June 2011 - clo-devel - mailman3.common-lisp.net

[clo-devel] Putting statistics back into service.
by Jean-Claude Beaudoin 30 Jun '11

30 Jun '11

Hello common-lisp.net admins, I would like to propose my services to put back into service the download statistics that used to be available on the "stats" page before the machine upgrade. The currently available files date back to May 7, about 7 weeks ago. I already have an account on common-lisp.net for my MKCL and CL+J projects. My user name is: jcbeaudoin So I guess that a simple change in its group memberships is all that is needed to give me the required access to do the work (ie: modify the cron script and maybe read the content of /var/log/apache2). Cheers, Jean-Claude Beaudoin

2 8

[clo-devel] common-lisp.net domain registration expiring soon
by Daniel Pezely 24 Jun '11

24 Jun '11

The domain registration for common-lisp.net expires in a couple of days. I'm willing to cover the cost of 5+ year renewal. If the Donate button on the site will get funds to the right person, we can handle it that way. Alternatively, I can call the registrar and simply give them my credit card info directly. Please confirm or suggest how to do the donation/payment, and tell me the amount. (For domain only services with the same registrar, I believe it's US$70 for one year US$135 for five and US$235 for nine, which is the maximum duration that they'll allow.) Is the plan to remain with the existing registrar? Why five or more years: For search engine optimization (SEO) reasons, it's better to have a far future expiration. That is, certain search engines factor this into their ranking. It's based upon belief that domains of official organizations intend to keep the domain and therefore renew for several years at a time. An examples, Wikipedia and Lisp.org have expirations of 2015 and 2016, respectively. Also, there have been issues in past of this particular domain expiring and someone (Xach, perhaps?) doing an emergency payment to keep it going. Let's just take care of it. Thanks, -Daniel -- first name at last name dot com

3 3

Re: [clo-devel] [cmucl-imp] URL wrong on CMU CL home page
by Chun Tian (binghe) 22 Jun '11

22 Jun '11

Hi, rtoy What's the relationship between CMUCL home page and common-lisp.net? I'm one assistant system administrator of cl.net now, please describe your need. --binghe 在 2011-6-20，20:35， Raymond Toy 写道： > > On Jun 20, 2011 4:41 AM, "Chun Tian (binghe)" <binghe.lisp(a)gmail.com> wrote: > > > > Hi, > > > > In the left sidebar of CMU CL home page, all URLs have a base URL of "http://www.cons.org/" but "http://www.cons.org/cmucl/". This caused all those links (Documentation, Download) are invalid. > > Dang. This all used to work fine, but something has changed on cl.net. I'll have to dig into it. > > > > P. S. Why not register "cmucl.cons.org", just like what CLISP did. > > That doesn't work, but try www.cmucl.org. That should work. > > Ray > > > > > --binghe > > > > _______________________________________________ > > cmucl-imp mailing list > > cmucl-imp(a)cmucl.cons.org > > http://lists.zs64.net/mailman/listinfo/cmucl-imp

2 6

Re: [clo-devel] Cron <root@common-lisp> nice -19 /custom/bin/cron.root.checkbot.sh (failed)
by Hans Hübner 20 Jun '11

20 Jun '11

What can it be that takes so long? Do all these huge, long-running batch jobs make sense at all? Would it not be better if we had one batch job that does all the required regular scans sequentially? -Hans On Mon, Jun 20, 2011 at 1:00 PM, Cron Daemon <root(a)common-lisp.net> wrote: > process already running, exiting >

1 0

Re: [clo-devel] Cron <root@common-lisp> /custom/bin/cron.root.update-viewvc-cvsroots.sh
by Erik Huelsmann 16 Jun '11

16 Jun '11

On Mon, Jun 13, 2011 at 9:34 AM, Hans Hübner <hans.huebner(a)gmail.com> wrote: > I've fixed this - The antik project owner did not want CVS (who wants > CVS nowadays anyway?) Good point. I've removed CVS repository creation from the 'add-project' script because of it (actually, just commented out, not deleted). Bye, Erik.

