- Good news
- PGSM Act.43 came out.
- Bad news
- Slept in, and missed my personal Iaido/Jodo time. And next week if I make it, I have to share the dojo with a Kendo team full of enthusiasm after today’s competition.
- Something new
- Sometimes, if you ignore a bug, it _does_ go away by itself.
Another Saturday morning, so another episode of PGSM. Even though we’re all terrified of what will happen after Act. 50, it’s taken nothing from the atmosphere of anticipation and jubliation on Saturday mornings on IRC.
And since I missed the dojo booking this morning, and haven’t had my gi pants repaired and couldn’t train in Jujutsu this afternoon, I spent the entire day on IRC having fun.
It was supposed to be my first Iaido and Jodo training in more than twelve months. And my first self-directed one in two years. Instead, I read Damia (by Anne McCaffery) until some stupid hour of the morning. >_< On the plus side, it’s a good book. ^_^
The other thing I played with today was my work on pulling apart the source code for WinNY that I found on the Internet once. It looks like it’s been decompiled from an unpacked executable, and large chunks are missing. The missing stuff is largely the interface support code, but half the software (including much of the network protocol) is done using Winsock callbacks, so I can only decode the first few messages (up to the first rekeying) of a conversation.
On the other hand, the file formats are quickly revealing their secrets to me now. I can now decode a nodelist into the node addresses, and pullall the metadata from a cache file. I actually had most of that already working, but the data I had was giving bad results. So I fired up the English-localised WinNY on my housemate’s laptop, which is based on the same version as the source code I had, and found that its cache files would decode fine. So I used them to complete the meta-data extracting code, and apart from the MD5 verification failing, it was good. Then I pointed it at the earlier data from a later version of WinNY, and instead of getting the ‘bad key’ I was getting before, I got output. O_O!! Anyway, I don’t know what was wrong with the old code. It’s gone now.
Oh, and in perl,
$value & 0xffff doesn’t seem to do what a C programmer like myself expects. Happily,
$value % 0x10000 does the job. ^_^