incubator-flex-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeff Conrad <>
Subject Re: What would it take to move to Git?
Date Sat, 11 Aug 2012 02:22:08 GMT
I went ahead and did it.

I was able to clone the Apache Flex git repo both from github and  It was a 174 MB download.  I've done this multiple times
over the last hour or so.  I was able to clone it on Windows 7 using Git
Bash, inside Eclipse using the EGit plugin, and I even cloned it also on a
vagrant box running CentOS 5.8 since I couldn't copy and paste the output
from Git Bash.

It took about 2-5 minutes to clone from the command line and significantly
longer--over 20 minutes--to clone from inside Eclipse using the EGit
plugin.  Below is the command line output from cloning the apache repo.

[vagrant@vagrant-centos ~]$ git clone git://
Cloning into 'flex'...
remote: Counting objects: 330636, done.
remote: Compressing objects: 100% (87765/87765), done.
remote: Total 330636 (delta 238772), reused 325026 (delta 233243)
Receiving objects: 100% (330636/330636), 174.64 MiB | 989 KiB/s, done.
Resolving deltas: 100% (238772/238772), done.

The git repository on Apache, and therefore the mirror on Github as well,
has the entire revision history.  There's so much history to page through
that Eclipse actually ran out of memory while I was trying to find the
first revision.  I found it, though, after a bit of digging on github.
 Here's the commit logs from the very first open source sdk revision on

It was on page 867 of the commit history:


2-5 minutes on the command line
174 MB download
You get a clone of the entire Apache Flex repository, with the complete
revision history.


On Fri, Aug 10, 2012 at 5:07 PM, Michael A. Labriola <> wrote:

> >OK, but the first clone will be huge?  I think our import file was in the
> 10's of GBs.  I'm afraid to try to clone the flex project over my home
> network.  I am testing by cloning the flex-site which worked from github.
> That is a big concern. I don't have a sense of how big it would be if all
> of the copies of files SVN has were made singular in the git method. We
> don't need to use github to figure it out though, we can make a local git
> repository and try this. I am willing to do so later next week when I
> should have time or, if someone else has time, they can do so first.
> Mike

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message