Bazaar still below the radar when evaluating VCS tools

[Parth Malwankar]

On Sun, Feb 28, 2010 at 12:36 PM, Gary van der Merwe <garyvdm at gmail.com> wrote:
> On Sun, Feb 28, 2010 at 8:03 AM, Teemu Likonen <tlikonen at iki.fi> wrote:
>> * 2010-02-26 22:25 (+0100), Roland Mas wrote:
>>
>>> Parth Malwankar, 2010-02-26 23:27:17 +0530 :
>>>> One thing I think that has hit bzr is the "its slow" tag. While this
>>>> is no longer the case this tag has stayed.
>>>
>>> Allow me to slightly disagree there. I'll be the first to admit the
>>> “throughput” performance has increased impressively over time, there
>>> are still places where “slow” is the correct term.
>>
>>> Bazaar has definitely left the “unusably slow” category, and everyone
>>> involved in that has my most grateful thanks. But it's still not in
>>> the “fast” category. “Fast enough for common cases”, yes; unqualified
>>> “fast”, not yet.
>>
>> My current use of Bzr is just pull/log/diff in the Emacs repository.
>> Besides "bzr pull" I often notice delays with log commands:
>>
>>    $ time bzr log -l10 -n0 >/dev/null
>>
>>    real    0m8.103s
>>    user    0m7.788s
>>    sys     0m0.268s
>>
>> When the nearly-official Git mirror is up-to-date enough I use it
>> instead:
>>
>>    $ time git log -10 >/dev/null
>>
>>    real    0m0.007s
>>    user    0m0.004s
>>    sys     0m0.004s
>>
>> Eight seconds is not unusably slow for me but I think everyone prefers
>> 0.007 seconds to 8.1 seconds (both numbers are with warm cache).
>>
>> Obviously not all branches and repositories are like Emacs' but I
>> believe that speed is still an important factor when evaluating whether
>> a VCS tool is "below [or above] the radar".
>>
>>
>
> I believe that you are comparing apples with oranges by using bzr log -n0.
>
>

Thats what I think, though I don't know much about git. Without the -n0
(include merges):

[trunk]% time bzr log -l10 > /dev/null
bzr log -l10 > /dev/null  0.30s user 0.06s system 99% cpu 0.364 total

[emacs.git]% time git log -10 > /dev/null
git log -10 > /dev/null  0.00s user 0.00s system 65% cpu 0.006 total

While technically git is still much faster in the above case, ~0.3
seconds seems good enough to me in practice.

Regards,
Parth