Hi,
Glenn Washburn wrote:
> > > if [ -z "$debug" ] && [ "$RET" -eq "$xfail" ]; then
I wrote:
> > If indeed in a successful --xfail test RET must have the value of "$xfail"
> > (i guess 1), then it is concise and sufficient.
> [...] the only acceptable xfail case is for RET=1. If
> RET is greate
Hi,
i wrote:
> > Assuming variable "xfail" is be set to a non-empty string exactly if
> > argument "--xfail" is given, i'd replace:
> >
> > if [ -z "$debug" ] && [ "${RET:-1}" -eq 0 ]; then
> > rm -rf "$lukstestdir" || :
> > fi
Glenn Washburn wrote:
> RET should never be undefined
Hi,
Glenn Washburn wrote:
> [...] grub-shell-luks-tester cleans up after
> itself, if it returns success. grub_cmd_cryptomount has a test that
> expects failure. But grub-shell-luks-tester doesn't know that this is
> an expected failure and should cleanup and grub_cmd_cryptomount doesn't
> ever cl
Hi,
i wrote:
> > I meanwhile forgot why i proposed
> > > +unset TMPDIR
> > but there was some theoretical reason for this ...
Glenn Washburn wrote:
> My guess was that you wanted to unexport TMPDIR.
Yes. The reason was that i wanted to limit the effect of exporting to
the run of grub-shell-l
Hi,
please ignore my proposal about "set -e".
My assessment of "set -e" in
tests/util/grub-shell-luks-tester.in
is obviously wrong, probably because "set -e" is not inherited by
sub-shells.
The line
#! @BUILD_SHEBANG@ -e
enables script abort on command error, but causes reply
errexit
Hi,
(adding Vladimir Serbinenko to Cc)
Glenn Washburn wrote:
> --- a/tests/util/grub-shell-luks-tester.in
> +++ b/tests/util/grub-shell-luks-tester.in
> @@ -143,6 +143,7 @@ fi
>
> # Make sure that the dm-crypto device is shutdown
> cleanup() {
> +RET=$?
> if [ -e "$luksdev" ]; then
>
Hi,
The patches apply without complaints by "git am".
But when i run (again as superuser, shudder):
make check TESTS=grub_cmd_cryptomount
i still get in /tmp the empty direcories /tmp/17*.LUKS2_*.
Minor nitpick:
> [PATCH 2/4] tests: Cleaup the cryptsetup script in grub_cmd_cryptomount
> un
Hi,
Pascal Hambourg wrote:
> When
> booting from a EFI partition on a regular disk, then $root is set to the
> EFI partition (hdX,Y). When booting from a EFI El Torito image on a CD,
> then $root is set to the whole CD (cdX), and the El Torito image is not
> visible.
That's a surprise to me.
I w
Hi,
Askar Safin wrote:
> I CC Thomas Schmitt, because my work seems to be related to xorriso.
I checked that grub-mkrescue variables efidir_efi and efidir_efi_boot
are not leading to arguments of the xorriso run.
Insofar this change is transparent from the view of xorriso.
I can confirm that ma
Hi,
i post this as base of discussion, not yet as sincere patch proposal.
It seems desirable to bundle the temporary data of each particular test
in tests/grub_cmd_cryptomount.in in a single directory, as it seems
intended by the code but currently is not achieved.
So i propose to default TMPDIR
Hi,
the submission of v2 is hampered by the fact that the artful composition
of TMPDIR in tests/grub_cmd_cryptomount.in does not influence the worker
tests/util/grub-shell-luks-tester.in because TMPDIR is not exported.
So grub-shell-luks-tester creates its workdirectory directly in /tmp
rather th
Hi,
thinking more i believe that the currently used mkdir option -p is
inappropriate in tests/grub_cmd_cryptomount.in .
It hampers proper cleanup because the script cannot know how many
directories in the path to TMPDIR were created and need to be removed.
It is unusual because about all other t
Hi,
i wrote:
> > Further delete each created directory as soon as the command of its
> > test case is finished.
> > [...]
> > mkdir -p "$TMPDIR"
> >
> > output=`"$@" 2>&1` || res=$?
> > +
> > +rmdir "$TMPDIR"
Daniel Kiper wrote:
> s/rmdir/rm -rf/?
This is equivalent to the question
Hi,
with the freshly submitted patch applied, i see as superuser:
# make check TESTS=grub_cmd_cryptomount
...
PASS: grub_cmd_cryptomount
...
# echo $?
0
#
No new directories appeared in the root directory. During the "make check"
run i could spot single files /tmp/*LUKS* which did
If not TMPDIR is set by the user then the test grub_cmd_cryptomount
creates about 20 directories named *LUKS*_test* in the root directory
and leaves them there when the test ends.
Initialize in the test script the variable TMPDIR to /tmp if it is not
set or if it set to empty text. To be consisten
Hi,
i believe to have found out what's wrong with TMPDIR in
tests/grub_cmd_cryptomount.in
It does not get assigned a default value and is used without such a
value. Many other tests have the following gesture before using $TMPDIR:
: "${TMPDIR=/tmp}"
But grub_cmd_cryptomount.in has not and u
Hi,
i find in the root directory of my system a lot of empty directories
like
/1678114331.LUKS1_test_with_twofish_cipher
/1678114333.LUKS1_test_key_file_support
I believe they come from
tests/grub_cmd_cryptomount.in
where i read
eval testcase "'LUKS1 test with twofish cipher:'" \
@bu
Hi,
sorry for goofing up "git send-email" again.
