pkgstats: contribution was never that easy

Pierre · 2008-11-09 19:08:07

Allan wrote:

creslin wrote:
Wonder how this happened.. pkgstats 99.76 %
Whats the bet Pierre ran his own copy of the script without the package installed?

No. Either some people have broken pacakge dbs or some funny guys just sent us garbage. But that does not matter at all, bacause most data seem valid.

Allan · 2008-11-09 19:16:14

I failed...

Mikko777 · 2008-11-09 19:44:33

Seems pretty much everyone has both qt/qtmod and gtk installed.

I find it weird that kdemod4 weren't stronger against extra kde4 tho.
guess its the competition/ confusion with 10 different versions/ providers.

miggy · 2008-11-09 20:37:41

I, for one, ran the script without having it installed. I thought it would be funny. Sorry if I ruined your statistics--everybody start over and resubmit.

(PS that was the only packages I fiddled with;))

wantilles · 2008-11-09 20:40:05

I have 3 different machines running Arch x86_64 here.

The two of them have roughly the same role (desktop) and packages, while the 3rd one has a different set for a different role (server).

They all have the same WAN IP (behind a broadband connection).

Will I be able to send over two statistics (one for each role)?

creslin · 2008-11-09 20:48:18

wantilles wrote:

I have 3 different machines running Arch x86_64 here.
The two of them have roughly the same role (desktop) and packages, while the 3rd one has a different set for a different role (server).
They all have the same WAN IP (behind a broadband connection).
Will I be able to send over two statistics (one for each role)?

http://bbs.archlinux.org/viewtopic.php? … 21#p445621

Send the desktop and server stats on different days. I have my laptop and desktop running monthly pkgstats on different days.

wantilles · 2008-11-09 20:51:48

creslin wrote:

wantilles wrote:
I have 3 different machines running Arch x86_64 here.
The two of them have roughly the same role (desktop) and packages, while the 3rd one has a different set for a different role (server).
They all have the same WAN IP (behind a broadband connection).
Will I be able to send over two statistics (one for each role)?
http://bbs.archlinux.org/viewtopic.php? … 21#p445621
Send the desktop and server stats on different days. I have my laptop and desktop running monthly pkgstats on different days.

Yes it just blocked me.

However I sent the desktop data today.

Tomorrow I will send the server data.

Spider.007 · 2008-11-10 08:57:35

frullino wrote:

I like this idea, great! But I think that recognize a single poc would be usefull. You don't have to store IDs or username, you can md5 some unique data, like mac address ok the ethernet card and save it to the server. Something like this:
ifconfig -a | head -n 1 | awk '{print $5}' | md5sum
[...]

When the script is plain-text; it is very easy to replace 'ifconfig -a' with some random echo statement

Pierre wrote:

Allan wrote:
creslin wrote:
Wonder how this happened.. pkgstats 99.76 %
Whats the bet Pierre ran his own copy of the script without the package installed?
No. Either some people have broken pacakge dbs or some funny guys just sent us garbage. But that does not matter at all, bacause most data seem valid.

My IP is probably in there as well, I scp-ed the script from another location because it wasn't on package-list yet.

iphitus · 2008-11-10 10:39:00

I think the idea with the IP is more to stop someone accidentally submitting it twice, rather than to prevent abuse. The script 10x as long and complicated and still be easily abused so simplicity was preferred over complexity.

Besides, with the raw data, there's other ways extraneous entries can be filtered out.

Hide · 2008-11-10 13:22:12

If I understood the statistics correctly, there is .08% *not* using pacman. How is it possible that the script worked for them/him/her? (not to mention the obvious: "Where's pacman, dude?" )

Not that it's important, I just found the fact interesting

Pierre · 2008-11-10 13:25:36

There are people using pacman-git or pacman-color.

Hide · 2008-11-10 13:32:15

I wasn't aware of this, thank you

sujoy · 2008-12-16 14:52:28

when can we see some stats? i am excited

dolby · 2008-12-16 15:08:41

http://www.archlinux.de/?page=PackageStatistics

I wonder who is the one who submitted stats having only 1 package installed.

highvoltage · 2008-12-20 19:15:28

Cool idea. I just submitted my desktop.

cosma · 2009-01-01 15:53:56

I'd like to contribute, but I'm an italian FastWeb user, and thus I'm always NATted by my provider, dynamically sharing my IP with a lot of other users..

I'll wait until you add something more than just the IP to identify an host (Windows-like hardware based identifier? )

fukawi2 · 2009-01-01 21:23:14

cosma wrote:

... thus I'm always NATted by my provider, dynamically sharing my IP with a lot of other users.

