Category: it could be called work

mozilla HTTP_REFERER workarounds

Bug 145579 – Website can see url of page visited after it (document referer used when loading images with javascript is incorrect while loading a new page)

Add the line

user_pref(“capability.policy.default.Window.onunload”, “noAccess”);

to your user.js file.

this:
user_pref(“network.http.sendRefererHeader”, 0);
also seems to be a valid workaround, though the first is preferred by content providers who track these things.

Mozilla HTTP_REFERER bug

Daily News – MozillaZine! Your Source for Mozilla News and Advocacy

Yesterday, ZDNet UK News reported that Mozilla has a privacy flaw involving HTTP referers. The flaw can be exploited using the onUnload JavaScript handler, which is triggered when a visitor leaves a page (for example, by clicking a link or using a bookmark). The problem is that the referer Mozilla sends is the URL of the page that the visitor is going to, not the page that he or she is exiting. This means that a site can discover where you are heading when you leave.

The security bug is present in the latest versions of Mozilla (including 1.0.1, 1.1 and 1.2 Alpha) as well as some Mozilla-based browsers, such as Netscape 6.x, Netscape 7.0, Galeon 1.2.x and Chimera 0.5. At the time of writing, no fix is available. A workaround is to disable JavaScript (Edit > Preferences > Advanced > Scripts & Plugins).

There is supposedly a fix for this, according to Slashdot but the page you need to see is suffering from the Slashdot Effect.

sad that this needs to be asked

How Mac and Windows Can Talk to Each Other

I have been a PC user since 1984. Until recently it has been my understanding that Mac files and Windows files are incompatible. Is that still true? If I should buy an iMac, could I still exchange files with people who use Windows 98 or XP?

This hasn’t been true for everyday applications for most of the 18 years this person has been harboring these delusions. No wonder Apple still struggles to get a foothold.

fear as a business strategy

Balancing Linux and Microsoft

After it bought Compaq this year, the combined company became the largest single buyer of Windows for personal computers and data-serving computers, and thus more dependent on Microsoft. A rising threat to Microsoft is GNU Linux, an operating system distributed free and developed using the open-source model in which communities of programmers donate their labor to debug, modify and otherwise improve the code.

After the merger with Compaq, Hewlett also became the largest vendor of Linux-based server computers, ahead of Dell Computer and I.B.M. Yet Hewlett’s bet on Linux still pales compared with its reliance on Microsoft. And after the merger, it was mainly former Compaq executives who took senior positions overseeing the Linux business.

According to the blurb at the Times, this article is supposed to be about how companies can balance their use of Microsoft and open source tools. The reality of the story is that even an organization as large as the merged HP/Compaq is unwilling to risk antagonising Microsoft.

Watch this space for the future of information retrieval

Nav4 Search Engine Patch Kit – Think Tank 23 Context Navigation Solutions

Give your search engine some words, and you’ll get back some documents that match one or more of those words. Give Nav4 a document, and you’ll get back the documents that match one or more of that document’s concepts. Is this search result close? Let Nav4 find others like it; let Nav4 get you closer.

Nav4 SEPK picks up where your search engine leaves off. Like what you’re reading? Want to know more? Nav4 gives you related documents, in context. You’re one click away from a Nav4 results page the additional documents you need. Or, why click? Embed Nav4 related documents right in your Web pages, with Nav4 Web services available in HTML and XML formats.

Look for this service to be impletmented here on this weblog soon: with each article you read, you’ll see links to other articles, here and on other weblogs, that are thematically similar. This is beyond the capabilities of keyword search, since it works like your mind does: it knows what a document is about, not how many times a given word appears.

I’m excited about it: this is the way information retrieval, aka search, should have always worked.

annoyances

  • apache2: there seems to be no easy way to migrate from apache’s httpd v.1 to version 2. I have tried a couple of times now and failed. The config file isn’t read properly and little things like SSI don’t work, cgi-bin isn’t active, and logging goes wrong. It sounds like the wrong config file is being read, but try as I might, I can’t get the one I want to be used. The apachectl script, for all its virtures, could benefit from either explicitly defining the file to be used or a path/directory in which to find it, like say /usr/local/apache2/conf.
  • Those online docs for the automounter are worse than bad, they’re dangerous. I found the one machine on which I had made those changes all but unusable. The only command I could run without getting an error that the system was unable to fork a process was echo which being shell built-in didn’t need any resources. I rebooted and messages had a whole lot of messages from the NFS process used by amd, so it seems to have just taken over the system. Needless to say, I have de-activated amd at boot time until I figure out if it’s worth documenting and then doing so.

browser standard compliance

I am in Windows 2000 right now and have looked at this weblog in IE 5.00.2920.0000 and Mozilla 1.0. IE manages to trash the layout of the page if the window is narrowed to squeeze the right column, while Mozilla just squeezes it all proportionally.

I suppose IE 6 fixes that, but this is IE 5, not 1 or 2.

my struggle with digital images

LED Casio QV Software

Arcgh. I needed to extract some images from a digital camera we used today (first day of kindergarten for my son and heir, so photos were required, the APS camera film hid from us, we took our old Casio QV-11 as well as my Nikon 8008).

So how to get the pictures out? Hmm, no modern Macs have serial ports, and this camera predates USB, so that’s out. I have used gPhoto before, but for some reason it failed to establish a serial connection. OK, I’ll see what I can do in Windows.

Hmm, the software that came with the ^*&^(*&)() camera doesn’t work: issues with the serial port. Now I’m getting annoyed. Windows says it can see the camera in its troubleshooting mode, so I Google up a freeware application from a UK software design consultancy, and by gum, it works. No serial issues, no whinging, just images, 78 of them (over a serial line, that takes a while).

Now, obviously the serial port is fine, so what happened? Why did two guys who did this just as proof of concept succeed where Casio and the gPhoto team were stymied? It’s especially annoying that gPhoto worked in the past and doesn’t now.

Hmm, so now I tried the Casio software and it worked. So once again, I have to wonder how these other guys managed to make this work, such that it works with other applications now. I did some power-cycling that presumably cleaned up any lingering connections, yet it failed until I tried QV (the freeware thing I found).

Interesting, in a frustrating sort of way.