[pkg-fso-maint] Commit style (was Re: [pkg-fso-commits] [SCM] Various non-packaged files branch, master, updated. d01a461083d18e8fd08fae1ec658b2086a9b6cb5)

[Steffen Moeller]

Hi Luca,

Luca Capello wrote:
> On Tue, 10 Mar 2009 23:32:05 +0100, Steffen Moeller wrote:
>> diff --git a/install.sh b/install.sh
>> index 5bd984d..2a7fc49 100755
>> --- a/install.sh
>> +++ b/install.sh
>> @@ -71,9 +71,9 @@ fi
>>  
>>  # general variables
>>  VERSION=1.1
>> -TESTHOST=www.debian.org
>> -RDATEHOST=ntp.fu-berlin.de
>> -WPA_SUPPLICANT_CONF=/etc/wpa_supplicant/wpa_supplicant.conf
>> +TESTHOST=${TESTHOST:-www.debian.org}
>> +RDATEHOST=${RDATEHOST:-ntp.fu-berlin.de}
>> +WPA_SUPPLICANT_CONF=${WPA_SUPPLICANT_CONF:-/etc/wpa_supplicant/wpa_supplicant.conf}
>>  DROPBEAR_KEYS="/etc/dropbear/dropbear_rsa_host_key /etc/dropbear/dropbear_dss_host_key"
> [...]
>>  # internal variables
>> @@ -221,8 +221,15 @@ ENVIRONMENT
>>                       (set to '$QI')
>>  	QI_VERBOSE_BOOT  set this to true if you want to see the kernel messages 
>>                       do this only when QI=true (set to '$QI_VERBOSE_BOOT')
>> +	RDATEHOST    machine with which to sync against with ntp (set to $RDATEHOST)
>>  	SINGLE_PART  set this to true if only one partition should be created
>>                       this is true by default for QI=true and false if SD_PART1_FS=vfat
>> +                     (set to '$SINGLE_PART')
>> +	TESTHOST     name of machine that should be tested for availability
>> +                     (set to '$TESTHOST')
>> +	WPA_SUPPLICANT_CONF
>> +                     local configuration which to transfer to neo
>> +                     (set to '$WPA_SUPPLICANT_CONF')
> 
> This is not the first time I noticed it, but I really do not like this
> type of commits: each commit should deal with a single issue, with a
> detailed explanation in the commit message if needed.  In this case,
> however, we have one single commit for three different things.
> 
> Please also note that this has nothing to do with the pkg-fso team: it
> is a general (and accepted) use to provide single patches, since they
> are easier to review.

accepted, I'll attempt to increase granularity, in my perception the three things were
only one thing (render variables with the character of being configurable to actually
be configurable from the command line and document it).

many greetings

Steffen