Wesley's Tool-Box

There's been some DB errors on the website today because there had been reboots and MySQL failed to automatically start each time ever since upgrading to OS X 10.10 Yosemite. I knew of this problem, but since the disruption was starting to get annoying I decided to find a way to fix it. Apparently, the fix goes all the way to 2010 for fixing the same issue in OS X 10.6 Snow Leopard by Marko Tomic. Then a simpler version for Yosemite came up a few weeks ago, which is largely similar to a 2013 entry at Apple Support Communities.

This was happening because MySQL has been using a very old way to autostart the service and never bothered to update it, making it susceptible to failure for several years already. Until there's an official fix, the following should be entered into the Terminal.app, which is essentially creating a plist file for the LaunchDaemon to use on boot.


Upon entering the "nano" editor, type (or paste) the following lines.


Press Ctrl-O, Enter, then Ctrl-X to save and quit. Then you need to set the file's owner & permissions and load into the LaunchDaemon.


Finally, restart the Mac and MySQL should autostart. Remember to restart immediately.

Celine and I had some fun making light trails at home during the night. The whole thing was spontaneous, but the results were pretty good. There weren't so much to prepare. All I needed to do was to have my iPhone 5S shoot at 1/2 second shutter speed, while Celine swung this LED-equipped electric fan all over the place.

By default, iOS devices like iPhones and iPads use 4-digit passcodes for the lockscreen. You could also use alphanumeric passcodes that are more secure, but they bring up a keyboard on the lockscreen. This is rather cumbersome compared to a keypad.

What's often overlooked is that there is in fact a nice compromise between the two. You could use a long numeric passcode, and it'll give you a keypad just like when you use a 4-digit passcode, as you can see here. You do need to press a button after entering the digits, but that's because it won't let someone guess how long it is.

I think this is even more useful when coupled with the Touch ID in iPhone 5S. You can still use the fingerprint to conveniently unlock most of the time, and you'll feel more safe in that there is less of a chance someone else could bypass that and unlock using a passcode instead. And the "autofill" of the passcode when you press your finger to unlock doesn't reflect how long the passcode is, either. It just shows nine dots all the time. A perpetrator won't be able to guess the length of the passcode by glancing at your screen.

Late last month, there was an RFID-themed expo called RFID/IoT World Congress 2013 in COEX. The entry badge happened to contain an RFID card that you could use to tag at a booth so the company running it could know that you visited. It turns out that the card is basically a blank RFID card. My contact info must have been paired to the card's serial number, and the terminals merely checked the number and pulled the paired contact data from a central database.

When I took it home and poked it around with the only NFC-enabled smartphone I have, LG Prada 3.0, the card was a generic MIFARE Ultralight card without any locks or data. These types of card can only contain 64 bytes of data, 48 of which is user-programmable, so it's apparently very limited and cheap. That seems to be a sensible choice for something to hand out to thousands of visitors.

Because it was initially unformatted, regular NFC apps on the phone refused to do anything with it. But with the NXP TagWriter app, I was able to format and write data into it at once. To test, I tried to program the URL of this website and its title into it. It ended up being 47 bytes total, and the card reserved 2 bytes for something else, so it went over the storage limit. I omitted the hyphen and I was able to write the data successfully. When I tag the phone with it, it would load this website.

Looks like I should try to find some cheap MIFARE Ultralight cards like this one and happily program all sorts of things to it.

If you're playing Simpsons: Tapped Out game (iOS / Android) on your spare time like me, and also happened to be a developer that upgraded your phone to iOS 7 (beta 1), then you may encounter a bizarre bug where you automatically get login to a completely random, but different account.

Some leafing through the web yielded some clues on what was likely going on. It seems to be caused by the way EA implemented the device and account management.

Apparently, EA decided to keep session information of both the device and the account that it's logged in on the server. When a user exits the game and comes back, the game checks the device and automatically login to the account it was last associated with. Seems harmless, right?

오늘 (한국시간 2010년 11월 23일 새벽) 애플의 iOS 4.2가 공식 출시되었습니다. 새로운 기능이 여럿 있는데, 그 중 베타기간 동안 별로 언급이 없었던 것이 바로 "나의 iPhone 찾기(Find My iPhone)" 서비스의 무료 전환입니다. 원래는 $99/년 하는 모바일미 서비스의 일부로 제공되었던 기능이지요. 아마 경쟁 OS에서 유사한 기능을 무료로 제공하기 시작한 점이 영향을 미치지 않았을가 추측해 봅니다.

어쨌든, 이 기능을 쓰게 되면 원격에서 iOS 기기의 위치를 추적하거나, 기기에 원하는 메시지를 띄우거나, 발견한 사람이 못 쓰도록 잠그거나 내용을 완전히 삭제할 수 있어서 기기를 분실했을 때 유용합니다. 그런데 안타깝게도 위치기반서비스(LBS) 관련 법적 문제로 공식적으로 국내에서는 아직까지 사용이 안되도록 조치되어 있습니다.

다행스러운 것은, 제한을 거는 방식이 USIM 기반이라는 점입니다. 국내에는 KT에서만 공식으로 아이폰과 아이패드를 출시한 관계로, KT USIM이 꽂혀있을 때에만 "나의 iPhone 찾기" 옵션이 안 나타납니다. SKT USIM을 꽂으면 비로소 나타나지요. USIM이 개통되지 않았더라도, 종류가 일반 USIM이 아닌 데이터쉐어링(OPMD)용 USIM이라도 SKT 것이면 상관 없습니다.

그럼 구체적으로 켜는 방법을 알아봅시다.

The new wearable computing project has now seen the first fruit of labour. The main system module, a completely functioning system that only needs power source, input device, and monitor, is now in working order with its own casing. As you can see, it is only as big as a novel, but houses a Pentium M 1.6GHz, 1GB of memory, 100GB hard disk, and a power supply that accepts 8V to 13.6V DC voltage.