This time not by omitting the empty line after the subject of a cover
letter but probably by leaving the commit id of the previous
"git format-patch" in the "git send-email" command line.
git send-email --to=grub-devel@gnu.org $directory $surplus
The UUID emitted by function grub_iso9660_uuid is derived from the
ISO 9660 Primary Volume Descriptor field "Volume Modification Date and
Time". But the error message about possible invalid content of this
field talks of "creation date" rather than of "modification date".
Signed-off-by: Thomas Sch
The UUID emitted by function grub_iso9660_uuid is derived from the
ISO 9660 Primary Volume Descriptor field "Volume Modification Date and
Time". But the error message about possible invalid content of this
field talks of "creation date" rather than of "modification date".
Signed-off-by: Thomas Sch
Hi,
the question is now:
Is it worth to correct an error text which i cannot get to show up ?
Vladimir 'phcoder' Serbinenko wrote:
> grub-fstest IMAGE ls -- -l
$ gunzip /tmp/iso9660_early_ce.iso
$ ./grub-fstest /tmp/iso9660_early_ce.iso ls -- -l
...
Device loop0: Filesystem type iso966
Hi,
i found a wrong word in an error message of grub-core/fs/iso9660.c
function grub_iso9660_uuid():
grub_error (GRUB_ERR_BAD_NUMBER, "no creation date in filesystem to
generate UUID");
The missing entity would be the modification date
data->voldesc.modified
rather than the creation
Hi,
somehow i managed to omit the first paragraph of the cover letter:
-
grub-fstest describes itself as "debug tool for filesystem driver" but
fails to forward to the user the grub_errno status and grub_errmsg of the
drivers und
Most return values of function execute_command() are ignored in
util/grub-fstest.c. Only one of two "loopback" occasions and the command
"cryptomount" care. They perform grub_util_error() if the return value
is non-zero.
So there seems to be a general lack of handling of problems found in the
drive
This part of tests/iso9660_test works on an intentionally bad ISO which
with older versions of GRUB will cause an endless cycle. Therefore the
test runs grub-fstest under the program timeout. A non-zero exit value
is taken as indication that the timeout was triggered. Non-empty output
from grub-fst
Check grub_errno after a command was performed. If not GRUB_ERR_NONE then
print grub_errno and grub_errmsg. Count the errors and exit 1 with a
message if the count is larger than zero when function fs_test() ends.
The number of printed messages is restricted to 3.
Introduce a new option -E which r
As reported by Victoriia Egorova in bug 65880, grub-mkrescue does not
verify that the expected argument of an option like -d or -k does really
exist in argv.
So check the loop counter before incrementing it inside the loop which
copies argv to argp_argv. Issue an error message similar to what older
Hi,
now that i know how to test grub-mkrescue out of the git clone, i also
gave your patch v2 a run.
Its mail form seems to be problematic:
- Alpine shows two leading blanks in the context lines instead of one.
Empty context lines show no leading blank. Long lines show intermediate
line break
Hi,
i wrote:
> > So i will start a new thread with the question:
> > How do i convince the git clone to produce programs and ISO for 64 bit
> > EFI.
Vladimir 'phcoder' Serbinenko wrote:
> ./configure --with-platform=efi --target=x86_64
Fast and concise as ever. :))
Thanks, Vladimir.
It works i
Hi,
Maximilian Stendler wrote:
> to keep the host installation clean, I would probably use a container.
Yes, a virtual machine came to my mind. Easy to clone and to dispose.
But there must be some better way to test a utility built from git
independenly of systemwide directories.
Vladimir 'phco
Hi,
on occasion of
https://savannah.gnu.org/bugs/index.php?65880
"heap-buffer-overflow in grub-mkrescue.c"
i try to get grub-mkrescue running from git.
My problem is that grub_util_get_pkglibdir() returns
/usr/local/lib/grub
and grub_util_get_pkgdatadir() returns
/usr/local/share/grub
whi
Hi,
Mingcong Bai wrote:
> I was testing for loongarch64-efi. As noted in the commit message, I
> found that Loongson's firmware incapable of handling non-upper-case EFI
> boot paths
If no test reports emerge about other platforms, then i would consider to
reduce the patch to what was tested and i
Hi,
Mingcong Bai wrote:
> Thanks for your review. I will now submit v2.
Well, it's not actually a review but rather pointing out a problem which
would probably cause a failure of the xorriso run, when the option
-hfs-bless-by i /System/Library/CoreServices/boot.efi
does not find "boot.efi" beca
Hi,
Mingcong Bai wrote:
> While FAT < 32 filesystems are not case sensitive (which grub-mkrescue
> creates
> as a FAT12 image via mformat with a size of 2.88MiB), it seems that
> some of Loongson's LoongArch-based firmware (namely those found on their
> latest XA61200 boards) seems to treat this f
34 matches
Mail list logo