2 2

[clo-devel] common-lisp.net downloading 100GB from itself?
by Erik Huelsmann 12 Jun '11

12 Jun '11

Looking at http://common-lisp.net/cgi-bin/awstats.pl I see that common-lisp.net is one of the main visiting domains of common-lisp.net. In fact, it's the main downloader: ~100GB. It's probably the checkbot crawler which does this, maybe because it's downloading all variants of output being generated by a cgi script or something. That would also explain why it keeps filling up the /tmp volume. The question would be: which cgi script, so we can prevent this behaviour... Bye, Erik.

1 0

[clo-devel] Historical web service on port 8000?
by Chun Tian 11 Jun '11

11 Jun '11

Hi, Today I temporary set up a web server on common-lisp.net's port 8000, and soon received lots of HTTP requests: 66.249.71.78 - - [2011-06-11 11:41:42] "GET /cl-xml/browser/tags/0.918?rev=11 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 11:41:56] "GET /cl-xml/browser/tags/0.918?rev=13&order=size&desc=1 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 11:42:10] "GET /cl-xml/browser/tags?rev=11&order=size&desc=1 HTTP/1.1" 404 066.249.71.78 - - [2011-06-11 11:42:24] "GET /cl-xml/browser/tags?rev=13&order=size&desc=1 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 11:42:39] "GET /cl-xml/browser/trunk/code/atn-lib?preselected=/cl-xml/browser/tags/0.918%3Frev%3D3 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 11:42:53] "GET /cl-xml/browser/trunk/code/atn-lib?preselected=/cl-xml/browser/branches/sbcl-0.9.x-testing HTTP/1.1" 404 066.249.71.78 - - [2011-06-11 11:43:07] "GET /cl-xml/browser/trunk/code/atn-lib?preselected=/cl-xml/browser/trunk HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 11:43:21] "GET /cl-xml/browser/trunk/code/atn-lib?preselected= HTTP/1.1" 404 0 ... 66.249.71.78 - - [2011-06-11 12:12:06] "GET /claw/browser/?rev=126&order=size HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 12:12:18] "GET /claw/browser/?rev=126 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 12:12:30] "GET /claw/browser/?rev=58&order=date HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 12:12:42] "GET /claw/browser/?rev=58&order=size HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 12:12:54] "GET /claw/browser/?rev=64 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 12:13:06] "GET /claw/browser/branches/claw-0.0.1/doc/CREDITS?rev=130&order=size&desc=1 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 12:13:18] "GET /claw/browser/branches/claw-0.0.1/doc/Makefile?rev=130&order=size&desc=1 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 12:13:30] "GET /claw/browser/branches/claw-0.0.1/doc/blender/figure1.blend?rev=144 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 12:13:42] "GET /claw/browser/branches/claw-0.0.1/doc/README?rev=130&order=size&desc=1 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 12:13:54] "GET /claw/browser/branches/claw-0.0.1/doc/blender?rev=130&order=size&desc=1 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 12:14:06] "GET /claw/browser/branches/claw-0.0.1/doc/blender/figure1.blend?rev=173 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 12:14:18] "GET /claw/browser/branches/claw-0.0.1/doc/blender?rev=173&order=date HTTP/1.1" 404 0 ... 66.249.71.78 - - [2011-06-11 13:10:18] "GET /clfswm/browser/clfswm/doc/keys.html?rev=180&order=size HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 13:10:31] "GET /clfswm/browser/clfswm/doc/keys.txt?rev=120&order=date&desc=1 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 13:10:44] "GET /clfswm/browser/clfswm/doc/keys.txt?rev=180&order=size HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 13:10:57] "GET /clfswm/browser/clfswm/doc/menu.html?rev=120&order=date&desc=1 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 13:11:10] "GET /clfswm/browser/clfswm/doc/menu.html?rev=180&order=size HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 13:11:23] "GET /clfswm/browser/clfswm/doc/menu.txt?rev=120&order=date&desc=1 HTTP/1.1" 404 0 66.249.71.78 - - [2011-06-11 13:11:35] "GET /clfswm/browser/clfswm/doc/menu.txt?rev=180&order=size HTTP/1.1" 404 0 All from the same IP (66.249.71.78), seems like it's checking out our SVN data? or, Did we have any web service on port 8000 in history? --binghe

