Re: Moving NetBSD from CVS to svn|git|hg|fossil

[Kamil Rytarowski <n54%gmx.com@localhost>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 19.03.2016 02:26, matthew sporleder wrote:
> On Fri, Mar 18, 2016 at 8:56 PM, Kamil Rytarowski <n54%gmx.com@localhost>
> wrote:
>> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256
>> 
>> On 17.03.2016 23:34, John Klos wrote:
>>> Hi,
>>> 
>>> There's been lots of discussion (although not here) about
>>> moving NetBSD from CVS to something else. There are many points
>>> to discuss, so it may be worth separating discussion in to one
>>> of several threads:
>>> 
>>> Discussion of the impact on the workflow of the NetBSD
>>> project.
>>> 
>>> Discussion about the long term maintainability of the commit 
>>> history of the project.
>>> 
>>> Discussion about the tools themselves, how they'll be
>>> integrated in to NetBSD, and how we will provide services for
>>> older and current NetBSD machines.
>>> 
>>> 
>>> Personally, I know nothing about any of these SCMs beyond basic
>>> use of CVS, so I'm mostly concerned that any new tools have
>>> minimal impact on the size of the OS (meaning not importing
>>> Perl / Ruby / python, et cetera, in to base) and operate
>>> decently even on modest machines (meaning one shouldn't be
>>> required to have hundreds of megabytes just to check out src).
>>> 
>>> John
>> 
>> I just tested checkout of FreeBSD's svn over https and the
>> process was using approximately 300M RAM (I shortly checked
>> top(1)).
>> 
>> I wasn't able to clone with depth=1 a NetBSD IIJ src with git
>> with 600M RAM.
>> 
>> Going out of cvs to svn/git will alienate significant part of
>> Tier 1 port machines and most of the Tier 2 ones.
>> 
>> Could you please list your wish list of features to add/improve
>> in cvs(1 )?
> 
> 
> Can you use jsonn/src instead?  https://github.com/jsonn/src
> 
> Also here is some ~/.gitconfig  for your smaller machine (I have
> more* below) [pack] threads = 1 [core] bigFileThreshold = 5m 
> [transfer] unpackLimit = 10 [push] default = simple
> 
> I can do a full clone in 415MB and a --depth 1 in far less.
> 

I confirm similar numbers.

It takes about 430MB for full clone (but I skipped resolving deltas
phase, it takes too long) and maybe 100MB less for a shallow clone.

> 
> 
> 
> *  experiment with the commented out stuff [pack] #windowMemory =
> 2m #packSizeLimit = 64m #deltaCacheSize = 2m #deltaCacheLimit = 25 
> #packSizeLimit = 2m threads = 1 [core] #packedGitWindowSize = 2m 
> #packedGitLimit = 2m #deltaBaseCacheLimit = 2m #compression = 1 
> #loosecompression = 1 bigFileThreshold = 5m [transfer] unpackLimit
> = 10 [push] default = simple
> 

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
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=pra+
-----END PGP SIGNATURE-----