http://www.postronic.org/h3/garbage in, garbage outenhttp://www.postronic.org/h3/images/feedlogo.gifhttp://www.postronic.org/h3/8831garbage in, garbage outhttp://www.postronic.org/h3/pid55.htmlhttp://www.postronic.org/h3/pid55.html... The beams were perfectly aligned, the large collider finally ready for the final test. One of the scientists was staring at the clock, while thinking about all those years passed in that lab, working so hard for such a long time, all for that moment. The chief engineer was adjusting the last parameters while all the others were sitting in front of the monitors, waiting in trepidation. Then the scientist said: > > A moment later there was nothing, then God said: "Let there be light", and there was light. The rest of this story has already been written a long time ago.One of the scientists was staring at the clock, while thinking about all those years passed in that lab, working so hard for such a long time, all for that moment.
The chief engineer was adjusting the last parameters while all the others were sitting in front of the monitors, waiting in trepidation.
Then the scientist said: << It's time! >>

The counter started at ten. As it went down to zero... the unexpected happened.
God came down to Earth.

As soon as he touched ground he gave a brief look around.
Some of the scientists were stuck. Some were kneeling, praying in tears. Many of them feeling guilty.
Then God said:
<< Holy shit! I was giving my usual look at the logs, and then I saw that horrible, HORRIBLE exception! >>
A couple of nerds looked at each other for a second, just a little less puzzled than the others.

He went on:
<< Thanks to the stack trace I could find this place, that I had mostly forgotten. You managed to make a big mess here guys. >>

Then one of the nerds took all of his courage and whispered:
<< ...we are so sorry God, please forgive us... >>

God gave a look at the monitors, the particles still perfectly blocked in the instant of the collision, as someone had paused a VCR, then he said:
<< No problem guys, it's ok to play around. It's the only way to learn something usefull, and I know, it's so fun! >>
<< I've already put a try/catch around the code, so that next time everything should be ok. >>

At these words the two nerds felt a little better.

He was already floating towards the sky when he added:
<< I had almost forgotten to tell you that I'll have to halt that cloud-computing thing and reboot your Universe. You know guys, hot code replacement never really works when you need it. >>

A moment later there was nothing, then God said: "Let there be light", and there was light.
The rest of this story has already been written a long time ago.]]>h32009-10-17T20:20:02+01:001http://www.postronic.org/h3/pid55.html#commentshttp://www.postronic.org/h3/pid54.htmlhttp://www.postronic.org/h3/pid54.htmlAnother new project of mine :) This is something i have been thinking about for a long time, because it's a tool I often need. So, here it is: Needlog (http://code.google.com/p/needlog/) Find a needle in a text-file-haystack! Needlog is a tool that can perform complex search queries on text-files of any size. Needldog aims to be user-friendly, cross-platform, and blazingly fast. This is the perfect tool for analyzing huge log-files.
This is something i have been thinking about for a long time, because it's a tool I often need.
So, here it is:

Needlog (http://code.google.com/p/needlog/)
Find a needle in a text-file-haystack!

Needlog is a tool that can perform complex search queries on text-files of any size.

Needldog aims to be user-friendly, cross-platform, and blazingly fast.

This is the perfect tool for analyzing huge log-files.]]>h32009-10-10T23:44:19+01:000http://www.postronic.org/h3/pid54.html#commentshttp://www.postronic.org/h3/pid53.htmlhttp://www.postronic.org/h3/pid53.html Lessons learned: Do not use woodUse a cheap camera

Lessons learned:

1. Do not use wood
2. Use a cheap camera

]]>h32009-07-26T11:44:10+01:000http://www.postronic.org/h3/pid53.html#commentshttp://www.postronic.org/h3/pid52.htmlhttp://www.postronic.org/h3/pid52.html

]]>h32009-07-21T20:36:23+01:004http://www.postronic.org/h3/pid52.html#commentshttp://www.postronic.org/h3/pid51.htmlhttp://www.postronic.org/h3/pid51.htmlToday I played again with the Google App Engine (Java preview release) for a couple of hours and found another bug. Seems like the MemcacheService namespaces are completely ignored in production. Bug 1364 already reported. The JCache API I used last week was just an higher-level interface built on top of the MemcacheService. I found that using the low-level API directly is a better approach because it exposes some additional functionalities with no drawbacks. Atomic-increment of cached entries for example is an extremely useful feature in a clustered environment.
Seems like the MemcacheService namespaces are completely ignored in production.

Bug 1364 already reported.

The JCache API I used last week was just an higher-level interface built on top of the MemcacheService.
I found that using the low-level API directly is a better approach because it exposes some additional functionalities with no drawbacks.
Atomic-increment of cached entries for example is an extremely useful feature in a clustered environment.]]>h32009-04-19T21:51:32+01:000http://www.postronic.org/h3/pid51.html#commentshttp://www.postronic.org/h3/pid50.htmlhttp://www.postronic.org/h3/pid50.htmlIn these days of holiday I've been playing with the Java version of the Google App Engine (yes I have been lucky to have access to the "early look" version). In general I find it wonderful. I really like how the Eclipse plugin integrates with the App engine and GWT. The included services are implemented using standard APIs whenever possible, but you are free to use the underlying technologies in a more low-level way if needed. My first idea has been to implement a sort of "comet" service for a GWT application, and I must admit: it's not easy because of the many limitations: 1) You can't create threads. 2) You can't use continuations. 3) All threads (generated from http requests) have a maximum lifetime of 30 seconds. Anyway finally I found a tricky way to implement a sort of slow-polling using the Memcache service, but.. I found a bug: JCache entries' expiration time is not reset when entries are updated! I think this is quite a big issue. Bug report 1284 already submitted :) In general I find it wonderful. I really like how the Eclipse plugin integrates with the App engine and GWT.
The included services are implemented using standard APIs whenever possible, but you are free to use the underlying technologies in a more low-level way if needed.

My first idea has been to implement a sort of "comet" service for a GWT application, and I must admit: it's not easy because of the many limitations:
1) You can't create threads.
2) You can't use continuations.
3) All threads (generated from http requests) have a maximum lifetime of 30 seconds.

Anyway finally I found a tricky way to implement a sort of slow-polling using the Memcache service, but..
I found a bug: JCache entries' expiration time is not reset when entries are updated! I think this is quite a big issue.

Bug report 1284 already submitted :)
]]>h32009-04-13T20:42:32+01:000http://www.postronic.org/h3/pid50.html#commentshttp://www.postronic.org/h3/pid49.htmlhttp://www.postronic.org/h3/pid49.htmlIt feels like opening the cage of your beloved pet and staring at him while he walks into the wild, finally free like wild animals. It's a bit scary and you hope he is ready for this step, but you know that this is good for him, in the long run. Knack has been my pet project for some years and starting from today it is Free Software. Please contribute!It's a bit scary and you hope he is ready for this step, but you know that this is good for him, in the long run.
Knack has been my pet project for some years and starting from today it is Free Software.

Please contribute!]]>h32008-11-08T21:55:02+01:000http://www.postronic.org/h3/pid49.html#comments