Advogato blog for dwmw2

29 Apr 2008

Tue, 29 Apr 2008 14:19:33 GMT

I pay my telephone bill to British Telecom by Direct Debit — it's taken from my bank account directly, under their control (albeit with fairly decent safeguards).

Strange, therefore, that I got a call yesterday from their missing payments department chasing up my last bill, which they hadn't bothered to take for some reason. They left a message with a number to call them back on, and a reference number. Yet when I called, the person there seemed unable to do anything useful like checking why they hadn't bothered to take the payment. She just said she'd have to get someone more clueful to call me. I wonder why I wasn't asked to call that person in the first place?

Immediately after the call I checked my bank statement, and it seems that the Direct Debit was actually taken — yesterday. So I helpfull called back and told them that, since they didn't seem clueful enough to work out for themselves what they were doing.

Today I got another phone call, and another British Telecom representative lied to me by telling me that the bill was still unpaid.

Fucking Useless Telco.

29 Feb 2008

Fri, 29 Feb 2008 22:47:52 GMT

I suppose we should try to keep the ppc64 ExcludeArch bug as empty as we've been keeping the 32-bit one, despite the fact that we don't really use much 64-bit userspace on Fedora/PPC.

So let's make a start with the fun bits... OCaml now builds on PPC64 Linux. Maybe I'll take a look at Modula-3 if I get a few moments to myself and need a little light relief.

I've looked at a few build failures on PPC/PPC64 this week, and (aside from the 'we need some nutter to port OCaml' bit) so far they've all turned out to be generic problems which just happened to bite here first. As usual, Fedora on other architectures benefits a lot from the mere fact that we also build it for PowerPC.

24 Feb 2008

Sun, 24 Feb 2008 10:45:51 GMT

Apparently travellers have voted Seoul Incheon airport the best in the world. I can't say I agree. I don't particularly enjoy airports at the best of times, but ICN was especially crappy.

For a start, there are almost no useful shops. Now, I can deal with airports without shops; there are a lot of those. But Incheon is just packed with perfume and shiny things and tat of all kinds. Although there's a huge concourse of shops, it's really disappointing to find that there's nothing but rubbish there, aside from two tiny bookstalls. And absolutely nobody selling DVDs, thus failing to provide the two basic ways of whiling away the time spent locked in tin cans. It was extremely disappointing.

If I'd wanted to buy 500 handbags from 17 separate but (to the untrained eye) identical shops, I'd have been perfectly happy though.

There was no Internet access either. Again disappointingly so -- it looked like there was, but on trying to pay for an hour's access and getting a few screens through the process you get a popup dialog saying "Try again, please after setting the Plugin!". Closer investigation shows that it seems to be trying some Windows-only plugin just to register for access!

Not a good airport. Hong Kong, from which I type this, is much nicer. Also lots of tat shops, but at least there are some real shops amongst them. And working (and free) wireless access. And a Virgin lounge, which helps... :)

7 Jan 2008

Mon, 7 Jan 2008 13:17:48 GMT

I'm slightly confused by Jeremy Clarkson's admission that he was wrong about the safety of publishing his bank details.

The Direct Debit Guarantee promises him a full and immediate refund in this situation. If he phones his bank, the money should be back in his account by the time he puts the phone down.

I have heard that Barclays Bank are very bad at honouring their obligations under the Direct Debit Guarantee, and one person tells me that's one of the reasons he now banks elsewhere.

But that's a reason for Jeremy to change banks, not to admit that he was wrong about publishing bank details -- which are, after all, on every cheque he's ever written, too.

