Revision of sensors-detect

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, 13 Mar 2009 17:29:08 +0100, Jean Delvare wrote:
> On Thu, 12 Mar 2009 19:41:30 +0200, Axel Thimm wrote:
> > On Thu, Mar 12, 2009 at 01:16:20PM +0100, Jean Delvare wrote:
> > > If not, it would be nice to have a workaround. I have one, which is a
> > > simple CGI script which fetches the file using the svn command line
> > > client and prints its contents:
> > > 
> > > #!/bin/sh
> > > echo "Content-type: text/plain"
> > > echo
> > > svn cat http://www.lm-sensors.org/svn/lm-sensors/trunk/prog/detect/sensors-detect
> > > 
> > > I've tested it locally and it works fine for me. I could run it on my
> > > home server, however I think it is better for the users if they get the
> > > script from lm-sensors.org. They really have no reason to trust my
> > > personal server (which may also disappear someday.)
> > > 
> > > Axel, is there a chance we could run this CGI script on lm-sensors.org
> > > and point users to it when we want them to test the latest version of
> > > sensors-detect?
> > 
> > I think I'd rather prefer a static solution, e.g. create a daily
> > checkout w/ keyword substitution. Could be part of the snapshot
> > creation script.
> 
> It isn't uncommon that I commit a fix and ask a user to give it a try
> right away. Having to wait for one day would be inconvenient in this
> case. I can live with it if there's no other way (that's probably
> better than the current situation) but I'd prefer a more frequent
> update rate.
> 
> Wouldn't it be possible to add an exception for sensors-detect to the
> post-commit hook? If file == sensors-detect then run "svn cat
> sensors-detect > some path" or something? That way we would ensure
> there is no delay between the commit and the snapshot, and also no
> daily checkout when the file doesn't change.
> 
> > CGIs are nasty in the sense that for security one may turn it off on a whitelist
> > basis (actually nothing uses CGI currenlty) and a simply CGI script
> > could be forgotten in a server update/hardening messing up the
> > download.
> 
> This wouldn't be a critical CGI, so if it breaks, we can simply fix it
> when we notice the breakage. But anyway, as the server admin it's up to
> you. If you say no CGI then no CGI.

Axel, any progress here?

-- 
Jean Delvare



[Index of Archives]     [Linux Kernel]     [Linux Hardware Monitoring]     [Linux USB Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Yosemite Backpacking]

  Powered by Linux