|
Yep,
Apparently that's what's been holding the internet together for the last thirty years. Let me explain;
I am currently working on a closed source DHCP client and server. A perfect implementation of RFC 951, RFC 1534, RFC 1542, RFC 2131, RFC 2132 and RFC 3046. The source code was very elegant and I was quite pleased. Verified and tested it against DNSMASQ, ISC DHCP and various other servers and clients. Everything works as expected!
But then...
I decided to test against older DHCP clients and servers... that's when I realized that for the last 25 years nobody even completely understood the DHCP protocol. Nearly 100% of the DHCP servers I am testing have bugs.
Older versions of Windows Server with a default install will send just the IP address and lease time. No T1 or T2 values which violates the RFC. It also does weird things like NULL terminate it's strings and identifiers. It also sends some of the DHCP values in the wrong network byte order. I'm finding some other DHCP implementations that implement broadcast/unicast incorrectly and ignore the flags. Some DHCP servers just SPAM the network with gratuitous DHCPINFORM messages.
In fact...excluding software written in the last five years... I can't seem to find any DHCP servers in the past 25 years that does NOT have any bugs.
My once beautiful code now has to check for wrong network byte order, NULL terminated strings, missing DHCP options, missing T1 T2 time values and various other weirdness.
You probably thought these old RFC written 30 years ago have been perfected. You were wrong.
Best Wishes,
-David Delaune
|
|
|
|
|
|
|
Internet Standard - Wikipedia
Quote: Engineering contributions to the IETF start as an Internet Draft, may be promoted to a Request for Comments, and may eventually become an Internet Standard.
The internet standard are listed here: Official Internet Protocol Standards » RFC Editor
Even after the document becomes a standard the correct nomenclature remains 'RFC'
Best Wishes,
-David Delaune
|
|
|
|
|
You forgot chicken wire and chewing gum.
Quote: You probably thought these old RFC written 30 years ago have been perfected Nope.
With all added complexity of everything else around, it seems to go the other direction.
|
|
|
|
|
And carrier-pigeons[^].
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
Yes, it is fun isnt it.
I put a DHCP server in a widows driver once. But at least it only had XP and 7 to deal with, and they are different enough!
|
|
|
|
|
Munchies_Matt wrote: widows driver Driving Miss Daisy?
Software Zen: delete this;
|
|
|
|
|
I had to look twice to get that!
"Go forth into the source" - Neal Morse
|
|
|
|
|
Like fixing valid HTML for Internet Explorer.
|
|
|
|
|
Should I start on automotive embedded code ? Mmmhhh... I better not start on automotive embedded code.
|
|
|
|
|
And what is wrong with Duct Tape & Super Glue? That was one the main structural components of some hardware I shipped to a customer...
|
|
|
|
|
You expected software to be good and bug free?
You should've known better...
|
|
|
|
|
We (i.e., where I work) were using VMWare's DHCP to identify users. Since they had already logged into the VM, they were already authenticated. All we need to know is who has a particular IP.
As it turns, it can lie when you ask it "who's got that IP right now?"
Particularly after any sort of upgrade. Even using a secondary identification, the uid assigned to the machine name. Both had to agree. Completely eliminated the lies since both must match or it failed.
Until the next upgrade.
So now, I wrote an internal uid/pwd system with SQL Server files for registering a user (and their Hashed pwd) and tracking last access time (in another table) - follows my rules.
Enough was way more than enough.
Ravings en masse^ |
---|
"The difference between genius and stupidity is that genius has its limits." - Albert Einstein | "If you are searching for perfection in others, then you seek disappointment. If you are seek perfection in yourself, then you will find failure." - Balboos HaGadol Mar 2010 |
|
|
|
|
|
This newfangled interweb thingy will never catch on, so it doesn{t matter.
CQ de W5ALT
Walt Fair, Jr., P. E.
Comport Computing
Specializing in Technical Engineering Software
|
|
|
|
|
61. My queen, right or wrong (5)
|
|
|
|
|
ERROR
ER Queen Lizzie
. R right
. OR
. wrong - anag.
Software rusts. Simon Stephenson, ca 1994. So does this signature. me, 2012
|
|
|
|
|
|
Isn't that a synonym not an anagram?
|
|
|
|
|
oops
Software rusts. Simon Stephenson, ca 1994. So does this signature. me, 2012
|
|
|
|
|
Just took dog for walk. Came back to the computer updating at 30% DURING ACTIVE HOURS!!! I believe I saved last Word document, but it shows 'recovered.' FMS!!!
And they wonder why Win 7 is still popular.
|
|
|
|
|
They seem so clueless don't they.
|
|
|
|
|
Apparently taking the dog for a walk is not an active hour
|
|
|
|
|
Precisely! The work morals of some people these days.
"If we don't change direction, we'll end up where we're going"
|
|
|
|
|
Didn't you know? Active hours only apply when you are actively using the computer. If you take a break to be physically active then you certainly obviously don't need your computer.
In all seriousness, I believe it will ignore the "Active Hours" setting if the updates require a restart to complete and the system is idle an arbitrarily decided number of hours.
Bit old but might still work... Active Hours doesn't work for Windows Update[^]
|
|
|
|