Re: [PATCH] Warn the users when more than 3 '-C' given.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Le Sat, 10 Apr 2010 12:12:58 -0700,
Junio C Hamano <gitster@xxxxxxxxx> a écrit :

> Bo Yang <struggleyb.nku@xxxxxxxxx> writes:
> 
> > Output a warning message to users when there are more than
> > 3 '-C' options given. And ignore the numeric argument value
> > provided by the additional '-C' options.
> 
> How were you bitten by the lack of this warning?  You gave four or
> five to see how output would change, spent sleepless nights but
> couldn't figure out what the differences between third and fourth
> levels are, and wasted too much time?

That sounding a bit harsh, I guess it is my turn to take the blame for
suggesting this in last week's thread :)


> IOW, what does this fix?

One practical advantage of this warning would be, in the very case of
adding meaning to an additional -C, that a user trying to use it on an
older version of git would get a warning that the program might not
indeed to what the user requested.

However, my first feeling was simply that, while it is usually harmless
to let the user specify a flag several time, when it changes nothing,
the situation is different when repetition of the flag is important -
it is closer to an invalid flag combination.

In fact, I even dislike that use of repetitive -C.  One could argue
that it is much like repetition of -v used in various programs to raise
verbosity.  But well, in our case, it is much more than just increasing
the level of details, it makes it use a different mechanism - even if
each time it is a superset of the previous one.

And what if someone comes with an idea of a "level of -C" that indeed
lays between two existing ones ?  Will we shift the meaning of the
existing ones ?  And what about one "level" that would not strictly fit
in the existing "superset" chain ?

What about instead using a more descriptive flag ?  That would be more
verbose typing, but then we can still keep the existing flags for
backward compatibility, and we also have shell command-line completion.

I'd think about something like:
-C -C     -> -Cunmodified (that one also for diff)
-C -C -C  -> -Chistory

I could also argue that "blame -M" could also be better placed as a -C
variant (it is also supposed to detect some copies), and could have as
fullname something like "blame -Csamefile".


> I personally do not see much value in this patch.  It would be just a
> hindrance to remember to remove this hunk when somebody improves the
> algorithm to add fourth level of detail to the inspection logic.

Well, the warning should trigger the 1st time that somebody tests his
fourth -C, right ?

-- 
Yann
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]