ISP fail

patroclo7 · 2009-01-01 22:28:01

I'm a fastweb user too, but I do not see the problem. Or is there actually another archer with you same IP? Ok, arch is popular, but this is really unlikely. Or don't you actually succeed in using pkgstats with you connection? I have had no problem.

cosma · 2009-01-02 00:20:22

patroclo7 wrote:

I'm a fastweb user too, but I do not see the problem. Or is there actually another archer with you same IP? Ok, arch is popular, but this is really unlikely. Or don't you actually succeed in using pkgstats with you connection? I have had no problem.

Haven't tried... thought that no data are better than wrong data.

zatricky · 2009-01-15 07:02:54

cosma wrote:

Haven't tried... thought that no data are better than wrong data.

I wouldn't consider it "wrong" data if you only submit once.

elzurk · 2009-02-03 06:48:50

Dang, talk about KISS. Noob arch user here, and truly happy to see this kind of behavior by the devs!

Chris.Giles · 2009-02-15 08:13:35

As well as providing a cron-job with 'pkgstats', how about moving it to 'core' and adding it as a dependency for 'pacman'?

This might sound extreme, but it'd help to give Arch developers a truer indication of: how many people use our wonderful distribution; the popularity of each package.

Chris.Giles · 2009-02-15 08:14:23

In addition to what I've written immediately above, perhaps we could instruct 'pacman' to call 'pkgstats' once in every hundred executions. A random number generator could be used to achieve this and we'd only need to add two lines of code to 'pacman'.

This method would overcome the problem of 'pkgstats' not being called when users haven't loaded 'cron' or have moved their monthly cron-job directory elsewhere. What do you guys think about all of this?

Allan · 2009-02-15 09:31:18

I think that you need to learn to use the edit button....

And given you repeated yourself:
http://bugs.archlinux.org/task/12991#comment39515
pkgstats is an Arch specific construct and pacman aims to be distro agnostic. So pacman will not be calling pkgstats.

crouse · 2009-02-15 22:05:06

Chris.Giles wrote:

As well as providing a cron-job with 'pkgstats', how about moving it to 'core' and adding it as a dependency for 'pacman'?
This might sound extreme, but it'd help to give Arch developers a truer indication of: how many people use our wonderful distribution; the popularity of each package.

Not to mention, some users would prefer to keep (for whatever reasons) everything private. That should be their choice as well.
Sending data back, without the consent of the user, nope, not in favor of that myself.

Arch Linux

#76 2008-11-09 19:08:07

Re: pkgstats: contribution was never that easy

#77 2008-11-09 19:16:14

Re: pkgstats: contribution was never that easy

#78 2008-11-09 19:44:33

Re: pkgstats: contribution was never that easy

#79 2008-11-09 20:37:41

Re: pkgstats: contribution was never that easy

#80 2008-11-09 20:40:05

Re: pkgstats: contribution was never that easy

#81 2008-11-09 20:48:18

Re: pkgstats: contribution was never that easy

#82 2008-11-09 20:51:48

Re: pkgstats: contribution was never that easy

#83 2008-11-10 08:57:35

Re: pkgstats: contribution was never that easy

#84 2008-11-10 10:39:00

Re: pkgstats: contribution was never that easy

#85 2008-11-10 13:22:12

Re: pkgstats: contribution was never that easy

#86 2008-11-10 13:25:36

Re: pkgstats: contribution was never that easy

#87 2008-11-10 13:32:15

Re: pkgstats: contribution was never that easy

#88 2008-12-16 14:52:28

Re: pkgstats: contribution was never that easy

#89 2008-12-16 15:08:41

Re: pkgstats: contribution was never that easy

#90 2008-12-20 19:15:28

Re: pkgstats: contribution was never that easy

#91 2009-01-01 15:53:56

Re: pkgstats: contribution was never that easy

#92 2009-01-01 21:23:14

Re: pkgstats: contribution was never that easy

#93 2009-01-01 22:28:01

Re: pkgstats: contribution was never that easy

#94 2009-01-02 00:20:22

Re: pkgstats: contribution was never that easy

#95 2009-01-15 07:02:54

Re: pkgstats: contribution was never that easy

#96 2009-02-03 06:48:50

Re: pkgstats: contribution was never that easy

#97 2009-02-15 08:13:35

Re: pkgstats: contribution was never that easy

#98 2009-02-15 08:14:23

Re: pkgstats: contribution was never that easy

#99 2009-02-15 09:31:18

Re: pkgstats: contribution was never that easy

#100 2009-02-15 22:05:06

Re: pkgstats: contribution was never that easy

Board footer