removing CVS in Handbook Updating and Upgrading chapter

Isaac (.ike) Levy ike at blackskyresearch.net
Sun Jan 27 20:52:33 UTC 2013


Warren, I owe you and Glen a drink for dealing with this exasperating thread.
I wouldn't continue if I didn't believe new users first installs are extremely important,

On Jan 27, 2013, at 3:15 PM, Warren Block wrote:
> On Sun, 27 Jan 2013, Isaac (.ike) Levy wrote:
> 
>>>> + ports, cvsup access end-of-lifed Feb 28
>> - great- drop all ports references to cvs/cvsup
>> 
>>>> - source, cvsup deprecated - no end-of-life date set
>> - please do not drop the current canonical base/src fetch method from the handbook!
>> 
>> However, cvsup is still merely deprecated, without a clear canonical replacement, (per notes in my previous email).
> 
> Understood, but really that is an argument against deprecating CVS, not removing it from the docs.
> 
>> Therefore, I hope I'm being clearly understood that it's cleaner for *users* to still fetch *base/src* with cvsup, until:
>> 
>> - canonical src fetch replacement is in place
>> - whenever cvsup for base/src has an end-of-life date set, (like ports does- Feb. 28!)
>> 
>> At that point, it seems rational that the last references to cvsup can disappear.  (The rest of CVS can disappear before that though?)
> 
> Say we leave CVS references in this chapter, and CVS for source goes away in (for example) six months.  Anyone who starts using CVS in that time will be forced to switch.  And they would be justifiably upset, because they were offered an option that was going away.

I can totally agree, all *CVS* verbiage can be removed from the doc.

I'm certain I don't have to explain to you that cvs(1) and cvsup/csup(1) are not at all the same thing.  I may be missing your point, but loads of users have used the cvsup for many many years, but many never really touched cvs proper.  Kill cvs in doc.

> My personal feeling is that continuing to suggest CVS for -STABLE or -CURRENT, even with a deprecation warning, is a disservice to those users.


I can certainly see that point.  cusp(1) has a man page, config files, etc… it's pretty self-evident to admins who want to use it.

--
However, the one place where people are getting tripped up, is the quick fetch required before buildworld/buildkernel:

I see nothing but benefit leaving one line before the buildworld/makeworld dance:
# csup /usr/share/examples/standard-supfile

And one more line after that,
"cvsup/csup(1) has been deprecated, and will be end-of-lifed when the replacement is completed in base"

It's a lot simpler than leading them down the SVN setup path, the state it's in today!

> Using CVS is still documented in Appendix A, along with a big deprecation warning.


Noted, for sure.

Best,
.ike




More information about the freebsd-doc mailing list