Svante Signell, le ven. 05 janv. 2024 23:20:54 +0100, a ecrit: > On Fri, 2024-01-05 at 21:58 +0100, Samuel Thibault wrote: > > Svante Signell, le ven. 05 janv. 2024 21:14:19 +0100, a ecrit: > > > --- a/debian/rules 2023-12-16 18:35:11.000000000 +0100 > > > +++ b/debian/rules 2024-01-02 12:49:12.000000000 +0100 > > > @@ -4,7 +4,14 @@ > > > include /usr/share/dpkg/pkg-info.mk > > > > > > export DEB_BUILD_MAINT_OPTIONS = hardening=+all > > > -export > > > _PYTHON_SYSCONFIGDATA_NAME=_sysconfigdata__$(DEB_HOST_ARCH_OS)_$(DE > > > B_HOST_MULTIARCH) > > > + > > > +ifeq ($(DEB_HOST_ARCH_OS), hurd) > > > +#FIXME: Replace gnu0 with either gnu or hurd in python3.11! > > > +#/usr/lib/python3.11/__pycache__/_sysconfigdata__gnu0_i386- > > > gnu.cpython-311.pyc > > > + export > > > _PYTHON_SYSCONFIGDATA_NAME=_sysconfigdata__gnu0_$(DEB_HOST_MULTIARC > > > H) > > > +else > > > + export > > > _PYTHON_SYSCONFIGDATA_NAME=_sysconfigdata__$(DEB_HOST_ARCH_OS)_$(DE > > > B_HOST_MULTIARCH) > > > > Probably better just ask python itself: > > > > export _PYTHON_SYSCONFIGDATA_NAME:=_sysconfigdata__$(shell python3 -c > > 'import os; print(os.sys.platform)')_$(DEB_HOST_MULTIARCH) > > Why does the os.sys.platform report gnu0? It should be gnu or > preferrably hurd??
Well, here this question does not matter: it reports what it should report, i.e. what is used to name the sysconfig data file. If the string should be changed, that should be discussed with upstream python, not this package in particular (and this package should just adapt to whatever os.sys.platform comes to expose). > > > @@ -33,7 +40,7 @@ > > > else > > > dh_auto_configure -- \ > > > SH_PATH=/bin/sh \ > > > - MODINFO=/sbin/modinfo \ > > > + MODINFO= \ > > > --disable-uinput --disable-devinput > > > > We do want modinfo on the linux platform, please make this os- > > specific. > > Note the else: > ifeq ($(DEB_HOST_ARCH_OS), linux) > dh_auto_configure -- \ > SH_PATH=/bin/sh \ > MODINFO=/sbin/modinfo \ > --enable-uinput --enable-devinput > else > dh_auto_configure -- \ > SH_PATH=/bin/sh \ > MODINFO= \ > --disable-uinput --disable-devinput > endif Ah, ok, I didn't have the broader context in the patch. Samuel