tom_newton

or

None of your "*" ones will work. Think carefully about what you are asking - substr takes a portion of the string - you are saying "are the first 18 characters of the URL equal to "http://192.168.0.*" - which they will never be, the * there is literal. Try "are the first 15 characters equal to "http://192.168." instead url.substring(0,15)=="http://192.168." isInNet is the proper way to do it for IPs... you'll still need to "deal with" DNS names though - the browser doesn't resolve them - one of your examples takes a stab here: isInNet(host, "10.134.32.0", "255.255.240.0") That's fairly specific isInNet(host, "10.0.0.0", "255.0.0.0") would catch all 10. addresses and isInNet(host, "192.168.0.0", "255.255.0.0") would catch all 192 addresses. dnsDomainIs(host, "myvle.org") || looks suspiciously like an example dnsDomainIs(host, "*misportal.net") || * won't work - replace it with a . to match all subdomains of misportal.net dnsDomainIs(host, "ats-ls01*") || dnsDomainIs(host, "ats-isa01*") || dnsDomainIs(host, "internet*") || dnsDomainIs(host, "intranet*") || These are fixable - but I won't tell you how until you tell me why Its a really bad idea to put in prefixes - it could cause you all sorts of grief. If you can give a good list of what you need to go direct, and what needs to go proxy, we may be able to help draw up a less confused proxy.pac

Yeah, but web protocols/flash may not be

Also: feel free to call & ask! (OR email.. or PM... or give an update number on here)

url.substring(0,10)=="http://192.168.0.*" will never match - as the second string is longer than 10 characters. Needs to be 0,17. Should probably be using "isinnet" tho...

Unfortunately, caching of realtime feeds isn't something that's feasible. Now if multicast had ever really caught on...

Well, the flights are booked - I will be there! In usual SmoothWall style, final staffing has been left to the last minute - but we can now confirm it'll be Claire and I from the Leeds office, and Chris from the US. This will be our first "joint" trade show - though BETT has felt like it for a few years now

This is an area i've got a lot of interest in. I always say, as soon as a company brings out an image filter that is: 1. Fast (and by this i mean fast enough for an edge device, not desktop) 2. Accurate - even 90% is not good enough when we are talking fase positives 3. Value for money - we're chasing edge cases here, it should not cost the earth ..at this point we will add that to our arsenal. Anyway - I haven't found it yet. I should imagine the best way to hunt down well-hidden student sites is to keep an eye on logs and reporting.

Java feels backwards becuase you're using a method from one object, but you're comparing two. Eg. Perl if ($foo eq $bar) C if (strcmp(foo,bar) == 0) Java if (foo.equals(bar)) So this one's the only one which "looks" uneven - counterintuitive, even. Not done java for ages, but it might be possible to do something like: if (string::equals(foo,bar))

John - if you could send em thru anyway, would be appreciated. If you could ever cadge the morning off work, remember you guys have an open invite at any of our offices to come and talk filters/drink our tea.

@john; so you're dealing with overblocks? The new system will lower these considerably further anyway If you two guys could send me your stuff... i'd appreciate it Also - this would not affect being able to ban search terms.... Still contemplating which bits we could have. Will have to talk to the devs again tomorrow...

@chrish; what "common content" was it if you dont mind me asking? @rinsewand; how fast would be "fast"? Under 24h do you? (though we may go to much more frequent updates RSN)

Ah, not talking about removing regex blocking - you'd still be able to block URLs by that means, and put in "straight" URLs. The "leeway" I refer to, is the ability to make the pagescanning faster AND use less RAM, at a cost of not being able to edit the files. We may well add in-page regex blocking though - which would block a page if it contained a particular expression. That's be a performance hit, but not bad if you only did a few.

Guys - do you ever add your own "custom content filtering" rules - and do you find it advantageous? Eg. do you ever "block all pages with the word rhubarb on them" - not just the URL block/allow rules. Thinking about limiting this ability to just our blocklist team, as it would allow us some.. leeway.. elsewhere. Thoughts?

Well, our Leeds office is very nice and its a great place to live. Can't comment on soton - the office is great (I am there now) but living at the travel binn sucks.

@DMcCoy; - which bit of the hours? You mean the 0830 start? Coz laties are (I am led to believe) becoming a thing of the past as the US team grows. @john; thanks again - it really means a lot to us that we have friends like you (it would seem rude to downgrade you to "customer" ) - it makes life most pleasant when the majority of the people we get to work with actually like working with us!

Chris, No - was fuuuull. Turns out I have avoided sharing a room - as Richard has found himself unable to attend - instead of Me, chris bragan and Richie, you've got me Chris and Claire. So we have had to shunt Chris out of the hotel, claire and I get a room each.. so all is good. Except I probably get lumbered with more "organisey" jobs.

Im still finding "new posts" to be less effective, but thats probably increased my page clicks Think i post less - but maybe because im so damn busy because no bugger wants to come and work here hint hint

Shame Southampton's "catchment area" is ruined by a large expanse of wet stuff. Even our one Islander applicant (Hi!) has decided its too far Still... gotta be some west yorks folk around, no?

OpenDNS?

Full disk? Perhaps? Again - this is the kind of issue that would benefit from support gaining access with a remote viewer tool - the reporting database has more stuff to go wrong than most places, and it is hard for me to give a direct answer to these questions - please believe that if I could I would If you raise support tickets and PM me the ID, I will keep tabs for you. 3gb db is not large - suggest running the optimise overnight?

Just a cheeky hoof to make sure this gets seen - would rather recruit through edugeek than elsewhere - I know you guys are all quality

Yes, but a redirect won't work on https - to varying degrees. Either the redirect won't apply to the encrypted traffic, or if it does, you'll get an HTTP certificate mismatch. You can't reliably redirect HTTPS->HTTP, except perhaps via MITM, which obviates the need for such tricks in the first place.

@AndyDavies: We'd love to help out with debugging any issues between Kerboodle and SmoothWall - whatever help you need, let me know, and if we, or our customers find anything "fun", we'll get in touch. PS. May be worth getting a 'sig which says you are a kerboodle type (which it appears you are?!)

@LeMarchand: send me your CV anyway - you never know - we're actually looking for 2 support people in Fareham, so there might be room for someone who needs a bit of training up