[5574] in Release_7.7_team
Re: Renaming "afslocker" to "attach" for non-Athena OpenAFS 1.4.1
daemon@ATHENA.MIT.EDU (William Cattey)
Fri Aug 11 12:17:37 2006
In-Reply-To: <2E432BF3-F431-41D2-9A74-4CD61C10D17E@mit.edu>
Mime-Version: 1.0 (Apple Message framework v624)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <0e44596bbd4172caf96589e96b884f3f@mit.edu>
Content-Transfer-Encoding: 7bit
Cc: openafs-release@mit.edu, release-team@mit.edu,
Alex T Prengel <alexp@mit.edu>
From: William Cattey <wdc@MIT.EDU>
Date: Fri, 11 Aug 2006 12:17:27 -0400
To: Jonathan Reed <jdreed@mit.edu>
X-Spam-Score: 1.217
X-Spam-Level: * (1.217)
X-Spam-Flag: NO
The traditional way to do this is to have a transition script that says,
"This way of doing things is no longer recommended use X instead. The
old way will stop working at event Y."
The script then proceeds to behave normally.
So I recommend we update the afslocker script to behave the same but
print to the standard error output:
With the OpenAFS 1.4.1 release, we offer "attach" as the replacement
for afslocker.
Please switch to using attach instead. The afslocker script will stop
working at the
next IS&T release of OpenAFS.
-wdc
"Like my father before me, I shall remain 5 years old till the day I
die!"
On Aug 11, 2006, at 11:59 AM, Jonathan Reed wrote:
>
> What do we want to do about machines which currently have 1.2 and
> afslocker? As I see it, we have a few choices:
>
> 1) Leave afslocker where it is.
> 2) Rename it and chmod a-x it
> (which we currently do if it exists and is a file as opposed to a
> symlink to the new version)
> 3) Replace it with a script which says "this is not what you want"
>
> I vote for #2, and perhaps a stock answer:
>
> Q: I installed OpenAFS 1.4, and now I can't run afslocker?
>
> A: afslocker was renamed to "attach"
>
> Thoughts?
>
> openafs-setup is a noarch RPM, so I can put out new builds while I'm
> away next week.
>
> -Jon
>
> On Aug 11, 2006, at 9:21 AM, William Cattey wrote:
>
>> Below, jdreed makes a case for renaming the "afslocker" script we
>> ship with non-Athena OpenAFS to "attach", and adding a "detach"
>> script. The rationale is that our attach does most of what Athena
>> attach does, and there is value to harmonizing the platforms.
>>
>> I have run this by Greg Hudson and he "sees no big pitfalls".
>>
>> I'm still hoping for a go-live of OpenAFS 1.4.1 for early next week.
>> I THINK we can make this change and still go live then.
>>
>> Here is my inventory of what's needed:
>>
>> 1. Revise the installer (again).
>> a. rename afslocker to attach
>> b. add detach
>> c. confirm installer still works after changes.
>> d. confirm attach and detach function
>> 2. Revise the detach stock answer.
>> (http://itinfo.mit.edu/answer.php?id=7017) to say "Yes." or just
>> delete it.
>> 3. Revise the Install doc
>> (http://itinfo.mit.edu/article.php?id=7297): rewrite penultimate
>> bullet item to call it "attach" and say it is of similar
>> functionality to "attach as provided by Athena".
>> 4. Revise the Getting Started doc
>> (http://itinfo.mit.edu/article.php?id=7754): rewrite section
>> "Navigating and afslocker" to make attach canon.
>> 5. Revise AFS at MIT: An Introduction
>> (http://itinfo.mit.edu/article.php?id=6845):
>> a. Change 4 instances of "attach or afslocker" to "attach"
>> b. Rewrite Navigation Shortcuts to delete "afslocker" references and
>> to detail any differences between Athena and non-Athena attach.
>>
>> Next steps:
>>
>> 1. Get sign-off from Bob Lang, Alex Prengel, and Heather Anne that
>> doing this is OK.
>> 2. Get sign-off from jdreed that the change can be made to the
>> installer.
>> 3. Milestone: change is made to the installer.
>> 4. The other 4 revision tasks are done.
>>
>> Bob, Alex, Heather Anne: What say you?
>>
>> -wdc
>>
>> "Like my father before me, I shall remain 5 years old till the day I
>> die!"
>>
>>
>>
>> On Aug 10, 2006, at 10:59 AM, Jonathan Reed wrote:
>>
>>> This is why I was wondering if we want to just call "afslocker"
>>> attach, to eliminate confusion, since it now mostly does everything
>>> attach does. I can easily add a "detach" command if we want one,
>>> which may be smarter that telling people to remove the symlink
>>> (especially if they're stupid and use a command which may follow the
>>> symlink and delete files in the locker)
>>
>
