[pkg-fso-maint] [Fwd: Re: install.sh: tasks functions

Steffen Moeller steffen_moeller at gmx.de
Tue Jun 16 11:56:53 UTC 2009


Ciao Luca, Brian,

Luca Capello wrote:

> Cc:ing the pkg-fso-maint@ mailing list, please in case reply there.
> 
> On Tue, 16 Jun 2009 10:23:42 +0200, Steffen Moeller wrote:
>> While the first patch seems fail-proof, I would need a spare
>> evening/weekend to test the 2nd bits which at the moment (read June) I
>> truly don't have. I'll happily apply them, but need the confirmation
>> of a successful run.  Luca, Joachim?
> 
> Apart from my comments below, I will try to perform some tests today or
> tomorrow: if everything is OK, I will apply them :-)

I have now applied the first patch.

>> On Mon 2009 Jun 15 12:52 you wrote:
>>> Luca Capello wrote:
>>>> On Sun, 14 Jun 2009 22:09:36 +0200, Brian DeRocher wrote:
>>>>> I don't see what calls action_all().  Not even line 1,496 will call it. 
>>>>> Also it won't do action_tasks() or action_cleanup().
>>>> Steffen is the master here, I would wait for his comments.
>>> the master is slightly embarrassed, Brian indeed found some dead code. It
>>> is (partly) nicer than the all) case of the main routine and the two should
>>> merge ... but not today ... as I don't even have time to send this reply.
>>>
>> Consolidated action_all() and all) case.  Dropped a couple informational 
>> comment because they are redundant.  Created two cases for partition and tasks 
>> because they have special stuff.
> 
> Please, consider each patch as a separate entity, thus do not mix
> changes that are not semantically common.
> 
>> diff --git a/install.sh b/install.sh
>> index 333ac20..d05c7bb 100755
>> --- a/install.sh
>> +++ b/install.sh
>> @@ -461,8 +461,6 @@ __END__
>>  action_all () {
>>  	cat <<__END__
>>  I: Running all stages with these settings:
>> -I: hostname set to $HOSTNAME
>> -I: microSD card device at $SD_DEVICE
> 
> What is the reason for that?  These are information bits, which IMHO are
> useful if you are paranoid, and preventing your question, I am ;-)

:)

it is partly redundant to information in action_configuration but I don't mind that,
in the contrary, I suggest to leave it in.

>> @@ -471,10 +469,25 @@ I: dash as default /bin/sh se to $DASH_BINSH
>>  
>>  __END__
>>  
>> -	for f in testing time partition format mount debian apt fso configuration 
>> kernel unmount
>> +	for f in testing time partition format mount debian apt fso configuration 
>> tasks kernel cleanup unmount
>>  	do
>> -		if [ -n "$VERBOSE" ]; then echo calling '$f'; fi
>> -		action_$f
>> +		case "$f" in
>> +			"partition")
>> +				if [ -n "$intern_no_partitioning" ]; then
>> +					echo " * Skipping partitioning since --no-partitioning flag set"
>> +					continue
>> +				fi
>> +				;;
>> +			"tasks")
>> +				if [ -z "$TASKS" ]; then
>> +					echo " * Skipping over installing tasks since none were specified"
>> +					continue
>> +				;;
>> +			*)
>> +				if [ -n "$VERBOSE" ]; then echo calling '$f'; fi
>> +				action_$f
>> +				;;
>> +		esac
>>  	done
>>  	cat <<__END__

the above seems fine to me

>>  
>> @@ -1467,25 +1480,8 @@ do
>>  			MOUNT_PROC="true"
>>  			;;
>>  		 "all")
>> -			action_testing
>> -			action_time
>> -			if [ -n "$intern_no_partitioning" ]; then
>> -				echo " * Skipping partitioning since --no-partitioning flag set"
>> -			else
>> -				action_partition
>> -			fi
>> -			action_format
>> -			action_mount
>> -			action_debian
>> -			action_apt
>> -			action_fso
>> -			action_configuration
>> -			if [ -n "$TASKS" ]; then
>> -				action_tasks
>> -			fi
>> -			action_kernel
>> -			action_cleanup
>> -			action_unmount
>> +			echo "Running all stages"
>> +			action_all
>>  			;;
>>  		 
>> "testing"|"time"|"partition"|"format"|"mount"|"debian"|"apt"|"fso"|"configuration"|"kernel"|"unmount"|"uboot"|"tasks")
>>  			if [ -z "$showHelp" ]; then
>> ----------
> 
> It seems OK for me and I have applied it to my local Git repository:
> once I have tested it, I will push it.

Please feel reminded about the removal of the action_all function.

Many thanks

Steffen



More information about the pkg-fso-maint mailing list