3 2

[clo-devel] Cron job to check available disk vs nagios?
by Erik Huelsmann 11 Jun '11

11 Jun '11

Hi, While working on turning back on all the cron jobs that we are running on cl-net, I found there's a cron job to check available disk space. Shouldn't we be using other software, like nagios, to monitor the services? Maybe we can write plugins/task scripts to do any of the cl-net specific tasks, but having a framework like that sounds better. Bye, Erik.

7 15

[clo-devel] On Volunteering
by Hans Hübner 10 Jun '11

10 Jun '11

Hi, I have received off-list email regarding my posting about using nagios vs. some home-brew disk space monitoring mechanism. I don't have the time to discuss this at length, but I'd urge everyone not to waste their time with meta discussions. We need disk space monitoring, and we need someone who wants to set it up and maintain it. binghe volunteered, and I am confident that he'll come up with something that works and that he likes to keep running. Most importantly, it is something that he likes to do. I find it rather unpleasant to read advice from people with respect to the tools. We are not an operation where some tells someone else what to do, we do what we like for a common good. If you feel like you're in a tell-me-what-to-do-operation, I'll gladly play boss for you. -Hans

1 0

Re: [clo-devel] /tmp full again
by ehuels＠gmail.com 09 Jun '11

09 Jun '11

Hi Chun, the program checks all pages for links which are invalid, 404 or otherwise and generates a report for them. Bye, erik. Sent from my Nokia phone -----Original Message----- From: Chun Tian (binghe) Sent: 09/06/2011, 18:15 To: Subject: Re: [clo-devel] /tmp full again What's the purpose of this script? (I'll write disk space monitor as soon as possible) --binghe 在 2011-6-9，23:57， Raymond Toy 写道： > > In case no one has noticed yet, but /tmp is full again. It appears > that the Checkbot (based on the size of the /tmp/Checkbot.nnnn > directory) has consumed all available space. > > Ray > > > _______________________________________________ > clo-devel mailing list > clo-devel(a)common-lisp.net > http://lists.common-lisp.net/cgi-bin/mailman/listinfo/clo-devel _______________________________________________ clo-devel mailing list clo-devel(a)common-lisp.net http://lists.common-lisp.net/cgi-bin/mailman/listinfo/clo-devel

3 2

[clo-devel] /tmp full again
by Raymond Toy 09 Jun '11

09 Jun '11

In case no one has noticed yet, but /tmp is full again. It appears that the Checkbot (based on the size of the /tmp/Checkbot.nnnn directory) has consumed all available space. Ray

2 1

Re: [clo-devel] Using git for a common-lisp.net project.
by ehuels＠gmail.com 07 Jun '11

07 Jun '11

Hi Philippe, I know Raymond Toy set up the Oct project with git and trac. maybe you can have a quick look at that project? Raymond, could you tell us what you did? Bye, erik Sent from my Nokia phone -----Original Message----- From: Philippe Brochard Sent: 05/06/2011, 23:36 To: Subject: [clo-devel] Using git for a common-lisp.net project. Hi, I'd like to know how I can setup a git repository for a project on common-lisp.net (namely: CLFSWM)? For now I'm using svn on common-lisp.net and git on repo.or.cz. But since I use mostly git for my other projects and most of clfswm developers use the repo.or.cz git repository to participate, I'd like to use only git coupled with the trac on common-lisp.net. I maintain both git and svn repositories with some homemade scripts, but its not an acceptable solution (even if that works for some years now). So, is there a way for coupling git and trac on common-lisp.net and can I use it there? Thanks for your job and for any reply. Philippe Brochard (hocwp) - CLFSWM developer. _______________________________________________ clo-devel mailing list clo-devel(a)common-lisp.net http://lists.common-lisp.net/cgi-bin/mailman/listinfo/clo-devel

