| Newsgroups |
gmane.linux.lfs.automated |
| Message-ID |
<CABJoABbW=236gLtaj9zww1Q-VckFFNRMFJPLr4Z_ufsYjd_=zw@mail.gmail.com> |
On Mon, Mar 18, 2024 at 1:17 PM Pierre Labastie <
[email protected]> wrote:
> On Mon, 2024-03-18 at 00:11 +0000, md wrote:
> > Hi,
> >
> > I'm using jhalfs d3df258eddf27b6f3601a87b7ae3ee4c0c5af9ec Adapt
> > packInstall.sh.porg to xdg-utils location.
> >
> > Somewhere between the following two commits to
> > https://git.linuxfromscratch.org/lfs.git, there is a change
> > that appears to make "Building target 803-glibc" hang:
> > b570133fecd04d6a835c81816b95316225a7f77d Update changelog and whatsnew fo
> > r12.1 - This commit works
> > a56a4fcd9c2e07dbddd1ef6270d7fbe0139ed9aa procps: Run chown -R tester .
> before
> > running the test as tester -This commit appears to hang at "Building
> target
> > 803-glibc"
> >
> > This is how I'm starting the make target:
> > cd $LFS/jhalfs && make BREAKPOINT=825-shadow | ansi2txt | ack --passthru
> > "stopped at 825-shadow as requested"
> >
> > Other observations, which may help in debugging:
> > 1) Already in b570133fecd04d6a835c81816b95316225a7f77d, I see these
> errors,
> > but I don't have older logs to check when they first appeared
> > Building target 821-attr
> > make: *** [Makefile:1035: 825-shadow] Error 1
> > make: *** [Makefile:113: mk_CHROOT] Error 2
> > [0 min. 6 sec|
> > Target 821-attr OK
> >
> > 2) Already in b570133fecd04d6a835c81816b95316225a7f77d, I see the
> following
> > ownership.
> > I'm not sure this ownership is expected, but the full jhalfs build
> succeeds
> > (including booting the image).
> > tree -ugsD $LFS/jhalfs
> > > -- [root root 0 Mar 17 18:58] 401-creatingminlayout
> > > -- [root root 0 Mar 17 18:58] 402-addinguser
> > > -- [ lfs lfs 0 Mar 17 19:00]
> 403-settingenvironment
> > ...
> > > -- [ lfs lfs 0 Mar 17 20:26] 617-gcc-pass2
> > > -- [root root 0 Mar 17 20:30] 701-changingowner
> > ...
> > > -- [root root 0 Mar 17 21:24] 825-shadow
> >
> > I believe hanging of "Building target 803-glibc" is not due to running
> long
> > tests, since:
> > 3) I'm disabling these tests with:
> > cd $LFS/jhalfs/lfs-commands/chapter08 && sed -i "/make -k check/d"
> 8*-glibc &&
> > sed -i "/test_summary/d" 8*-glibc
>
> Have you looked a t the script after that? Maybe some part of the command
> is
> still there (like "su -tester"), which could explain the hang...
>
The lfs-git commit 887bf34f49879964ae9614cef6ecd0f94c222cae introduced
in lfs-commands/chapter08/803-glibc a grep for "Timed out", and this was
hanging
when only the "make -k check" was removed. Removing the "Timed out" line
avoids hanging
make -k check >> $TEST_LOG 2>&1 || true
grep "Timed out" $(find -name \*.out) >> $TEST_LOG 2>&1 || true
>
> If you don't want the tests, I'd suggest "unticking" the "Run testsuite"
> check
> box... This still allows you to enable some tests by "uncommenting" them.
>
I see, the "uncommenting" still means going to
individual lfs-commands/chapter*/ files, which originally contain
# make -k check >> $TEST_LOG 2>&1 || true
I see that when tests are disabled in the jhalfs configuration file, then
lfs-commands/chapter08/803-glibc has both
# make -k check ...
and
# grep "Timed out" ...
lines commented out as expected.
>
> > cd $LFS/jhalfs/lfs-commands/chapter08 && sed -i "/make -k check/d"
> 8*-gcc &&
> > sed -i "/test_summary/d" 8*-gcc
> > 4) After the make BREAKPOINT=825-shadow, which uses
> > a56a4fcd9c2e07dbddd1ef6270d7fbe0139ed9aa expires,
> > I see the test-logs/803-glibc-2.39 file (owned by root) is 29 bytes of
> size
> > and contains a date,
> > so the 803-glibc tests either were not running or didn't write any
> output.
>
> Seems correct since "make check" is removed...
>
>
>
> Pierre
> --
> http://lists.linuxfromscratch.org/sympa/info/alfs-discuss
> Unsubscribe: See the above information page
--
http://lists.linuxfromscratch.org/sympa/info/alfs-discuss
Unsubscribe: See the above information page