[SOLVED] Why such large download for patches?

CollieJim · 03-28-2015, 04:37 AM

Hi,

I'm setting up a cross development environment for the Beaglebone Black.

I'm following the instructions at http://dev.gentoo.org/~armin76/arm/b...k/install.xml:

Code:

Code Listing 3.6: Obtaining the kernel

# git clone https://github.com/beagleboard/kernel.git

(Change to the directory)
# cd kernel

(Switch to an up-to-date branch)
# git checkout origin/3.8 -b 3.8

(Download a kernel.org and patch it, may take a while)
# ./patch.sh

What I'm seeing:

Code:

/data/devel/CrossDev # git clone https://github.com/beagleboard/kernel.git
Cloning into 'kernel'...
remote: Counting objects: 31953, done.        
remote: Total 31953 (delta 0), reused 0 (delta 0), pack-reused 31953        
Receiving objects: 100% (31953/31953), 17.52 MiB | 182.00 KiB/s, done.
Resolving deltas: 100% (21508/21508), done.
Checking connectivity... done.
/data/devel/CrossDev # cd kernel/
/data/devel/CrossDev/kernel # git checkout origin/3.8 -b 3.8
Branch 3.8 set up to track remote branch 3.8 from origin.
Switched to a new branch '3.8'
/data/devel/CrossDev/kernel # ./patch.sh
Cloning into '.'...
remote: Counting objects: 4051665, done.        
remote: Compressing objects: 100% (7516/7516), done.        
Receiving objects:  22% (914610/4051665), 373.13 MiB | 283.00 KiB/s

At the rate the download is going, .patch.sh is going to fetch 1.6GB !!!

Why is such a huge download needed to patch the kernel?

Puzzled,
Jim

veerain · 03-28-2015, 11:13 PM

Your patch script is using git to clone the whole repository of software perhaps a master branch plus a required branch. Usually a patch is just diff output of two directories so it has only changes. But in your case it's downloading the whole of source/ plus all of the intra changes. And then perhaps it may be using git diff to get patches or something else. You should read about git. Also see your patch.sh file to know what it is doing.