Friday, December 14, 2012

C buildpack for Heroku done!

In a previous blog post, I mentioned starting development on C buildpack for Heroku. It is done now!

https://github.com/atris/heroku-buildpack-C

Peace,

Atri

Tuesday, December 11, 2012

Autodoc Generator for PostgreSQL prototype available


The first prototype of my Autodoc Generator for PostgreSQL is available now.

Please note that as it is a prototype, it has very limited functionalities at the moment. I will be adding more functionalities soon.

I will be grateful for any feedback/comments.

The project link is:

https://github.com/atris/Autodoc-Generator

I will keep you all updated on the progress of the project.

Peace,

Atri

Tuesday, December 4, 2012

C buildpack for Heroku

This gets a bit personal.I love Heroku(most people who know me know that) but I have always found the lack of support of C a bit annoying.I recently voiced my concern over this on Heroku's official Facebook page.To my delight,they responded,saying that if I build a buildpack for that,they will promote it.So,this starts now then.

Let the hacking begin!

I will keep you posted.

Peace,

Atri

Saturday, December 1, 2012

My LinkedIn profile

My LinkedIn profile,updated recently.Please do check it out.

Atri Sharma's LinkedIn profile

Peace,
Atri

Autodoc generator for PostgreSQL

The software will generate a set of HTML files for a PostgreSQL database with inter linking.This shall allow users to conveniently see the overall structure of their database.The information will include tables,views,functions,schemas,sequences and more.

Later on we are also planning to interface ERD generation with our software.

I am enjoying doing the project like anything.I am learning a lot of SQL and how PostgreSQL maintains its data and metadata.It is so awesome.

I will also have to learn Python for the project.This should be interesting.

I will keep you all updated on the progress.

Peace,

Atri

Patch for backend speed up for OLAP style workloads for PostgreSQL

A patch that gives speed up upto 25% in the backend for workloads that involve a lot of data being touched by the same transaction.The patch tries to mitigate the cost of I/O in this case by trying to mitigate the I/O.It involves maintaining a cache that can be used for determining the commit status of the transaction that last touched a tuple in cases of workloads as described above.

The idea belongs to my mentor,Merlin Moncure.My work is a simpler implementation of the idea,simpler as compared to Baba(I call Merlin as Baba)'s implementation.

The patch is under review.Thanks a ton Amit Kapila and Hari Babu.

More on it soon.

Peace,

Atri