diff mbox

[02/31] download: Add KDE mirrors.

Message ID 20160801181342.16203-3-david@craven.ch
State New
Headers show

Commit Message

David Craven Aug. 1, 2016, 6:13 p.m. UTC
* guix/download.scm (%mirrors)[kde]: Add kde mirrors.
---
 guix/download.scm | 47 ++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 46 insertions(+), 1 deletion(-)

Comments

Andreas Enge Aug. 1, 2016, 7:54 p.m. UTC | #1
On Mon, Aug 01, 2016 at 08:13:13PM +0200, David Craven wrote:
> * guix/download.scm (%mirrors)[kde]: Add kde mirrors.

This is a good idea! I tried it out and it worked. I think you can push
this already now.

But I think the "[kde]" should not appear in the commit message?
It would be good if someone more knowledgeable could comment.

Andreas
Ludovic Courtès Aug. 2, 2016, 9:55 a.m. UTC | #2
Andreas Enge <andreas@enge.fr> skribis:

> On Mon, Aug 01, 2016 at 08:13:13PM +0200, David Craven wrote:
>> * guix/download.scm (%mirrors)[kde]: Add kde mirrors.
>
> This is a good idea! I tried it out and it worked. I think you can push
> this already now.
>
> But I think the "[kde]" should not appear in the commit message?
> It would be good if someone more knowledgeable could comment.

It’s fine, but honestly it doesn’t matter much.  :-)

Please push, David!

Thanks,
Ludo’.
David Craven Aug. 2, 2016, 10:08 a.m. UTC | #3
Does the ordering of the mirrors matter?

I had them all in in the order listed on the kde website, but that
ended up causing really slow downloads from africa... =)
Leo Famulari Aug. 2, 2016, 3:47 p.m. UTC | #4
On Tue, Aug 02, 2016 at 12:08:22PM +0200, David Craven wrote:
> Does the ordering of the mirrors matter?
> 
> I had them all in in the order listed on the kde website, but that
> ended up causing really slow downloads from africa... =)

Do they offer a load balancer that we can put first in the list? Usually
such a machine would take care of such things.
David Craven Aug. 2, 2016, 3:52 p.m. UTC | #5
> Do they offer a load balancer that we can put first in the list? Usually
> such a machine would take care of such things.

I can't find one. Guix could also pick a mirror at random to
distribute the load.

[0] http://download.kde.org/extra/mirrors.html
Leo Famulari Aug. 2, 2016, 4:27 p.m. UTC | #6
On Tue, Aug 02, 2016 at 05:52:40PM +0200, David Craven wrote:
> > Do they offer a load balancer that we can put first in the list? Usually
> > such a machine would take care of such things.
> 
> I can't find one. Guix could also pick a mirror at random to
> distribute the load.
> 
> [0] http://download.kde.org/extra/mirrors.html

download.kde.org itself seems to redirect the request to one of the
mirrors. Maybe it should go first in the list?
David Craven Aug. 2, 2016, 4:40 p.m. UTC | #7
It's not a dns load balancer I don't think. dig download.kde.org
always resolves to 195.135.221.71. I don't know how to get curl to
show me the final destination after port forwarding and http
redirects.

Not that it isn't a good idea that is. I'm just testing my
understanding of networks that's all.
David Craven Aug. 2, 2016, 4:54 p.m. UTC | #8
download.kde.org itself seems to redirect the request to one of the
mirrors. Maybe it should go first in the list?

Oh yes you are right!

curl http://download.kde.org/stable/frameworks/5.24/attica-5.24.0.tar.xz
        ~
<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<html><head>
<title>302 Found</title>
</head><body>
<h1>Found</h1>
<p>The document has moved <a
href="http://www-ftp.lip6.fr/pub/X11/kde/stable/frameworks/5.24/attica-5.24.0.tar.xz">here</a>.</p>
<hr>
<address>Apache/2.2.22 (Debian) Server at download.kde.org Port 80</address>
</body></html>
diff mbox

Patch

diff --git a/guix/download.scm b/guix/download.scm
index 73c0e89..fa6fa45 100644
--- a/guix/download.scm
+++ b/guix/download.scm
@@ -224,7 +224,52 @@ 
        "http://ftp.de.debian.org/debian/"
        "http://ftp.fr.debian.org/debian/"
        "http://ftp.debian.org/debian/"
-       "http://archive.debian.org/debian/"))))
+       "http://archive.debian.org/debian/")
+      (kde
+       ;; Mirrors from http://files.kde.org/extra/mirrors.html
+       ;; Europe
+       "http://mirror.easyname.at/kde"
+       "http://mirror.karneval.cz/pub/kde"
+       "http://ftp.fi.muni.cz/pub/kde/"
+       "http://mirror.oss.maxcdn.com/kde/"
+       "http://ftp5.gwdg.de/pub/linux/kde/"
+       "http://ftp-stud.fht-esslingen.de/Mirrors/ftp.kde.org/pub/kde/"
+       "http://mirror.klaus-uwe.me/kde/ftp/"
+       "http://kde.beta.mirror.ga/"
+       "http://kde.alpha.mirror.ga/"
+       "http://mirror.netcologne.de/kde"
+       "http://vesta.informatik.rwth-aachen.de/ftp/pub/mirror/kde/"
+       "http://ftp.rz.uni-wuerzburg.de/pub/unix/kde/"
+       "http://mirrors.dotsrc.org/kde/"
+       "http://ftp.funet.fi/pub/mirrors/ftp.kde.org/pub/kde/"
+       "http://kde-mirror.freenux.org/"
+       "http://mirrors.ircam.fr/pub/KDE/"
+       "http://www-ftp.lip6.fr/pub/X11/kde/"
+       "http://fr2.rpmfind.net/linux/KDE/"
+       "http://kde.mirror.anlx.net/"
+       "http://www.mirrorservice.org/sites/ftp.kde.org/pub/kde/"
+       "http://ftp.heanet.ie/mirrors/ftp.kde.org/"
+       "http://ftp.nluug.nl/pub/windowing/kde/"
+       "http://ftp.surfnet.nl/windowing/kde/"
+       "http://ftp.icm.edu.pl/pub/unix/kde/"
+       "http://ftp.pbone.net/pub/kde/"
+       "http://piotrkosoft.net/pub/mirrors/ftp.kde.org/"
+       "http://mirrors.fe.up.pt/pub/kde/"
+       "http://ftp.iasi.roedu.net/pub/mirrors/ftp.kde.org/"
+       "http://ftp.acc.umu.se/mirror/kde.org/ftp/"
+       "http://kde.ip-connect.vn.ua/"
+       ;; North America
+       "http://mirror.its.dal.ca/kde/"
+       "http://mirror.csclub.uwaterloo.ca/kde/"
+       "http://mirror.cc.columbia.edu/pub/software/kde/"
+       "http://mirrors-usa.go-parts.com/kde"
+       "http://kde.mirrors.hoobly.com/"
+       "http://ftp.ussg.iu.edu/kde/"
+       "http://mirrors.mit.edu/kde/"
+       "http://kde.mirrors.tds.net/pub/kde/"
+       ;; Oceania
+       "http://ftp.kddlabs.co.jp/pub/X11/kde/"
+       "http://kde.mirror.uber.com.au/"))))
 
 (define %mirror-file
   ;; Copy of the list of mirrors to a file.  This allows us to keep a single