[OE-core] wpa-supplicant strangeness after pull

[Khem Raj]

On Wed, Feb 29, 2012 at 2:05 PM, Steve Sakoman <sakoman at gmail.com> wrote:
> I'm continuing to see failing builds or defective packages on package
> rebuilds triggered by the recent change to sstate hash.
>
> The wpa-supplicant package is a good example of the latter -- the
> rebuild completes successfully but the resulting package is defective.
>  There were no recent changes to the recipe, so this was clearly a
> "just to be safe" rebuild based on sstate hash (correct me if I am
> wrong!)
>
> In particular /usr/share/dbus-1/system-services/fi.w1.wpa_supplicant1.service
> now contains:
>
> Exec=/usr/usr/sbin/wpa_supplicant -u
>
> rather than:
>
> Exec=/usr/sbin/wpa_supplicant -u
>
> Rebuilding the package after a -c cleansstate restores the expected result.
>
> Is it "just me" or are others seeing this kind of breakage?
>

you are just ahead of others :)
and this is yet another case where basichash exposes an underlying problem

see the sed command in below snippet

install -m 644 ${S}/dbus/dbus-wpa_supplicant.conf
${D}/${sysconfdir}/dbus-1/system.d
install -d ${D}/${datadir}/dbus-1/system-services
sed -i -e s:${base_sbindir}:${sbindir}:g ${S}/dbus/*.service
install -m 644 ${S}/dbus/*.service ${D}/${datadir}/dbus-1/system-services

that command can only be run correctly once unless you delete ${S}
since its sedding something in ${S}
and we know base_sbindir (/sbin) and sbindir(/usr/sbin) are close
enough expressionwise


> I'm beginning to suspect I ought to just do a clean build after every
> pull in order to remain productive!

That will hide these kind of issues you bring forward :)

>
> Steve
>
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core at lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core