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 <tests/iso9660_early_ce.iso.gz >/tmp/iso9660_early_ce.iso $ ./grub-fstest /tmp/iso9660_early_ce.iso ls -- -l ... Device loop0: Filesystem type iso9660 - Label `ISOIMAGE' - Last modification time 2023-03-04 16:28:35 Saturday, UUID 2023-03-04-16-28-35-00 - Sector size 512B - Total size 68KiB ... Now zeroizing the modification date (offset 830 in PVD at offset 32768): $ dd if=/dev/zero of=/tmp/iso9660_early_ce.iso conv=notrunc bs=1 count=17 seek=33598 $ ./grub-fstest /tmp/iso9660_early_ce.iso ls -- -l ... Device loop0: Filesystem type iso9660 - Label `ISOIMAGE' - Sector size 512B - Total size 68KiB ... $ Well, it did not crash and there is no UUID. But i don't see the error message either. Switching to the branch with my proposed error text improvements of grub-fstest ... Still nothing to see. Obviously grub_errno is not set after function execute_command of grub-fstest.c. I checked that a bad ISO still causes a visible error message $ /grub-fstest /tmp/iso9660_ce_loop2.iso ls / -- -l ls : GRUB error 9 with message 'suspecting endless CE loop' ./grub-fstest: error: encountered 1 error during command execution. $ I also zeroized creation date at offset 32768+813 of the original ISO and verified that grub-fstest still shows an UUID. So the only thing i could prove by my test run is that indeed GRUB depends on ISO 9660 modification time for its UUID production. I failed to hack the function grub_error in grub-core/kern/err.c so that it shows grub_errmsg to me. Neither fprintf(3) nor write(2) seem in reach. The attempt to #include /usr/include/stdio.h or /usr/include/unistd.h fails with not found sub-header files. Have a nice day :) Thomas _______________________________________________ Grub-devel mailing list Grub-devel@gnu.org https://lists.gnu.org/mailman/listinfo/grub-devel
