Older blog entries for dwmw2 (starting at number 105)

Must try to remember Advogato password. Have been manually extending lifetime of cookie in cookies.txt for the last three years.

hadess: Nat West's online banking system works fine with Mozilla if you fake the user-agent ident string to be something it likes.

Since Mozilla doesn't allow you to do that per-site, you have to do it globally, and then you stop showing up in site statistics anywhere as Mozilla; you show up as whatever you're masquerading as to make Nat West happy.

Apparently the Mozilla people think this is a good thing, which is why they refuse to implement per-site user-agent masquerading.

Since masquerading also breaks the Java plugins, it's probably best just to use Konqueror for accessing the Nat West online banking stuff.

robilad: Don't confuse -ko and -kb. Remember that even if there are false positives occasionally they're trivial to filter out of a diff. CVS won't even complain about them if they change when you commit. Extra information, within reason, is good. Lost information cannot be reclaimed.

I shall decline your kind offer to call you a vegetable right now, but if you ever import my code and eat its CVS tags, then one of your customers comes to me for help with it and I can't tell what versions they've got, I reserve the right to change my mind :)

RCS tags are evil

I disagree. Why should they give you 'bad diffs' when there's more than one person working on the project? If you're sharing a repository it just works as $DEITY intended, surely?

And if another project imports your sources, presumably they do it with -ko to avoid tag expansion. And if you disagree with that you might just as well assert that linefeeds are evil because some vegetable might work with your source on a DOS box and convert CR to CRLF.... hence you should write all your code on a single line, perhaps?

Finally, CVS tags _do_ give me information which the CVS command can't tell me on demand. When someone sends me some source code, or is trying to describe to me what some vendor has given them, if it has my $Id$ tags in I can refer immediately back to the real version that what they've got was _based_ on, and hence see the _changes_ which were made elsewhere. There's no other way to do it than to embed version information in the file which is going to survive through being merged into various other trees and source control repositories, and make it back to me, via whatever torturous route it takes, intact.

"Couldn't sleep and now I have ssh-method.c working. Going to clean up the code a bit (still will be a bit hackey when I turn it in) and also add a password dialog box so you won't have to put your password in the uri."

Alternatively, you could just ensure that you detach from your controlling tty to work around the SSH bug, and then SSH will happily invoke ssh-askpass for you if it needs a password/passphrase/skey/etc.

See how Evolution does it when invoking SSH.

These requirements apply to the modified work as a whole. If identifiable sections of that work are not derived from the Program, and can be reasonably considered independent and separate works in themselves, then this License, and its terms, do not apply to those sections when you distribute them as separate works. But when you distribute the same sections as part of a whole which is a work based on the Program, the distribution of the whole must be on the terms of this License, whose permissions for other licensees extend to the entire whole, and thus to each and every part regardless of who wrote it.

Thus, it is not the intent of this section to claim rights or contest your rights to work written entirely by you; rather, the intent is to exercise the right to control the distribution of derivative or collective works based on the Program.

UK Copyright law is amusing.

§301 of the Copyright, Designs and Patents Act 1988...

301. The provisions of Schedule 6 have effect for conferring on trustees for the benefit of the Hospital for Sick Children, Great Ormond Street, London, a right to a royalty in respect of the public performance, commercial publication, broadcasting or inclusion in a cable programme service of the play "Peter Pan" by Sir James Matthew Barrie, or of any adaptation of that work, notwithstanding that copyright in the work expired on 31st December 1987.

Looking at implementing an internal IPv6 network in parallel with the existing IPv4 network on RFC1918 addresses.

At first glance it looks nice and simple. IPv6 has its equivalent of RFC1918 addresses -- with even better semantics. It has 'site scope' addresses, and machines can be given _both_ 'site' and 'global' scope addresses, and basically do the right thing in all circumstances... using the internal addresses for internal communication, and the external addresses for everything else.

So each site on the intranet can have its own _global_ IPv6 link, with optimal routing out the local connection to the Real World, while tunnelling the site-scope addresses over the CIPE links with the RFC1918 IPv4 traffic. None of this NAT-from-internal-to-public-addresses crap which we have for IPv4.

Unfortunately it seems that the concept of site scope is to be abolished. Discussion of the various possible alternative implementation plans has failed to turn up any possibilities which are even _close_ to being as simple and effective.

I wonder if the martians are to blame. They know that TCPv4 isn't good enough for us to communicate with our Martian-invasion fleet, and they're determined to scupper our IPv6 deployment efforts by fucking it up for us... so they placed mind-control anal probes into the IPv6 WG and made them abolish the part which was going to actually let people deploy IPv6 sanely in large organisations.

Or maybe not. Perhaps there are just telcos involved?