[EMAIL PROTECTED] on 02/28/2000 02:32:11 PM
>[ On Monday, February 28, 2000 at 09:21:42 (-0500), Noel L Yap wrote: ]
>> Subject: Re: removing the need for "cvs add file" to contact the server....
>>
>> I wasn't asking for a vote. I was asking for a poll to gauge how my opinion
>> compares with that of the populous. I've always said that "cvs co", "cvs
>> export", and "cvs up" should be the only commands treating empty hierarchies
in
>> any special way.
>
>I'm telling you it's simply not a matter of "opinion". It's done deal
>in the current implementation. Period. Changing this would not only
>require a complete re-implementation but also some significant re-design
>of the repository structure.
Since it's just an optimisation, I don't see how it should require a complete
reimplementation. In any case, we'll see. Once you post the patch, I'll rip
out the stuff I don't like and post a repatch for those interested.
You're right. Popular opinion doesn't matter when it comes to what goes into
the standard source. It does, however, matter when it comes to what patches are
actually being used.
>> Packages that have any (ie empty or not) directories they wish not to be "CVS
>> managed" (by your definition) should not be within any CVS-managed
directories.
>> They should fix their build to reflect this.
>
>Wrong. This is the job for the "ignore" facility. It works A-OK too.
Again, you're absolutely right. Users shouldn't rely on the optimisation you
propose. They should instead be using the ignore facility.
>> Since "cvs add empty-hier" is an optimisation, it is not *necessary*.
>
>Who said that!?!?!? The optimisation is in not doing anything *after*
>not finding any files to add. There's no "problem" with actually
>running the command "cvs add empty-hier" -- it just won't cause any
>changes in the current workspace. Please try to keep your logic
>straight.
My logic is completely straight. I don't know what you're reading into my
statement. It's very simple. Optimisations are mathematically unnecessary.
IOW, the same functionality can be achieved without the optimisation.
>> It may
>> even be dangerous 'cos you give the false impression that there is no way to
get
>> CVS admin subdirectories within local empty directory hierarchies.
>
>Huh? What kind of hyperbole is that? If you want to add a file then
>add a file! If you don't then be happy that CVS will be careful not to
>do anything "extra" so that you won't end up wasting a potentially
>massive amount of time and effort with every subsequent command you ever
>run in that workspace.
You said it yourself. If people don't want CVS doing all this extra work, they
should be using the ignore facility.
Furthermore, your optimisation *will* cause CVS to do extra work when it comes
to empty directory hierarchies.
>> If so, "cvs
>> ci" will also have to create CVS/fileattr within the repo for each of the
>> directories it has created.
>
>If necessary, yes. However that's a side issue that only affects a
>secondary set of sub-commands. Normally there is no CVS/fileattr stuff
>in the repository unless you've used one of the sub-commands that
>specifically require it.
This is not a side issue. CVS commands are CVS commands. You cannot be biased
towards/against one "set" of commands just 'cos you don't use them.
Noel
