ildasm has problems with non-ASCII characters

[nnn1590]

Description

ildasm gives incorrect results if the input or output filename contains non-ASCII characters.

Reproduction Steps

cd "$(mktemp -d)"
dotnet new console --name bug --output .
dotnet add package runtime.linux-x64.Microsoft.NETCore.ILDAsm --version 6.0.0
dotnet build  # -> bin/Debug/net6.0/bug.dll

~/.nuget/packages/runtime.linux-x64.microsoft.netcore.ildasm/6.0.0/runtimes/linux-x64/native/ildasm -output=123１２３.il bin/Debug/net6.0/bug.dll
~/.nuget/packages/runtime.linux-x64.microsoft.netcore.ildasm/6.0.0/runtimes/linux-x64/native/ildasm bin/Debug/net6.0/bug.dll -output=ASCII_and_ＮＯＮ−ＡＳＣＩＩ＿ＣＨＡＲＡＣＴＥＲＳ.il
~/.nuget/packages/runtime.linux-x64.microsoft.netcore.ildasm/6.0.0/runtimes/linux-x64/native/ildasm -output=ASCII_and_ＮＯＮ−ＡＳＣＩＩ＿ＣＨＡＲＡＣＴＥＲＳ.il bin/Debug/net6.0/bug.dll

Expected behavior

File names are processed correctly.

Actual behavior

$ ~/.nuget/packages/runtime.linux-x64.microsoft.netcore.ildasm/6.0.0/runtimes/linux-x64/native/ildasm -output=123１２３.il bin/Debug/net6.0/bug.dll
error : File 'bin/Debug/net6.0/bu' not found or not a PE file
$ ~/.nuget/packages/runtime.linux-x64.microsoft.netcore.ildasm/6.0.0/runtimes/linux-x64/native/ildasm bin/Debug/net6.0/bug.dll -output=ASCII_and_ＮＯＮ−ＡＳＣＩＩ＿ＣＨＡＲＡＣＴＥＲＳ.il
("ASCII_and_ＮＯＮ−ＡＳＣＩ" generated instead of "ASCII_and_ＮＯＮ−ＡＳＣＩＩ＿ＣＨＡＲＡＣＴＥＲＳ.il")
$ ~/.nuget/packages/runtime.linux-x64.microsoft.netcore.ildasm/6.0.0/runtimes/linux-x64/native/ildasm -output=ASCII_and_ＮＯＮ−ＡＳＣＩＩ＿ＣＨＡＲＡＣＴＥＲＳ.il bin/Debug/net6.0/bug.dll
(The usage shown)

Regression?

No response

Known Workarounds

Don't use non-ASCII characters in the path.

Configuration

• Which version of .NET is the code running on?
  .NET 6.0.101 (dotnet-sdk-6.0/focal,now 6.0.101-1 amd64 installed)
• What OS and version, and what distro if applicable?
  Ubuntu 20.04 LTS amd64
• What is the architecture (x64, x86, ARM, ARM64)?
  x64
• Do you know whether it is specific to that configuration?
  I don't know.

Other information

I'm sorry if my english is wrong.

Tagging subscribers to this area: @JulieLeeMSFT
See info in area-owners.md if you want to be subscribed.

Issue Details

---

Description

ildasm gives incorrect results if the input or output filename contains non-ASCII characters.

Reproduction Steps

cd "$(mktemp -d)"
dotnet new console --name bug --output .
dotnet add package runtime.linux-x64.Microsoft.NETCore.ILDAsm --version 6.0.0
dotnet build  # -> bin/Debug/net6.0/bug.dll

~/.nuget/packages/runtime.linux-x64.microsoft.netcore.ildasm/6.0.0/runtimes/linux-x64/native/ildasm -output=123１２３.il bin/Debug/net6.0/bug.dll
~/.nuget/packages/runtime.linux-x64.microsoft.netcore.ildasm/6.0.0/runtimes/linux-x64/native/ildasm bin/Debug/net6.0/bug.dll -output=ASCII_and_ＮＯＮ−ＡＳＣＩＩ＿ＣＨＡＲＡＣＴＥＲＳ.il
~/.nuget/packages/runtime.linux-x64.microsoft.netcore.ildasm/6.0.0/runtimes/linux-x64/native/ildasm -output=ASCII_and_ＮＯＮ−ＡＳＣＩＩ＿ＣＨＡＲＡＣＴＥＲＳ.il bin/Debug/net6.0/bug.dll

Expected behavior

File names are processed correctly.

Actual behavior

$ ~/.nuget/packages/runtime.linux-x64.microsoft.netcore.ildasm/6.0.0/runtimes/linux-x64/native/ildasm -output=123１２３.il bin/Debug/net6.0/bug.dll
error : File 'bin/Debug/net6.0/bu' not found or not a PE file
$ ~/.nuget/packages/runtime.linux-x64.microsoft.netcore.ildasm/6.0.0/runtimes/linux-x64/native/ildasm bin/Debug/net6.0/bug.dll -output=ASCII_and_ＮＯＮ−ＡＳＣＩＩ＿ＣＨＡＲＡＣＴＥＲＳ.il
("ASCII_and_ＮＯＮ−ＡＳＣＩ" generated instead of "ASCII_and_ＮＯＮ−ＡＳＣＩＩ＿ＣＨＡＲＡＣＴＥＲＳ.il")
$ ~/.nuget/packages/runtime.linux-x64.microsoft.netcore.ildasm/6.0.0/runtimes/linux-x64/native/ildasm -output=ASCII_and_ＮＯＮ−ＡＳＣＩＩ＿ＣＨＡＲＡＣＴＥＲＳ.il bin/Debug/net6.0/bug.dll
(The usage shown)

Regression?

No response

Known Workarounds

Don't use non-ASCII characters in the path.

Configuration

• Which version of .NET is the code running on?
  .NET 6.0.101 (dotnet-sdk-6.0/focal,now 6.0.101-1 amd64 installed)
• What OS and version, and what distro if applicable?
  Ubuntu 20.04 LTS amd64
• What is the architecture (x64, x86, ARM, ARM64)?
  x64
• Do you know whether it is specific to that configuration?
  I don't know.

Other information

I'm sorry if my english is wrong.

Author:	nnn1590
Assignees:	-
Labels:	area-ILTools-coreclr, untriaged
Milestone:	-

[am11]

This looks like we have an issue with PAL's MultiByteToWideChar.

Minimal repro (excerpt from src/coreclr/ilasm/main.cpp#L855):

int main(int argc, char* str[])
{
    g_pszExeFile = str[0];
    if (0 != PAL_Initialize(argc, str))
    {
        fprintf(stderr,"Error: Fail to PAL_Initialize\n");
        exit(1);
    }

    const char* test = "１２３４５６"; // U+FF11U+FF12U+FF13U+FF14U+FF15U+FF16
    printf("before conversion (LPSTR): %s\n", test);

    int length = MultiByteToWideChar(CP_ACP, 0, test, -1, NULL, 0);
    LPWSTR result = new (nothrow) WCHAR[length];
    length = MultiByteToWideChar(CP_ACP, 0, test, -1, result, length);

    wprintf(W("after conversion (LPWSTR): %s\n"), result);
}

Output:

before conversion (LPSTR): １２３４５６
after conversion (LPWSTR): １２

cc @janvorli

[am11]

It was actually a problem in PAL initialization, fixed by #64146.

wprintf not printing whole string with %s specifier and WCHAR arg, seems to be a separate issue.

[nnn1590]

Thank you!