3 6

[clo-devel] Using git for a common-lisp.net project.
by Philippe Brochard 07 Jun '11

07 Jun '11

Hi, I'd like to know how I can setup a git repository for a project on common-lisp.net (namely: CLFSWM)? For now I'm using svn on common-lisp.net and git on repo.or.cz. But since I use mostly git for my other projects and most of clfswm developers use the repo.or.cz git repository to participate, I'd like to use only git coupled with the trac on common-lisp.net. I maintain both git and svn repositories with some homemade scripts, but its not an acceptable solution (even if that works for some years now). So, is there a way for coupling git and trac on common-lisp.net and can I use it there? Thanks for your job and for any reply. Philippe Brochard (hocwp) - CLFSWM developer.

1 2

[clo-devel] Beirc cvs broken?
by Zach Beane 07 Jun '11

07 Jun '11

I tried to check out beirc from CVS today, but got this: cvs export: failed to create lock directory for `/project/beirc/cvsroot/beirc' (/var/lock/beirc/beirc/#cvs.lock): No space left on device What should I do? Zach

3 3

[clo-devel] RT ticket #13: eliminating the need for /var/www/project(s)
by Erik Huelsmann 05 Jun '11

05 Jun '11

Hi, Today, I fixed Chun's rewrite rule (the regex was too greedy), meaning that it's no longer required to have the project and projects directories sitting in /var/www. I've already renamed them to project.OLD / projects.OLD and the SLIME site still works :-) The only scripts which seems to depend on the directories existing is cron.root.nag_about_empty_webpages.sh - however, that's not enabled (yet). Bye, Erik.

1 0

[clo-devel] Double version control in /custom/bin: _darcs and RCS?
by Erik Huelsmann 05 Jun '11

05 Jun '11

There seems to be double version control applied to /custom/bin, since there are 2 VC system admin directories. Which one do we choose? RCS is file-based, but good enough for our purpose, I'd say, especially since most of the files we're version-controlling are stand-alone. RCS then? Bye, Erik.

2 1

[clo-devel] Removed FTP cron jobs: FTP service long unsupported / no 'ftp' user on the system
by Erik Huelsmann 05 Jun '11

05 Jun '11

I just removed these files and directories: /custom/bin/cron.root.ftp_stats.sh /custom/bin/cron.root.ftp-update.sh /custom/bin/cron.root.chown_ftp_dirs.sh /var/ftp/pub/project/* (there were only the two most recent projects there, I assume created during project creation). Adapted: add-project.sh del-project.sh Bye, Erik.

1 0

[clo-devel] deprecated: add-viewcvs.sh
by Erik Huelsmann 05 Jun '11

05 Jun '11

Due to the way ViewVC has been set up, there's no need to specifically add projects to the ViewVC config anymore. The 'add-viewcvs.sh' script can thus be deprecated. Can I just delete it or do others hold different opinions? Bye, Erik.

2 2

[clo-devel] setup-svn.sh
by Erik Huelsmann 05 Jun '11

05 Jun '11

Just adjusted setup-svn.sh: it used to add configuration to websvn configuration files, which we don't have any longer. The specific bits have been removed. Bye, Erik.

