
(_gpgrt_process_spawn): Likewise.
--
On NetBSD, a POSIX semaphore is somehow shared among processes, even
if sem_init is called with PSHARED=0.
GnuPG-bug-id: 8065
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
| |
--
|
| |
|
|
|
|
|
|
|
|
|
| |
* build-aux/texinfo.tex: Fix typos.
--
Apply the part of changes in libgcrypt's commit of:
9b651fb632f3697e70685c9ee340ab0cb2274bdf
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
| |
* m4/ax_cc_for_build.m4.: Fix for no arg.
--
Incorporate the change of libgcrypt commit:
e3b441214f93d8f61875b8223480e57afa2a3f10
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
| |
* build-aux/libtool.patch: Remove.
--
Fixes-commit: 8dd210e9e15905f8c6d155c64b09378d65382335
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
| |
* tests/t-stringutils.c: [HAVE_W32_SYSTEM] (mygethome): Use
fnameconcat to get HOME, instead of the environment variable.
--
Signed-off-by: NIIBE Yutaka <[email protected]>
|
| |
|
|
|
|
|
|
|
| |
* src/spawn-w32.c (_gpgrt_process_terminat): Care about the case
it's INVALID_HANDLE_VALUE.
--
Signed-off-by: NIIBE Yutaka <[email protected]>
|