[uml-devel] Re: Auto-updating Jeff Dike tree

[Frank Sorenson <frank@tuxrocks.com>]

[-- Attachment #1: Type: text/plain, Size: 869 bytes --]

BlaisorBlade wrote:
> I use the attached script (questionable as much as you want, i.e. it kind of 
> works here, no guarantee, not even "I think it works") to download the 
> patches. It parses the patches.html in a very rough Way. I mean 
> "DownloadAllPatches".
> 
> Also, the generated series file must be edited: it appends the new patch list, 
> so if you use it as-is it has duplicate patches and won't work.

BlaisorBlade,

I've modified your DownloadAllPatches script a bit, and thought you 
might want the changes as well.  This version doesn't require editing 
for whatever versions are on the page (now auto-detects), rewrites the 
series file, and is quieter.

Frank
---------------------------------------------------------------------------
Frank Sorenson - KD7TZK
Systems Manager, Computer Science Department
Brigham Young University
frank@tuxrocks.com

[-- Attachment #2: DownloadAllPatches --]
[-- Type: text/plain, Size: 818 bytes --]

#!/usr/bin/perl -w

$| = 1;

$hostName = "user-mode-linux.sourceforge.net";
$host = $hostName;

print "Downloading latest patchlist\n";
system("wget -N -q http://$host/patches.html");

$KernelVer = 'none';

open PATCHES, "patches.html";
while (<PATCHES>)
{
	if (/name="(2.[^"]+)"/)
	{
		$KernelVer = $1;
		print "Kernel " . $KernelVer . "\n";
		$outDir = $KernelVer;
		mkdir $KernelVer;

		open PATCHLIST, "> $KernelVer-series";
		$currDate = `date`;
      chomp($currDate);
		print PATCHLIST "#\n#Downloading from Jeff Dike web-site on " . $currDate . "\n#\n";
	}

	next if ($KernelVer eq 'none');

	/href="(work[^"]+)"\>\s*([\w-]+)\<\/a\>/ or next;
	$url = "http://$host/$1";
	$name = "$2.patch";

	print PATCHLIST "$name\n";

	print "...patch $name\n";
	system("cd $outDir && http_proxy= wget -N -q -t 10 $url");
}