Control: reassign -1 src:efibootguard 0.21-1
Control: tags -1 ftbfs patch upstream
On Mon, Jan 05, 2026 at 07:51:45PM +0100, Chris Hofstaedtler wrote:
> On Mon, Jan 05, 2026 at 06:53:35PM +0100, Chris Hofstaedtler wrote:
> > the following packages differ between the last good and first bad timestamp:
> > binutils 2.45-8 -> 2.45.50.20251125-1
>
> bisecting binutils master points to this commit:
>
> 5e83077d552ed6f81dbc092eb3ccf827a43de42c is the first bad commit
> commit 5e83077d552ed6f81dbc092eb3ccf827a43de42c
[..]
> The mentioned NEWS entry says:
>
> * Internal changes to plugin support, and stricter target checking may result
> in some errors being exposed in user options passed to the various
> binutils.
> For example objcopy --target=TARGET now will only work if the input file is
> for TARGET whereas prior versions of objcopy accepted other target input
> files and produced a TARGET output. If you do in fact want the old
> behaviour the correct usage is objcopy --output-target=TARGET.
>
> So maybe what is seen in efibootguard is indeed intentional.
The following patch fixes the efibootguard FTBFS. Reassigning back
to efibootguard.
diff --git i/Makefile.am w/Makefile.am
index fcd66ed..06542c6 100644
--- i/Makefile.am
+++ w/Makefile.am
@@ -303,13 +303,13 @@ efi_ldflags = \
$(GNUEFI_LIB_DIR)/crt0-efi-$(ARCH).o
if ARCH_IS_X86
-objcopy_format = --target=efi-app-$(ARCH)
+objcopy_format = --output-target=efi-app-$(ARCH)
else
if OBJCOPY_USE_BINARY_COPY
objcopy_format = -O binary
efi_ldflags += --defsym=EFI_SUBSYSTEM=0xa
else
-objcopy_format = --target=efi-app-$(ARCH)
+objcopy_format = --output-target=efi-app-$(ARCH)
endif
endif