On Wed, 2010-06-16 at 15:50 +0200, Jonas Meurer wrote: > thanks, i've applied some changes to the check scripts in svn trunk, > please take a look. unfortunately (as written in the previous mail), > your scripts are cluttered. these scripts really should be as simple as > possible, and the code in most cases is obvious enough that > documentation isn't required. Wow it's really strange how you refuse to simply use well-documented/tested code as is... especially for such a security sensitive package where strict documentation should be considered to be a must, and where uncertainties and doubts are killers.
Especially as really no one is harmed by such documentation,... > all of these are applied, take a look at the svn trunk. well... - You still use sed for no reason I can see - Not sure whether you check e.g. for the availability of blkid which is nearly guaranteed to be there, but not for whether the device exists is readable and a block device. > > Other changes I've made: > > - No longer check for the existence of blkid (this should be always there > > as util-linux > > is essential/required). > kept this one. Is there a special reason? If not I don't see how this fits your usual arguments when rejecting my code, about dropping unnecessary documentation and/or safety-checks. > checks aren't supported in cryptroot at all so far. feel free to provide > a patch, Well.... I guess it's not to difficult to do this for you :),... any I guess my patches would be rejected anyway (and I do currently not consider to change my coding style ;) ) With respect to my replacements for the check scripts: > given that it is kept simple, small I guess they were even simpler as the previous versions, not only regarding the stuff I've dropped (sed/dd/etc), but also regarding how to read the code e.g. splitted up the "compelex" boolen expressions in the if-statements to simpler expressions with a nested if-statement. Apart from that, the code was basically identical. > and uses common coding > style. Uhm,.. that's the coding and documentation style used throughout many manpages, all POSIX,... so it's not that uncommon, is it? > vol_id check scripts don't need to be > depreciated. either one still has vol_id from udev available, or one > doesn't. Don't understand this... I thought vol_id was completely dropped from udev, thus it should be generally not available starting with everything post lenny?! Cheers, Chris.
smime.p7s
Description: S/MIME cryptographic signature
