Does a git pull have to be so big?

Does a git pull have to be so big?

[Nigel Cunningham]

Hi.

I try to do pulls reasonably often, but they always seem to be huge downloads 
- I'm sure they're orders of magnitude bigger than a simple patch would be. 
This leads me to ask, do they have to be so big? I'm on 256/64 ADSL at home, 
did a pull yesterday at work iirc, and yet the pull this morning has taken at 
least half an hour. Am I perhaps doing something wrong?

I'm using cogito .16-2 (ubuntu) and git 1.0.6.

Regards,

Nigel

#cg-fetch 
Fetching head...
Fetching objects...
progress: 114 objects, 256992 bytes
Getting alternates list for 
http://www.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git/
progress: 376 objects, 1413225 bytes
Getting pack list for 
http://www.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git/
progress: 453 objects, 1924312 bytes
Getting index for pack 221c50e73e5ab65afededc14f1df0541b59ebdd5
Getting pack 221c50e73e5ab65afededc14f1df0541b59ebdd5
 which contains 62727f8969438d99c3c34415d16611cf86f16140

(Still going)

Re: Does a git pull have to be so big?

[Brian Gerst]

Nigel Cunningham wrote:
> Hi.
> 
> I try to do pulls reasonably often, but they always seem to be huge downloads 
> - I'm sure they're orders of magnitude bigger than a simple patch would be. 
> This leads me to ask, do they have to be so big? I'm on 256/64 ADSL at home, 
> did a pull yesterday at work iirc, and yet the pull this morning has taken at 
> least half an hour. Am I perhaps doing something wrong?
> 
> I'm using cogito .16-2 (ubuntu) and git 1.0.6.
> 
> Regards,
> 
> Nigel
> 
> #cg-fetch 
> Fetching head...
> Fetching objects...
> progress: 114 objects, 256992 bytes
> Getting alternates list for 
> http://www.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git/
> progress: 376 objects, 1413225 bytes
> Getting pack list for 
> http://www.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git/
> progress: 453 objects, 1924312 bytes
> Getting index for pack 221c50e73e5ab65afededc14f1df0541b59ebdd5
> Getting pack 221c50e73e5ab65afededc14f1df0541b59ebdd5
>  which contains 62727f8969438d99c3c34415d16611cf86f16140
> 
> (Still going)
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 
> 

Use git://git.kernel.org/... instead of http.

--
				Brian Gerst

Re: Does a git pull have to be so big?

[Jeff Garzik]

On Fri, Jan 13, 2006 at 08:45:29AM +1000, Nigel Cunningham wrote:
> I try to do pulls reasonably often, but they always seem to be huge downloads 
> - I'm sure they're orders of magnitude bigger than a simple patch would be. 
> This leads me to ask, do they have to be so big? I'm on 256/64 ADSL at home, 
> did a pull yesterday at work iirc, and yet the pull this morning has taken at 
> least half an hour. Am I perhaps doing something wrong?

Two answers here:

1) Every so often, you download the entire kernel history all over
again, if you are using pack files, since most repositories are repacked
occasionally.

2) Every change sends the full updated (albeit compressed) file,
not a patch.

	Jeff

Re: Does a git pull have to be so big?

[Nigel Cunningham]

Hi.

On Friday 13 January 2006 08:54, Jeff Garzik wrote:
> On Fri, Jan 13, 2006 at 08:45:29AM +1000, Nigel Cunningham wrote:
> > I try to do pulls reasonably often, but they always seem to be huge
> > downloads - I'm sure they're orders of magnitude bigger than a simple
> > patch would be. This leads me to ask, do they have to be so big? I'm on
> > 256/64 ADSL at home, did a pull yesterday at work iirc, and yet the pull
> > this morning has taken at least half an hour. Am I perhaps doing
> > something wrong?
>
> Two answers here:
>
> 1) Every so often, you download the entire kernel history all over
> again, if you are using pack files, since most repositories are repacked
> occasionally.

Thanks for the reply. Can I avoid using pack files with Linus' tree? If so, 
how?

> 2) Every change sends the full updated (albeit compressed) file,
> not a patch.

Ok. I can cope with that. Redownloading the whole history however, I'd like to 
stop.

Regards,

Nigel

Re: Does a git pull have to be so big?

[Nigel Cunningham]

Hi.

On Friday 13 January 2006 08:48, Brian Gerst wrote:
> Nigel Cunningham wrote:
> > Hi.
> >
> > I try to do pulls reasonably often, but they always seem to be huge
> > downloads - I'm sure they're orders of magnitude bigger than a simple
> > patch would be. This leads me to ask, do they have to be so big? I'm on
> > 256/64 ADSL at home, did a pull yesterday at work iirc, and yet the pull
> > this morning has taken at least half an hour. Am I perhaps doing
> > something wrong?
> >
> > I'm using cogito .16-2 (ubuntu) and git 1.0.6.
> >
> > Regards,
> >
> > Nigel
> >
> > #cg-fetch
> > Fetching head...
> > Fetching objects...
> > progress: 114 objects, 256992 bytes
> > Getting alternates list for
> > http://www.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git/
> > progress: 376 objects, 1413225 bytes
> > Getting pack list for
> > http://www.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git/
> > progress: 453 objects, 1924312 bytes
> > Getting index for pack 221c50e73e5ab65afededc14f1df0541b59ebdd5
> > Getting pack 221c50e73e5ab65afededc14f1df0541b59ebdd5
> >  which contains 62727f8969438d99c3c34415d16611cf86f16140
> >
> > (Still going)
> > -
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel"
> > in the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at  http://www.tux.org/lkml/
>
> Use git://git.kernel.org/... instead of http.
>
> --
> 				Brian Gerst

Ok. I'll give it a try - is it related to the packed files thing Jeff spoke 
of?

