Re: [PATCH] sdk: Avoid parsing errors when DISTRO_ARCH is not yet set

[Uladzimir Bely <ubely@ilbers.de>]

In mail from Monday, 27 February 2023 13:41:37 +03 user Jan Kiszka wrote:
> On 22.02.23 17:28, Jan Kiszka wrote:
> > From: Jan Kiszka <jan.kiszka@siemens.com>
> > 
> > In multiconfig setups, it can happen during some parsing steps that a
> > machine and, thus, a concrete DISTRO_ARCH is not yet set, leading to
> > 
> > The stack trace of python calls that resulted in this exception/failure
> > was: File: '<code>', lineno: 4, function: <module>
> > 
> >      0001:__anon_158__build____work_isar_meta_classes_base_bbclass(d)
> >      0002:__anon_162__build____work_isar_meta_classes_sstate_bbclass(d)
> >      0003:__anon_56__build____work_isar_meta_classes_image_bbclass(d)
> >  
> >  *** 0004:__anon_56__build____work_isar_meta_classes_sdk_bbclass(d)
> >  
> >      0005:__anon_21__build____work_isar_meta_classes_buildchroot_bbclass(d
> >      )
> >      0006:__anon_84__build____work_isar_meta_classes_imagetypes_wic_bbclas
> >      s(d)
> >      0007:__anon_24__build____work_isar_meta_classes_imagetypes_container_
> >      bbclass(d)
> >      0008:__anon_305__build____work_isar_meta_classes_image_bbclass(d)
> > 
> > File: '/build/../work/isar/meta/classes/sdk.bbclass', lineno: 52,
> > function: __anon_56__build____work_isar_meta_classes_sdk_bbclass> 
> >      0048:    distro_arch = d.getVar('DISTRO_ARCH')
> >      0049:    if mode == "0" or d.getVar('HOST_ARCH') ==  distro_arch:
> >      0050:        toolchain = "build-essential"
> >  
> >      0051:    else:
> >  *** 0052:        toolchain = "crossbuild-essential-" + distro_arch
> >  
> >      0053:    if d.getVar('ISAR_ENABLE_COMPAT_ARCH', True) == "1":
> >      0054:        toolchain += " crossbuild-essential-" +
> >      d.getVar('COMPAT_DISTRO_ARCH') 0055:    d.setVar('TOOLCHAIN',
> >      toolchain)
> >      0056:}
> > 
> > Exception: TypeError: can only concatenate str (not "NoneType") to str
> > 
> > Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
> > ---
> > 
> > Found in mtda: https://github.com/siemens/mtda/pull/283
> > 
> >  meta/classes/sdk.bbclass | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/meta/classes/sdk.bbclass b/meta/classes/sdk.bbclass
> > index 0a98ea04..01eb3a67 100644
> > --- a/meta/classes/sdk.bbclass
> > +++ b/meta/classes/sdk.bbclass
> > @@ -46,7 +46,7 @@ SDK_PREINSTALL += " \
> > 
> >  python __anonymous() {
> >  
> >      mode = d.getVar('ISAR_CROSS_COMPILE', True)
> >      distro_arch = d.getVar('DISTRO_ARCH')
> > 
> > -    if mode == "0" or d.getVar('HOST_ARCH') ==  distro_arch:
> > 
> > +    if mode == "0" or d.getVar('HOST_ARCH') == distro_arch or distro_arch 
== None:
> >          toolchain = "build-essential"
> >      
> >      else:
> >          toolchain = "crossbuild-essential-" + distro_arch
> 
> Hmm, I thought you have fast path for small fixes now?
> 
> Jan

Hello.

Yes, such an obvious thing might have been merged faster than usually. But it 
happened to conflict with other patches we were testing at the same time, at 
least with "[v2,0/4] Fix ccache issues" that seems to absorb your patch. 
That's why merging was delayed, considering both "fast" and "full" CI still 
pass OK.

I guess, you have the issue in some downstream that doesn't set distro_arch.

Anyway, ccache patchset will require new version by some other reasons, so we 
could merge your patch and consider it in ccache patchset v3.