Re: CVS and Jar files: Should you import Jar into the Repository? Why or why not

[Richard Caley]

In article <address@hidden>, Greg A Woods (gaw) writes:

>> The abiliy 
>> to get a diff is nice but way, way down the list. 

gaw> Until/unless you understand the fundamental nature of RCS files and the
gaw> fact that CVS is merely an RCS front-end, you won't even have a clue why
gaw> you are so wrong.

I know it (even processsed them a couple of times), but it's
irrelevent. Storing as diffs is just an optimisation, not a necessary
feature. 

Consider: if I have 10 versions of a non diffable file I could store it
as 10 files plus an extra one giving verson and date information,
or I could store it concatinated in an RCS file. If anything the
latter is probably slightly more efficiant. 

>> I want to be able to build the system as it was last week.

gaw> Very good -- that's one of the reasons for using version control in a
gaw> software configuration management process.

Main one so far as I am concerned.

gaw> Clearly you don't seem to understand that CVS is just one of the many
gaw> tools you will need if you wish to accomplish all you desire.

Indeed, but tools should not be multiplied beyond reason. Unless there 
exists a tool which will diff the files CVS can't, and which can
trivially be kept in sync with CVS, CVS is the best tool for the job. 

(clearly if something else was used in place of CVS for the remainder
 of a system, the balance may be different)

gaw> CVS does not, and never ever will (unless it becomes something different
gaw> than it is today), be suitable for use in tracking changes to
gaw> non-diff-able and non-patch-able files. 

Works better than the non existant alernative.

Works perfectly well for practical use.

Feel free to propose an alternative, if it's better everyone will be
happy to hear. 

-- 
Mail me as address@hidden        _O_
                                                 |<