Older blog entries for dwmw2 (starting at number 88)

Hmmm. Got bored last week so I played with improving JFFS2 mount time. By delaying data node and data payload CRC checks till after the mount, on the basis that we don't actually _need_ them done that early, and by using a pointer directly to the flash rather than copying into RAM every time, we get a fairly dramatic improvement in mount time -- from ~3s to ~0.3, and from ~5s to ~0.75s on a iPAQs with 14MiB and 31MiB file systems, respectively.

Not bad. Now the iPAQ bootloader is slower to scan the JFFS2 to find the kernel and parameters in it than the kernel is it mount it later :)

Just as I thought it was going alright
I find out I'm wrong when I thought I was right
Always the same, it's just a shame
That's All

Is it just me, or is this kind of cluelessness from someone in that position plain fucking scary?

Hmmm. He said 'From: <>' not 'MAIL FROM: <>' -- the header not the envelope.
Too hasty to condemn someone?

Me?

moray: Not all, but many. Personally, I find I'm physically incapable of typing my initials without the '2' any more.

Decided I should try to overcome my irrational hatred of C++.

So I started playing with Opie; developing a little program for playing with GSM phones.

I chose to use gsmlib for the actual communication with the phone. Upon first glance, it seemed to be quite nice.

Unfortunately it seems its effective footprint on the iPAQ is about 2MiB - layers upon layers of derived classes with gratuitous autorefcounting for objects which are only ever going to be allocated and discarded once, all based on libstdc++.

Strictly speaking, the project seems to be well on course to achieve the stated goal - I shall be left with a deep-seated but perfectly rational hatred of C++. :)

Ankh: Yes, I know. That was the point - perhaps I should have said "By the time you leave, you ought to know the difference."

In fact, I think I shall go back and change it :)

In kindergarten, practise and practice are the same.

By the time you leave, however, you ought to know the difference.

Ilan: We do have GOTOs in modern programming languages. C99, for one. And it's a useful tool which allows cleaner, more efficient and more readable code to be produced. If used with caution, of course - but that goes for most things.

Got annoyed with the Red Hat kernel including Rik's rmap patch and not working on ARM, so I ported the rmap VM to ARM.

Then I got annoyed with it dying horribly on 'dbench 10' on my 32MiB swapless nfsroot box, when the Red Hat 2.4.13 kernel worked flawlessly. It OOM-kills just about everything except the dbench threads. That one I suspect I'm going to have to leave to Rik.

VM is hard. Let's go shopping.

Free wavelan in airports rocks.