Any idea why the git repo is so huge

Any idea why the git repo is so huge

[Leo]

Hello,

The git repo I just checked out a minute ago is 54M, 20 times bigger
than the org lisp files. Any idea why it is so large? Thanks.

Leo

Re: Any idea why the git repo is so huge

[Eric S Fraga]

At Tue, 02 Feb 2010 09:01:57 +0000,
Leo wrote:
> 
> Hello,
> 
> The git repo I just checked out a minute ago is 54M, 20 times bigger
> than the org lisp files. Any idea why it is so large? Thanks.

ummm because there is a great deal of active development?  Each change
requires an entry in the repository with the changes that entry was
responsible for.

Re: Any idea why the git repo is so huge

[Ian Barton]

> 
> The git repo I just checked out a minute ago is 54M, 20 times bigger
> than the org lisp files. Any idea why it is so large? Thanks.
> 
> Leo
> 
Don't forget it has a complete history of all org development since the 
initial import of the source.

You may be able to reduce the size of your local repo by running:

git gc

See http://kernel.org/pub/software/scm/git-core/docs/v1.6.0.6/git-gc.html

Ian.

Re: Any idea why the git repo is so huge

[Leo]

On 2010-02-02 12:29 +0000, Ian Barton wrote:
>>
>> The git repo I just checked out a minute ago is 54M, 20 times bigger
>> than the org lisp files. Any idea why it is so large? Thanks.
>>
>> Leo
>>
> Don't forget it has a complete history of all org development since
> the initial import of the source.
>
> You may be able to reduce the size of your local repo by running:
>
> git gc
>
> See http://kernel.org/pub/software/scm/git-core/docs/v1.6.0.6/git-gc.html
>
> Ian.

Of course. But git is supposed to be very efficient in storing history
and other things. I remembered when it first appeared when Linus was
still maintaining it, it could have the Linux kernel's source tree with
complete history and other things within 1.5 times the size of the net
source.

Git gc won't help because 54M is the size of the first checkout.

Leo

Re: Any idea why the git repo is so huge

[Richard Riley]

Leo <sdl.web@gmail.com> writes:

> On 2010-02-02 12:29 +0000, Ian Barton wrote:
>>>
>>> The git repo I just checked out a minute ago is 54M, 20 times bigger
>>> than the org lisp files. Any idea why it is so large? Thanks.
>>>
>>> Leo
>>>
>> Don't forget it has a complete history of all org development since
>> the initial import of the source.
>>
>> You may be able to reduce the size of your local repo by running:
>>
>> git gc
>>
>> See http://kernel.org/pub/software/scm/git-core/docs/v1.6.0.6/git-gc.html
>>
>> Ian.
>
> Of course. But git is supposed to be very efficient in storing history
> and other things. I remembered when it first appeared when Linus was
> still maintaining it, it could have the Linux kernel's source tree with
> complete history and other things within 1.5 times the size of the net
> source.
>
> Git gc won't help because 54M is the size of the first checkout.
>

I was under the impression that the one thing that git does not
advertise is space efficiency - it stores total file blobs for each and
every change. So I think its efficiency is in terms of speed, not in
terms of disk usage. And to be honest, if you worry about 54M in a
development scenario then I'd be surprised when you consider how cheap
disk is these days.

Re: Re: Any idea why the git repo is so huge

[Christopher Allan Webber]

Richard Riley <rileyrgdev@gmail.com> writes:

> I was under the impression that the one thing that git does not
> advertise is space efficiency - it stores total file blobs for each and
> every change. So I think its efficiency is in terms of speed, not in
> terms of disk usage. And to be honest, if you worry about 54M in a
> development scenario then I'd be surprised when you consider how cheap
> disk is these days.

Git does store those as individual files initially, but once you run git
gc it'll compress them and store only the deltas in the pakfiles.

After I run git gc my orgmode .git/ directory is compressed to 44MB.
Not a huge difference, but a difference nonetheless.
 - cwebb