hey christoph, i just tried to fix most of the documentation issues you spotted. would you mind to take a look at the current svn version of cryptdisks.default and doc/crypttab.xml and see whether that satisfies you?
On 12/06/2010 Christoph Anton Mitterer wrote: > I might have again spotted several issues in the documentation which could > be improved: > > In /etc/default/cryptdisks: > 1) CRYPTDISKS_ENABLE > Clarify all allowed values, defaults if unset and most important > what this exactly does (I guess it simply disables the initscript) and > whether it only affects cryptdisks or also cryptdisks-early. > > > 2) CRYPTDISKS_MOUNT > How this exactly works,... are these simply mountpoints from fstab > and all options/etc. are taken from there? > Again, the missing default if the variable isn't set at all both fixed. > btw: Do you plan to deprecate this on the long term run? I guess passdev is > a more appropriate way for doing this. i didn't think about it yet. no need to depreciate it now, decision postponed. > 3) CRYPTDISKS_CHECK/CRYPTDISKS_PRECHECK > I guess this needs to be clarified in the following way: > The crypttab manpage already tells defaults. Are these the defaults set in > that file? > Or what exactly? This should be documented more exactly in both manpage/and > /etc/default/cryptdisks. > > I'd have expected CRYPTDISKS_PRECHECK to be set to "un_blkid"?! > > And in the /etc/default/cryptdisks it should also be documented that PRECHECK > is > only used for non-LUKS. fixed. > If you want assistance, I could write a better documented > /etc/default/cryptdisks, > perhaps with assistance from you by answering questions :) this is the same problem everytime you contribute: i would appreciate patches, but they have to be simple. everytime you send patches, they're cluttered with documentation, spaces and empty lines. i don't like the formatation of your code. please try to accept the coding and documentation style that is used in cryptsetup packages right now. i'm not the only one who decided to use the current style, several other contributors (david, michael, ...) used the same style in the past. and i don't accept changes that break this style. > btw: The manpage notes for both checks that the program is searched in > /lib/cryptsetup/checks/: > 1) Is it still ok to specify an absolute path to some other location? fixed in documentation. > 2) User's are not meant to put anything in /lib/ (by FHS). > So not sure how interesting this info on the search path is. > On could add functionallity that also some .../local/lib/cryptsetup/checks/ > path is searched, but there is no one that corresponds to /lib just one for > /usr/lib, > and then we should also add the same thing for keyscripts (searching them in > a "local"- > path). > These are of course not the most important features, but I guess it woul make > everything more perfect/conformant. that's a fair point. but i don't want to add even more options right now. maybe one day the whole checks/prechecks and keyscripts thing needs to be reimplemented in a cleaner way. but currently i'd like to leave it as is. greetings, jonas
signature.asc
Description: Digital signature
