Re: [PATCH] Riscv ld-elf/stab failure, and fake label cleanup.

[Kamil Rytarowski <n54 at gmx dot com>]

On 23.11.2017 12:18, Nick Clifton wrote:
> Hi Jim,
> 
>>> 	* testsuite/gas/all/err-fakelabel.s: New.
> 
> Sorry for the late notice, but I have just found new gas testsuite failures
> for a couple of architectures triggered by this test:
> 
>   Checking Binutils in: vax-netbsdelf ... GAS REGRESSION: gas/all/err-fakelabel.s
>   Checking Binutils in: ns32k-netbsd ...  GAS REGRESSION: gas/all/err-fakelabel.s   
>   Checking Binutils in: mmix-mmixware ... GAS REGRESSION: gas/all/err-fakelabel.s
> 
> The failures all happen because the assembler backends are producing fatal 
> error messages:
> 
> [vax]    gas/testsuite/gas/all//err-fakelabel.s:3: Fatal error: Ignoring statement due to "Unknown operator"
> [ns32k]  gas/testsuite/gas/all//err-fakelabel.s:3: Fatal error: No such opcode
> [mmix]   gas/testsuite/gas/all//err-fakelabel.s:3: Fatal error: internal: unhandled label label
> 
> Frankly I do not see a great need to change the backends to cope with this
> situation, so please could you add a #not-target line to the test, along
> with a comment explaining why it is needed ?  A patch to do so is pre-approved.
> 
> Cheers
>   Nick
> 

For the record, vax-netbsdelf is actively used. ns32k-netbsd is active
too but not in the mainstream NetBSD project (people try to recreate
hardware).

Attachment: signature.asc
Description: OpenPGP digital signature