Erwin Waterlander <water...@xs4all.nl> writes:
>>>> wcd -a
>>>> wcd -a
>>>> cat ~/.treedata.wcd
>>>>
>>> This is not a problem. Duplicate entries in the data file don't result
>>> in duplicate matches, because duplicate matches are filtered out.
>>>
>>
>> I understand that this is not a problem in the program, but it is
>> presents a problem for the user side:
>>
>> 1. Put directory $WCDHOME into version control
>> (take your pick: RCS, Cvs, Svn, Hg, Bzr, Git ...)
>>
>> Any chnage since last save to version control
>> will be reported "modified sources".
>>
>> 2. Add same directory again
>>
>> wcd -a
>>
>> => Not the .treedata.wcd has now changed content
>> => Version control reports changed files
>>
>> committing to version scontrol would be wrong thing to do.
>>
>> The multiple entries effective prevent using version control effectively
>> to save the states.
>>
>> If the program checked the entry before adding it to file, the
>> WCDHOME would stay pristine.
>
> Why would anyone add wcd data files to a version control system?
If you have 1TB disk and 100 wcd's, you do want to have backup. And you
want to notice chnages. Version does incremental backup control
efficiently with added binus that you can tag the versions and compare
chnages since etc.
> I also don't expect that a user types 'wcd -a' twice. The program does
> what the user asks it to do. Wcd is as intelligent as the user.
The program should not add entry twice. It is better to:
1) substitute the old entry with new
2) *or* warn the user about existing entry with the same name,
Consider variable assignment:
a=1
a=3
There is no two a's. Analogous to 'alias' commands in shells:
alias ll="ls -l"
alias ll="ls -la"
Jari
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
