Hi, > >> 4 - documentation does not look right, it's "cyl/mb/gb/tb" (or > >> "CYL/MB/GB/TB"). "c/m/g/t" are not valid according to the various > >> strncasecmp() in the code. > > Because of the 'n' it is - only the length of the modifier of the number > > is taken, so this works. > > Indeed. Waow, perhaps dehumanize_number(3) was not available at the time > fdisk(8) was written. and now you can't change it without breaking current behaviour (or producing more code than necessary). > Minor knits I just noticed ((more for the sake of being clean): > > - meaningful errors when missing an argument in -s > - check for 'id' validity before converting via strtoll(3) Ok, thank you! Btw, something I wondered about before: change_part is given cbootmenu as an argument, though cbootmenu isn't even defined if BOOTSEL is not selected. Strange enough, that compiles. Perhaps I'm just too dumb, or the compiler too smart for me, but imho that should be fixed. Seems... unclean. I don't know the preferred way to fix it, so I won't do it. The most obvious way would be to define change_part differently, depending on BOOTSEL, or to check for the last param being NULL and then defining cbootmenu without respect to BOOTSEL. Regards, Julian
Attachment:
signature.asc
Description: PGP signature