1 0

[clo-devel] Added support: Anonymous checkout of Subversion repositories through HTTP
by Erik Huelsmann 05 Jun '11

05 Jun '11

Last year, there were some people asking for anonymous checkout of some of the Subversion repositories on common-lisp.net, because they were behind corporate proxies. While projects can set up the equivalent for Git themselves, this is not the case for Subversion. To support wider distribution of project sources from common-lisp.net, I've set up anonymous checkout for Subversion-using projects on http://svn.common-lisp.net/<project> The same URL allows browsing the HEAD revision of the repository, although that doesn't seem too relevant anymore, since ViewVC is now available for Subversion browsing too. The svn.common-lisp.net root directory listing is empty because the repositories are collected into one directory using symlinks instead of being true directories; there's not much I can do about that (other than patching the sources and rebuilding mod_dav_svn, which I'm not about to do). Bye, Erik.

1 0

[clo-devel] Subversion repository browsing enabled through ViewVC
by Erik Huelsmann 05 Jun '11

05 Jun '11

Turns out setting up ViewVC to do Subversion repository browsing was trivial once we had a mechanism to collect repository roots in one place. For Subversion, I set up /var/lib/svnroots which has the same purpose as /var/lib/cvsroots. I set up a cron job to fill the svnroots directory with symlinks on a daily interval. All symlinks which points to repositories which don't have a "revision 1" are deleted, assuming that's a good indicator of which projects use Subversion. The http://common-lisp.net/viewvc listing contains CVS and Subversion repositories sorted by alphabet. Bye, Erik.

1 0

[clo-devel] Updating the cl.net website?
by Erik Huelsmann 05 Jun '11

05 Jun '11

What's the preferred way to update the website? I think there are many files sitting in /var/www which don't have an equivalent in /project/clo/public_html. Should we start using rsync with the --delete flag to sync the two to make sure they are exactly the same? I think most of the time the only correct way to edit the clo website would be to edit the files in /project/clo/public_html, isn't it? Bye, Erik.

2 2

[clo-devel] Drop support for ArchZoom?
by Erik Huelsmann 05 Jun '11

05 Jun '11

Since I'm installing repository viewers, I had to look at the list of repository viewers we currently support. We decided to drop ViewSVN, basically because Trac is a great repository viewer and we can easily add more Trac instance, which will provide projects much more than just repository viewing. Since GitWeb, ViewVC and darcsweb have now been set up, the only repository viewer not set up (but supported before) is ArchZoom. Should we drop support for ArchZoom? Are there projects and people using Arch? [I can't even find arch in the list of Debian packages!] If it's not clear from the above, if it were up to me, I'd drop Arch and ArchZoom. Bye, Erik.

3 2

[clo-devel] ViewVC (formerly ViewCVS) set up
by Erik Huelsmann 05 Jun '11

05 Jun '11

Just finished setting up ViewVC. In order to make it work, I had to create a directory /var/lib/cvsroots which contains symlinks to all cvsroots in all projects. This takes advantage of ViewVCs feature to list all repositories in a specific directory. To keep the list current, I installed a nightly cron job: cron.root.update-viewvc-cvsroots.sh. No actual copying takes place, just symlinks to the original repositories. Additionally, I had to install a RedirectMatch because of the name-change. Bye, Erik.

1 1

[clo-devel] Subversion post-commit hook fixed & enhanced
by Erik Huelsmann 04 Jun '11

04 Jun '11

After the migration to new cl.net, Subversion commit mails were broken. Today, I fixed the issue (all links were pointing to a subversion location in /usr/local which didn't get ported over). The links have been changed to all point to /custom/bin/svn-post-{commit,revprop-change} Additionally, I've created a 'mailer.py' configuration file in /custom/conf/svn-mailer.conf which contains all the configuration to work with current and future repositories as long as they are being set up in the /project hierarchy. In the armedbear project, I've long maintained a custom post-commit hookscript which invoked Trac's post-commit script to close or update tickets. This functionality has now been built into the default script, if there is a 'trac' directory residing next to the 'svn' directory within the project. Bye, Erik.

1 1