PR 25562: New binutils testsuite failures

[Jozef Lawrynowicz]

On Mon, 30 Mar 2020 10:48:32 +0100
Nick Clifton <nickc@redhat.com> wrote:

> Hi Jozef,
> 
>   I am seeing some new binutils testsuite failures for various different
>   targets:
> 
> Checking Binutils in: arm-wince-pe ... BIN REGRESSION: objcopy executable (pr25662)  
> Checking Binutils in: i686-pc-cygwin ... BIN REGRESSION: objcopy executable (pr25662)  
> Checking Binutils in: i686-pc-mingw32 ... BIN REGRESSION: objcopy executable (pr25662)  
> Checking Binutils in: mcore-pe ... BIN REGRESSION: objcopy executable (pr25662)  
> Checking Binutils in: mips-elf ... BIN REGRESSION: objcopy executable (pr25662)  
> Checking Binutils in: mips-sgi-irix6 ... BIN REGRESSION: objcopy executable (pr25662)  
> Checking Binutils in: mmix-mmixware ... BIN REGRESSION: objcopy executable (pr25662)  
> Checking Binutils in: tx39-elf ... BIN REGRESSION: objcopy executable (pr25662)  
> Checking Binutils in: x86_64-pc-cygwin ... BIN REGRESSION: objcopy executable (pr25662)  
> Checking Binutils in: x86_64-pc-mingw64 ... BIN REGRESSION: objcopy executable (pr25662)  
> 
>   Would you mind taking a look please ?

Hi Nick,

The test is working as expected, for those targets an objcopy of an executable
is not creating an exact copy of the original file.

Alan Modra has already done a triage of the failures (see this thread
https://sourceware.org/pipermail/binutils/2020-March/110422.html), and has setup
XFAILs for some targets. For the remaining failures I think the maintainers of
those targets need to investigate to see if XFAILs are necessary or if there is
a real bug in the toolchain.

For the PE targets the datestamp in the executable file is not being copied by
objcopy.

For the MIPS targets, objcopy has added the names of output sections to the
string table, where in the linked executable they were not in the string table.

I can file bugzillas for these if you would like.

Thanks,
Jozef
> 
> Cheers
>   Nick
>   
>