8 Dec 2007

Sat, 8 Dec 2007 11:15:07 GMT

(Oops. Updated to use anoncvs as I originally intended. Not everyone has an account)

Why is it that the easy way of obtaining and building a Fedora kernel isn't documented anywhere? It looks something like this:


$ cvs -d :pserver:anonymous@cvs.fedora.redhat.com:/cvs/pkgs co kernel/F-8
$ cd kernel
$ cvs -d :pserver:anonymous@cvs.fedora.redhat.com:/cvs/pkgs co common
$ cd F-8
$ cat > GNUmakefile <<EOF
ppc: DIST_DEFINES += --without smp
ppc64: DIST_DEFINES += --without kdump
i686: DIST_DEFINES += --without pae --without xen
include Makefile
EOF
$ make $(uname -m)

There are HOWTOs out there, but they seem to recommend that you download the SRPM and extract it, instead of working directly from the original. And presumably if you ever want to update it you need to download a whole new SRPM and do it all over again, instead of just 'cvs update; make $ARCH'. I cannot fathom why anyone would ever want to work directly with SRPMS like that, for any package. Even when I'm just building local hacks I wouldn't do it that way; I'd always take a copy of the Makefile from CVS and keep everything together in its own directory rather than scattered around ~/rpmbuild mixed together with potentially conflicting files from other packages.

The bit with the GNUmakefile is optional, but useful -- you can disable the builds you aren't interested in to save time, and more to the point if the last package built is the one you're interested in, you have the fully compiled source tree left behind (in kernel-$VERSION/linux-$VERSION.$ARCH) when it's done. This makes it easier for you to hack on it and debug it. After installing the built RPM on the target machine, you can individually rebuild and replace files -- both modules and the kernel image.

For the development ('rawhide') version, you'd use 'kernel/devel' in the first cvs checkout instead of 'kernel/F-8'. You can obviously have both at the same time, or even just checkout 'kernel', which should get you the common/ directory too. And a bunch of other 'branches' of the kernel.

29 Nov 2007

Thu, 29 Nov 2007 17:37:59 GMT

At rsync://bombadil.infradead.org/f8-efika there is a slightly updated Fedora 8 tree which installs on Efika.

It contains a new kernel with support for the built-in Ethernet -- which used to be present in Fedora 7 but somehow got dropped before Fedora 8, although it should now be in the next official Fedora 8 kernel update.

It also contains a few installer-related changes which are in the process of going into the Rawhide development tree. And a fix for the moronic bug which broke PS3 installations at the last minute.

You can burn it all to DVD, or just burn the images/boot.iso and boot from that. Or if you don't have a USB CD/DVD drive for your Efika, you can do it all over the network with TFTP. If someone reminds me precisely how to do that, I might update this post -- or better still just put it straight in the Release Notes Wiki. Another option is just to copy the two files you need to a USB storage device and use them from there.

To install, proceed as follows:

Run the FORTH script from /ppc/efika.forth. This fixes up a bunch of issues in the device-tree, and allows Linux to work correctly.
Boot the installer, by booting the "netboot" image from /images/netboot/ppc32.img.
Go through the install as usual. (It'll run in text mode.)
After the install finishes and you allow it to reboot, you need to make sure that it runs the efika.forth script at every boot. You do this by editing the nvramrc. Use the nvedit command in OpenFirmware, and add the following at the end of the script:
s" hd:0 /efika.forth" $boot
You exit the nvram script editor by hitting Ctrl-C, and then use the nvstore command to save it.
Finally, set the boot-device and boot-file environment variables in OpenFirmware so that it knows what to boot:
ok setenv boot-device hd:0
ok setenv boot-file /yaboot/yaboot

(That's assuming you used the default partitioning, with a separate /boot partition as the first partition. If you used a single ext3 partition and no LVM, which would be a sensible thing to do, then it might be 'hd:0 /boot/efika.forth' in step 4, and you might want to set boot-file to /boot/yaboot/yaboot instead. Use brain, or consult adult.)

