Entries Tagged 'GNU/Linux' ↓

Our own Kartik Mistry is on one of the foss.in/2007 promotional posters! Kartik has been contributing to FOSS for over 4 years now and is active contributor to Localization and Debian.

Congratulations Kartik, and keep contributing!

And there are many other friends in the posters: Gopal, Karunakar, Naba Kumar, Shreyas, Pradeepto, Tejas and Suparna. Good work guys! And good work Hari, on the wonderful posters!

Just saw this:

Google took its long-awaited plunge into the wireless world today, announcing that it is leading a broad industry alliance to transform mobile phones into powerful mobile computers that could accelerate the convergence of computing and communications.

Essentially:

• Google will not do hardware
• It will rather do the software that runs the phones
• The software is based on Linux and Java
• And is open source

This looks pretty interesting! Though I am interested in seeing what this software looks like!

Been trying out different embedded databases for the past few days. And am not sure which one is the best. The background is this:

We have a desktop application in C#, and are sending data from desktop to the PHP backend on the server. Since the data is going to be large, we want to send only the changed data to the server. That too in the most optimized manner. We are using CSV file as data storage and doing a diff on it to create the delta that we can send on the server. We are even zipping the delta to make sure it’s good on the bandwidth.

And here’s the problem:

We want to pre-process the delta. So we convert the unified diff to CSV again. Now we want to push it to a temporary table, and update a few fields. E.g. we want to mark all new rows (+ in diff) as “I” (inserted), removed rows (- in diff) as “D” (deleted) and updated rows (- and + both in diff) as “U” (updated). To do this, we load the data into a DataTable and perform some queries on it.

But there’s a catch. The .Net DataTable can’t perform updates on CSV files. So we need a real database.

That’s where our search for embedded databases began.

We want something free and something that does not restrict redistribution in a commercial application.

The options were limited. SQL Server Express, Firebird and SQLite (didn’t want Access/MDB).

Frankly, it’s not been a simple ride so far. We are considering only Firebird and SQLite (open source etc). Neither had enough ready samples / documentation that could solve the problem for us readily. Firebird was more difficult - with a lot of confusion about setting it up and opening the fdb file from a GUI database administrator application. SQLite was easier.

We have to do INSERT queries on the table, as we do not know how many fields will be there in the CSV file, and we couldn’t figure out an easy way to load a CSV file into a new table.

At this stage, I am not sure which one to go for. We need something that will work fast. I know SQLite has an option to load CSV files directly into a table through command line, but we don’t want to do a system call.

Any suggestions?

Was working with log files and wanted a quick way to empty them and recreate them. rm and touch would do it, but found something simpler:

As simple as that!

Wanted to deploy an application on a Windows server. The application had features that would work only if cron was setup to call a particular php page at fixed intervals. Doing this on Linux is a job of two minutes. But I did not know how to do cron on Windows. And this is something that I have not even bothered to find out for years!

Always told people that you should be on Linux if you want such features. Windows is not acceptable

And if the client persisted, we did a SQL server job setup or Windows Scheduled Task feature!

Today I thought otherwise. Thought I must find some cron replacement for Windows. Couple of searches later, I found pycron. It's a great free solution for setting up crons just the way you would do on Linux. Comes also with a GUI!

So next time you want to setup a cron on Windows, try out pycron!