Re: [Gnu-arch-users] patch: automatic remove the corrupt revlib revision, second try

[Stephen J. Turnbull]

>>>>> "ams" == ams  <address@hidden> writes:

    Derek> Addressing Matthieu's concern. Now it prints a warning
    Derek> message. Also does a rename before the rm.

    >>    *sigh* Another good project succumbs to CADT syndrome.
    >> Now even tla developers ignore Tom.  Whut a revoltin'
    >> dervelopment.

Sorry, I keep forgetting there are people in this world who didn't
grow up watching Bugs Bunny.  That was a joke.

    ams> If anything is revolting, it is your attitude in discouraging
    ams> people from sending patches that they have in their tree.

Your inability to comprehend anybody else's posts before replying is
shocking.  While I dislike the patch, I harbor no objection to him
sending it, discussing it, and revising it---I'm often enough wrong
about technical details---and I made none in my post.

What I do have a strong objection to is the way he defends his patches
by addressing the letter rather than the substance of objections.
Take the warning.  It's useless AFAICS.  To those who don't want their
revlibs to disappear without their permission, it's rubbing salt into
the wound---too late to do anything about it unless they want to do
some serious forensic work.  ls -i is no longer useful.  To those
(like Derek and perhaps Mathieu) who _do_ want the automatic deletion,
the warning is a useless annoyance because they have no intention of
doing anything about it.

Also, I object to making specious arguments that other people's
heuristics are bad, so they shouldn't impede implementing his
heuristics in the mainline.  Sure, as Derek points out, the "inode is
file identity" heuristic is not very reliable in some situations.  In
others it is.  So the status quo is a useful (and conservative)
heuristic.  Derek's proposed patch is a useful (but extremely
aggressive and quite destructive) heuristic.  I know which side I want
my SCM to favor!

By constrast, Andy's suggestion of mv'ing the junk revisions into the
junk namespace is kind of idea I'm used to seeing in Arch.  It may
turn out to have weaknesses, but it takes the best from both lines of
thought, rather than grasping at any bogus attack on a valid
objection.

    ams> it might just be that Derek finds this type of functionality
    ams> useful and wants to share his patch with people.

That's what Derek himself says, of course.  Very plausible.  But Andy
seemed (and still seems) to think that Derek was submitting it for
merge into the mainline.

-- 
School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
               Ask not how you can "do" free software business;
              ask what your business can "do for" free software.