25 Nov 2007

Sun, 25 Nov 2007 20:16:25 GMT

Sweetness, sweetness I was only joking
When I said by rights you should be
Bludgeoned in your bed

And now I know how Joan of Arc felt
Now I know how Joan of Arc felt

14 Sep 2007

Fri, 14 Sep 2007 21:31:59 GMT

I was in a meeting today, and a colleague spoke of wanting to buy a 4GiB SD card to use in his Nokia 800. I pointed out that such a beast doesn't exist -- if it's 4GiB, it can't be SD-compliant; it has to be SDHC (or just broken).

I was actually being serious -- it's quite possible that SDHC cards might fail to work in some places where SD works. But I think he thought I was just being pedantic -- his response was to say that I should be taken out back and shot.

Should I make a formal complaint about this threat of violence? Do you think the HR department would take it seriously? Or is it just a harmless figure of speech which I've even used about myself on occasion?

17 Aug 2007

Fri, 17 Aug 2007 13:28:31 GMT

My employer really does suck sometimes. Or at least certain parts of it do.

A few months ago we started hearing rumours that their Cambridge office (from which I nominally work, when I'm actually in the country) was going to close. It's kind of been on the cards for a while, since there's only a few people working there and it's very expensive to maintain -- but after they had building work done recently and we were explicitly told that there were no plans to close it, we did sort of think it would probably last until their 10-year lease was up, in 2011.

But then we started hearing these rumours of imminent closure -- dates on which the IS folks were due to shut down the systems for the final time, etc. There were various dates, most of which are in the past now. But the rumours were persistent enough that I approached my manager and asked about it. He knew nothing. A few weeks and a few more rumours later, I approached the HR department too. They also knew nothing, but promised to look into it and get back to me within two weeks.

Almost a month later, after being chased up at least once, the HR lady finally admitted to having some information.... but actually refused to pass it on!. Instead, she insisted on flying to Cambridge a week after that for a meeting... when I was in Belgium.

Ever since then, she's been refusing to just send email like any normal person would. I'd let her use the phone if she really insisted, although I'd much prefer email -- but so far she's failed to schedule a time for the call when all three necessary parties (including my manager) can be present.

I hear more rumours (all I know at this stage is rumours) that they've sent out snail mail... which is impressive of them, since I was scheduled to be in Asia for the whole of the month. The archaic practice of making marks on a piece of mashed tree and physically transporting it to my house doesn't really work very well when I'm a third of the way round the globe.

I really don't understand why they can't just send me an email. You know -- the primary communication method of everyone in the company, and in fact most of the western world these days. Even my father can handle email; at least on a good day.

So far, it's been over a month since she first said that she has information, but refuses to pass it on. The only thing I have directly from her is "nothing will happen this month or next". The current rumoured closure date is October 1st :)

[Final sentence was once redacted, due to an excessive overreaction on the part of the people responsible for the above fiasco. But since they've been telling people that I've been making 'death threats', I figured I should put it back again to set the record straight. It really is just a figure of speech, which is in frequent use in the company in both internal and external communication:]

I swear, some people should just be taken out back and quietly shot. But instead they come and work in our HR and IS departments.

16 Jun 2007

Sat, 16 Jun 2007 16:21:02 GMT

seth writes:

rpm –rebuilddb
Then go to /var/lib/rpm and make sure __db* files are gone.

For me it was the other way round, just a few days ago. I'm so used to RPM being totally broken with respect to locking, and always hanging up, that it's a habit now to just kill RPM (with ^Z, kill -9 %1, because it doesn't even respond to ^\, ffs), rm -f /var/lib/rpm/__db* and then run it again.

Unfortunately, I stupidly did this (and installed more packages) while 'yum update' was running. Not very clever of me, I suppose -- almost an ESR moment. But not quite, because I really should never have been forced into the habit of removing those lockfiles manually; that RPM bug really ought to have been fixed a number of years ago.

Thankfully, although I was very scared for a moment, the database rebuild was successful. I then logged into another remote machine to update it, and what did I find? 38 copies of the nightly cron job all blocked on rpmq because of the same crappy locking issue. And I had to think really carefully before I removed the __db* files that time :)