Ok, I just hit the ninja error (same as before)
===[hook_eval:130 (last $?=0)> sort
==[hook_eval:131 (last $?=0)> '[' 5 ']'
==[hook_eval:131 (last $?=0)> eval './configure.py --bootstrap'
===[hook_eval:131 (last $?=0)> ./configure.py --bootstrap
fl_wrapper.so: write outside basedir (/opt/t2-trunk):
/usr/lib/python3.7/encodings/__pycache__/__init__.cpython-37.pyc.139965801578800
Due to previous errors, no 0-ninja.log file!
(Try enabling xtrace in the config to track an error inside the build
system.)
--- BUILD ERROR ---
Creating file list and doing final adaptions ...
Searching for orphaned files ...
Found 5 files for this package.
Found 0 orphaned files for this package.
Clear (old) md5sums ...
Creating md5sum files ... done.
Creating package description ...
Making post-install adaptions.
== 02/04/20 04:55:14 =[0]=> Aborted building package ninja.
Applying the attached patch, and re-building python allows the ninja
build to succeed.
Jan
On 2/3/20 11:50 PM, Jan Rovins wrote:
Ok, Rene, I have added symlinks for arch & uname into bin (I used the
ones from /usr/embutils)
And now the Build-Target is off and running, got past glibc, so I
will just let it run.
for scsijon: I am at version 49198 and am using
t2-minimal-glibc-gcc-x86-64-r49128.iso as the reference for the chroot
binaries.
Jan
On 2/3/20 4:21 AM, René Rebe wrote:
Hi,
as per my long YT development & test livestream this is coreutils not
being installed correctly causing uname and arch not being in /bin
You can probably move them over and re build coreutils to fully fix
this.
Yes, it is also rather annoying and time consuming for myself that
each time we spin an ISO there are upstream regressions. Warning:
t2/trunk glibc update breaks glib and thus many other depends if you
build a large package selection. I will try to hunt down what is
wrong between glib and glibc another day, Google and mailing lists
did not yet had an answer for that @GLIBC PRIVATE symbol. For now I
would locally revert the glibc update…
René
On 3. Feb 2020, at 06:04, Jan Rovins <[email protected]> wrote:
Hi Rene,
Can't seem to get at anything useful to trace this, it's failing
very early in the T2 wrappers, before the glibc debug dir is created.
ERROR-LOG attached.
Jan
On 2/2/20 11:08 PM, Jan Rovins wrote:
Hi Rene & List.
Just got my new fast build server set up, and saw that the new iso
was out, so I figured that I'd try a fresh T2 chroot build with the
latest from svn.
Attached are the scripts that I'm using to set up the chroots, if
you wan to review them & share them on the site.
I will send outy notes on each regression as I hit them.
The build is a minimal generic, without anything special.
The first thing that I ran into was with glibc stage 0:
===[eval_config_command:284 (last $?=0)> ../configure
--prefix=/opt/t2-trunk/build/default-9.0-svn-generic-x86-64-linux/usr
--enable-hacker-mode
--with-headers=/opt/t2-trunk/build/default-9.0-svn-generic-x86-64-linux/usr/include
--with-libpam --with-pam --enable-libpam --enable-pam --with-tls
--with-__thread --enable-obsolete-rpc --disable-profile
--disable-multi-arch --enable-kernel=2.6.32 --with-fp
--with-gnu-binutils --build=-nocross-linux-gnu
--host=x86_64-t2-linux-gnu --cache-file=./config.cache --without-gd
configure: loading cache ./config.cache
checking build system type... config.sub: invalid option
-nocross-linux-gnu
Try `config.sub --help' for more information.
configure: error: /bin/sh ../scripts/config.sub -nocross-linux-gnu
failed
Due to previous errors, no 0-glibc.log file!
(Try enabling xtrace in the config to track an error inside the
build system.)
--- BUILD ERROR ---
Will start on debugging this next, and send an update shortly.
Jan
<ERROR-LOG.txt>-----------------------------------------------------------
If you wish to unsubscribe from this mailing, send mail to
[email protected] with a subject of: unsubscribe t2
Index: package/python/python/python.conf
===================================================================
--- package/python/python/python.conf (revision 49140)
+++ package/python/python/python.conf (working copy)
@@ -32,6 +32,8 @@
fi
+# sed can't find Modules/Setup.dist, amd the build fails
+# so we will skip calling this
python_modules() {
if pkginstalled gdbm; then
sed -i 's:#gdbm gdbmmodule.c -I/usr/local/include -L/usr/local/lib:gdbm gdbmmodule.c -I/usr/include -L/usr/lib:' Modules/Setup.dist
@@ -52,7 +54,8 @@
sed -i 's:#zlib:zlib:' Modules/Setup.dist
fi
}
-atstage toolchain || hook_add preconf 5 "python_modules"
+# FAILED. skip.
+#atstage toolchain || hook_add preconf 5 "python_modules"
# We can savely ignore this dependencies.
var_append flistrfilter "|" ".*: /usr/lib.*/python.*/site-packages/.*"
@@ -59,7 +62,7 @@
# hardcoded paths
fix_python_libdir() {
- sed -i "s,/lib\([[:blank:]]\),/${libdir##*/}\1,g" Modules/Setup.dist
+ #sed -i "s,/lib\([[:blank:]]\),/${libdir##*/}\1,g" Modules/Setup.dist
ln -svnf ../lib/python${ver:0:3} $root/$libdir/python${ver:0:3}
}
[[ $libdir = *lib ]] || hook_add preconf 9 "fix_python_libdir"
@@ -68,6 +71,9 @@
if atstage toolchain; then
export LDFLAGS="-Wl,-rpath,${root}${libdir}"
#hook_add postmake 5 'cp Parser/pgen $root$bindir/python-pgen'
+
+ # env/python fails without this, and other builds that call this will fail.
+ hook_add postmake 5 'ln -s $root$bindir/python3 $root$bindir/python'
fi
# cross-compiling
-----------------------------------------------------------
If you wish to unsubscribe from this mailing, send mail to
[email protected] with a subject of: unsubscribe t2
