Clay Baenziger wrote:
> Hi Joe,
>       Here's my findings:
>       On line 200 directy -> directory

fixed

>       On line 232 you set tmpdir_existed='NO', since you set the error
>          handler on line 249 but don't check to set tmpdir_existed='YES'
>          until 270 as this script evolves a race could develop where
>          we'd blow away the tmpdir unless I'm missing something. (Why
>          not just set tmpdir_existed to the correct state on
>          initialization or to yes for safety's sake and change line 270)

good catch.

fixed

> 
>       These two comments are probably best rolled into the one below for
>         lines 254-255
>       On line 255 why not use (( )) ksh arithmetic as used elsewhere,
>          so if (( $? != 0 )); then...?
>       On line 254 this checks for a uid=0, uid=1, uid=2... uid=9 it
>          seems I think you want id |grep -q 'uid=0(root)'?

done


> 
>       To replace lines 254-255 why not combine the two lines as:
>          "if $(id | grep -q "uid=0(root)"); then..."
>          (otherwise not being root causes an error from grep exiting
>           uncaught with a non-zero exit and invokes the error handler
>           instead of printing the you must be root error)
> 
>       On line 271, 276; if mkdir fails we won't get to the if statements
>          below checking to see if the directories exist (as the error
>          handler will be invoked and kick us out first (with no mention
>          of where we died (why not just check the return of mkdir or
>          are there cases where mkdir won't return an exit status on
>          failure? As is if mkdir fails, we'd just get:
> ------------------------------------------------------
> Error:
> 
> /export/backup/usbgen
>  Error or interrupt encountered. Exiting
> ------------------------------------------------------

As we discussed... I've tested this and it works well as is...

> 
>       On line 297 shouldn't this use the (( )) ksh arithmetic as used
>          elsewhere, so "if (( usb_size == 1 )); then...fi"?

fixed




>       On line 346 why use printf when print is used everywhere else?
>          For example, print "=== $0 finished at $(date)".

fixed. consistently using print


> 
>       Very cool job rescripting this in ksh93 and I like the trapping
> and other cool bits. Let me know where I probably got confused.
> 
>                                                       Thank you,
>                                                       Clay



Huge thanks you! for the help!

Joe


> 
> On Thu, 15 May 2008, Joseph J VLcek wrote:
> 
>> Joseph J VLcek wrote:
>>> Roland, and anyone else interested...
>>>
>>> Please review:
>>>
>>> http://cr.opensolaris.org/~joev/bug1462
>>>
>>>
>>> which addresses
>>>
>>>
>>> http://defect.opensolaris.org/bz/show_bug.cgi?id=1462
>>>
>>> This change represents a fairly extensive re-write of usbgen.
>>>
>>>
>>> Testing done:
>>>
>>> [1] Ran full Distro Constructor build
>>> [2] Built usb images using OpenSolaris-2008-05_rc3.iso
>>> [3] Because I added error handling I manually induced error into the
>>> script to confirm the error handling. I then removed the errors.
>>> [4] usbcopy-ed usb image to usb-stick, booted on test machine and ran
>>> full Install.
>>>
>>>
>>> Any and all input/suggestions most welcome.
>>>
>>> Thanks to Roland for the input he has already provided.
>>>
>>> Joe
>>>
>> I've addressed all review input received so far. Thanks Jean.
>>
>> The new webrev is at:
>>
>> http://cr.opensolaris.org/~joev/bug1462_2nd
>>
>> These changes addresses two bugs:
>>
>> http://defect.opensolaris.org/bz/show_bug.cgi?id=1402
>> http://defect.opensolaris.org/bz/show_bug.cgi?id=1462
>>
>>
>>
>> _______________________________________________
>> caiman-discuss mailing list
>> caiman-discuss at opensolaris.org
>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
>>


Reply via email to