Bheja Fry!

Last week, me and my team were working for a product's demo for our customer. In the mid of the week, we almost hit our target. But not yet there fully. I had an optimization change to the project, which theoratically will fill that target performance gap. I integrated the changes and made the code work without errors. The result was a big surprise, the performance were in same level as before. Hell! I couldn't crack the puzzle, the code looks same as in my standalone testing system. But the integrated product still run at the level below target.

Helluva lot of tests done, I roped in my team lead for a review. Went back to my testing system, searched for any loopholes there. Nothing was leading us to the end of the tunnel. I was sitting with irritation for the whole day.....bad, bad day ofcourse. I decided to scrap out the changes and planned to do it afresh on saturday morning and left for the day. On saturday morning, I entered my cube with fresh mind and motivation to nail down the error. I started reverting back my changes, and came to the project's makefile. Something caught up in eyes, it was a macro difference, literally MACRO difference. I was using a macro to configure my files and that macro wasn't consumed by the compiler at all! Alas! Silly little integration mistake..........

After nailing that stupid error, the demo crossed the target comfortably. OMG! this isn't the best experience to trace down those errors. Thought I caught up the tail of tiger that leads to big issue somewhere, but definitely an ironical end. Now we are laughing at this badly here!

Read more!