Regards,

Nigel

Re: Does a git pull have to be so big?

[Brian Gerst]

Nigel Cunningham wrote:
> Hi.
> 
> On Friday 13 January 2006 08:48, Brian Gerst wrote:
>> Nigel Cunningham wrote:
>>> Hi.
>>>
>>> I try to do pulls reasonably often, but they always seem to be huge
>>> downloads - I'm sure they're orders of magnitude bigger than a simple
>>> patch would be. This leads me to ask, do they have to be so big? I'm on
>>> 256/64 ADSL at home, did a pull yesterday at work iirc, and yet the pull
>>> this morning has taken at least half an hour. Am I perhaps doing
>>> something wrong?
>>>
>>> I'm using cogito .16-2 (ubuntu) and git 1.0.6.
>>>
>>> Regards,
>>>
>>> Nigel
>>>
>>> #cg-fetch
>>> Fetching head...
>>> Fetching objects...
>>> progress: 114 objects, 256992 bytes
>>> Getting alternates list for
>>> http://www.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git/
>>> progress: 376 objects, 1413225 bytes
>>> Getting pack list for
>>> http://www.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git/
>>> progress: 453 objects, 1924312 bytes
>>> Getting index for pack 221c50e73e5ab65afededc14f1df0541b59ebdd5
>>> Getting pack 221c50e73e5ab65afededc14f1df0541b59ebdd5
>>>  which contains 62727f8969438d99c3c34415d16611cf86f16140
>>>
>>> (Still going)
>>> -
>>> To unsubscribe from this list: send the line "unsubscribe linux-kernel"
>>> in the body of a message to majordomo@vger.kernel.org
>>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>> Please read the FAQ at  http://www.tux.org/lkml/
>> Use git://git.kernel.org/... instead of http.
>>
>> --
>> 				Brian Gerst
> 
> Ok. I'll give it a try - is it related to the packed files thing Jeff spoke 
> of?
> 
> Regards,
> 
> Nigel
> 

Yes.  If the objects are packed then the only way to get the objects by
http are to download the whole pack.

--
				Brian Gerst

Re: Does a git pull have to be so big?

[Nigel Cunningham]

Hi.

On Friday 13 January 2006 09:44, Brian Gerst wrote:
> Nigel Cunningham wrote:
> > Hi.
> >
> > On Friday 13 January 2006 08:48, Brian Gerst wrote:
> >> Nigel Cunningham wrote:
> >>> Hi.
> >>>
> >>> I try to do pulls reasonably often, but they always seem to be huge
> >>> downloads - I'm sure they're orders of magnitude bigger than a simple
> >>> patch would be. This leads me to ask, do they have to be so big? I'm on
> >>> 256/64 ADSL at home, did a pull yesterday at work iirc, and yet the
> >>> pull this morning has taken at least half an hour. Am I perhaps doing
> >>> something wrong?
> >>>
> >>> I'm using cogito .16-2 (ubuntu) and git 1.0.6.
> >>>
> >>> Regards,
> >>>
> >>> Nigel
> >>>
> >>> #cg-fetch
> >>> Fetching head...
> >>> Fetching objects...
> >>> progress: 114 objects, 256992 bytes
> >>> Getting alternates list for
> >>> http://www.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git/
> >>> progress: 376 objects, 1413225 bytes
> >>> Getting pack list for
> >>> http://www.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git/
> >>> progress: 453 objects, 1924312 bytes
> >>> Getting index for pack 221c50e73e5ab65afededc14f1df0541b59ebdd5
> >>> Getting pack 221c50e73e5ab65afededc14f1df0541b59ebdd5
> >>>  which contains 62727f8969438d99c3c34415d16611cf86f16140
> >>>
> >>> (Still going)
> >>> -
> >>> To unsubscribe from this list: send the line "unsubscribe linux-kernel"
> >>> in the body of a message to majordomo@vger.kernel.org
> >>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >>> Please read the FAQ at  http://www.tux.org/lkml/
> >>
> >> Use git://git.kernel.org/... instead of http.
> >>
> >> --
> >> 				Brian Gerst
> >
> > Ok. I'll give it a try - is it related to the packed files thing Jeff
> > spoke of?
> >
> > Regards,
> >
> > Nigel
>
> Yes.  If the objects are packed then the only way to get the objects by
> http are to download the whole pack.

Ah. Well, I've followed the advice and switched to git://. Thanks again!

Nigel