> > Hi Stefan, > > Pendl Stefan wrote: > > I had the following wrong match: > > > > 2009-09-29 16:21:38, DEBUG : Found network service: > {3FA06DD1-FB69-4CE5-B782-A7692479FBDD} > > 2009-09-29 16:21:38, DEBUG : Reading DHCP address. > > 2009-09-29 16:21:38, DEBUG : Found DHCP address: 129.0.40.227 > > 2009-09-29 16:21:38, DEBUG : Found profile > 'GL[01].\|PRM0.\|INO[01].' matching IP '129.0.40.227' > > 2009-09-29 16:21:38, DEBUG : Profiles applying to the > current host:|DesktopPC| > > This resulted in removing software from several hosts with > different IP addresses too. > > > > I do not use IP matching, so I do not understand, what the > cause of the problem is. > > I do have several other host patterns with the pipe for > multiple different host names. > > > > Running WPKG 1.1.2 > > > > Thanks for any insight. > > It looks like you hit a very rare case of overmatching. Your > host definition > "GL[01].\|PRM0.\|INO[01]." contains exactly 3 period > characters ("."). WPKG > first tries a 1:1 match to the hostname (which obviously > fails with your > regexp). Then it tries IP matching. Obviously your regexp can > be splitted into 4 > parts which makes it look like it qualifies for IP matching: > > GL[01] . \|PRM0 . \|INO[01] . > > So it tries IP-matching. Unfortunately JScript does not throw > an error when > trying to parse "GL[01]" as an integer. So finally it thinks > that your IP is in > range. > > The work-around for WPKG 1.1.2 is easy. Just add another "." > and WPKG will not > think any more it could be an IP. > > So you might change > "GL[01].\|PRM0.\|INO[01]." > to > "GL[01].\|PRM0.\|INO[01].|dummy." > > > Meanwhile I've checked in a change which is supposed to > detect non-numeric > values and then the IP match fails (as expected in this case). > Get it from here (version 1.1.3-RC1) if you like to test it: > > > br, > Rainer > Hi Rainer, I will surly test the new version, but I will wait till Friday, so I can make sure to not break anything once more. I think I have to reinstall three of our CAD workstations :-P BTW, the change.log and changes file say, there was an error for matching four dots in a pattern, but it actually are only three dots. Four numbers separated by three dots: n.n.n.n ;-) --- Stefan |