There is a `__wia_h__` macro check in wiadef.h. The new wia.h does not
have that macro. I am not sure if the check should be removed in
wiadef.h or be changed something else. Please feel free to change it
as you like. Test case
https://github.com/msys2/MINGW-packages/issues/17451
From 02bd0321d8388
--
Best regards,
LIU Hao
From 8959c30982e0802fc3d12c9e1688145ac5708cf0 Mon Sep 17 00:00:00 2001
From: LIU Hao
Date: Fri, 14 Feb 2025 10:57:36 +0800
Subject: [PATCH] crt/fesetexceptflag: Fix MXCSR operation
It should set the exception bits, rather than the exception mask bits.
Code on line 38-4
On 2025-02-12 Martin Storsjö wrote:
> (One could maybe even consider the struct to be a struct with an
> unspecified length of the d_name field? That'd make it clearer to
> callers that they'd better not try to allocate it themselves with a
> static size.)
This is the intention in POSIX. d_name ha
On Thursday 13 February 2025 14:16:06 Martin Storsjö wrote:
> On Fri, 7 Feb 2025, Pali Rohár wrote:
>
> > This will report runtime error if _CRT_INIT(DLL_PROCESS_DETACH) is called
> > by the current thread from some atexit_table hook.
>
> Btw, I forgot to ask; how would one practically do in orde
On Thursday 13 February 2025 14:10:44 Martin Storsjö wrote:
> On Fri, 7 Feb 2025, Pali Rohár wrote:
>
> > This will report runtime error if _CRT_INIT(DLL_PROCESS_DETACH) is called
> > by the current thread from some atexit_table hook.
>
> In which way will it report a runtime error? Do you mean t
On Thursday 13 February 2025 14:26:30 Martin Storsjö wrote:
> On Fri, 7 Feb 2025, Pali Rohár wrote:
>
> > This macro is used only on few places in .c files, not in .h files.
> > So remove its usage and dead code.
> > ---
> > mingw-w64-crt/crt/crtdll.c | 7 ---
> > mingw-w64-crt/crt/crtexe.c
On 13.02.2025 11:17, Jacek Caban wrote:
That’s unfortunate. If anyone has a better idea, please let me know.
Ideally, I’d like to restore the previous behavior for delay load
imports. Wine doesn’t use dlltool to create delay load import
libraries; it generates them itself. In theory, this gives
On Fri, 7 Feb 2025, Pali Rohár wrote:
Function with name DllEntryPoint is entry point of the DLL library,
an alternative or aliased name to DllMainCRTStartup. So calling entry point
(DllEntryPoint) from entry point (DllMainCRTStartup) is just wrong.
Default DLL entry point function varies betwe
On Fri, 7 Feb 2025, Pali Rohár wrote:
This macro is used only on few places in .c files, not in .h files.
So remove its usage and dead code.
---
mingw-w64-crt/crt/crtdll.c | 7 ---
mingw-w64-crt/crt/crtexe.c | 1 -
mingw-w64-crt/crt/dll_argv.c | 4
mingw-w64-crt/crt/dllargv.c | 4
Consider the following code:
test-fesetexceptflag.c:
#include
#include
#include
#include
#pragma STDC FENV_ACCESS ON
static void
handle_signal (int sig)
{
printf ("SIGFPE\n");
_Exit (1);
}
int main()
{
fexcept_t fexcept;
fegetexceptflag(&fexcept, FE_ALL_EXCE
On Fri, 7 Feb 2025, Pali Rohár wrote:
This will report runtime error if _CRT_INIT(DLL_PROCESS_DETACH) is called
by the current thread from some atexit_table hook.
Btw, I forgot to ask; how would one practically do in order to trigger
this case? I know it's a kinda pathological case that one i
On Fri, 7 Feb 2025, Pali Rohár wrote:
This will report runtime error if _CRT_INIT(DLL_PROCESS_DETACH) is called
by the current thread from some atexit_table hook.
In which way will it report a runtime error? Do you mean the call to
_amsg_exit(31), which gets called if (__native_startup_state
On 13.02.2025 12:01, LIU Hao wrote:
在 2025-02-13 18:17, Jacek Caban 写道:
This isn't entirely true for other linkers. While those linkers do
place .idata sections, including the IAT itself, in .rdata, they
don’t use .idata sections for delay imports. Instead, they implement
delay loading within
On Thu, 13 Feb 2025, LIU Hao wrote:
A writeable IAT is a conspicuous target for code injection.
If MSVC makes it read-only, LLD makes it read-only, and there have been few
issues with MSVC in practice (see the Old New Thing article), then there's no
reason for us to not do that; do you agree?
在 2025-02-13 18:17, Jacek Caban 写道:
This isn't entirely true for other linkers. While those linkers do place .idata sections, including
the IAT itself, in .rdata, they don’t use .idata sections for delay imports. Instead, they implement
delay loading within the linker itself rather than relying
Hi,
Thanks for looking into this. We probably need something like that in
mingw-w64, but unfortunately, it's more complicated for Wine.
On 11.02.2025 04:18, LIU Hao wrote:
Since Binutils 2.44, the import address table (abbr. IAT) is now read-only,
in alignment with Microsoft LINK and LLD. It
16 matches
Mail list logo
