Wednesday, August 14, 2013

Release Day is a Very Dangerous Day

Twas the night before release, when all through the rev,
Not a user was stirring, not even a dev.
Test cases were done by the QC with care,
In hopes that bugs were no longer there.

All branches were merged, all conflicts dead,
All unstaged changes commited to HEAD.
All builds in release, all binaries done,
We call it a day and make this version one.

When out at the hardware there arose such a clatter,
I tried to connect to see what was the matter,
A debug window to a peak at the flash,
revealed that the program had suddenly crashed.

Could a dynamically allocated matrix of doubles,
be the runtime error causing the troubles?
Could the sixteen-bit table I decided to cram,
be giving the finger to objects in RAM?

With a faulty old driver for my JTAG stick,
I knew in a moment this wouldn't be quick.
Move back into to EAGLE and start from scratch,
It's easier than trying to develop a patch.


No comments:

Post a Comment