11-19-2022, 11:06 PM
Cycle 3, stage 4, day 19:
I have been working the whole week with excellent productivity. One of the thing that I have been working on is upgrading my logging subsystem with a new C++20 feature that I have heard about by pure luck. I did stumble on the idea when I did watch a video presenting the feature.
Converting my whole codebase to use the new system is a huge time commitment. I started by making a small proof of concept and it has been conclusive. With the new feature in place I have seen my best timing ever recorded....
The objective is to eventually migrate totally to the new system but there is no urgency to reach that goal. The proof of concept did migrate the most critical performance wise part. I just want to migrate the whole system for the sake of consistency but the low hanging benefit has already been collected. I guess that I could refactor 1 module per day, I could be done in about 1 month or so...
I have not updated my server for at least a week. I am overdue for an upgrade but before doing so, I wanted address at least 2 pending issues that would make an upgrade worthwhile...
My latest discovery just kept me busy the whole week. I did work on putting in place the proof of concept from Monday to Wednesday... On deployment, as soon as I have found the good result I was getting with the new system, I wanted to migrate more of my code to use the new system. Initially, I was thinking that a single day would be enough but as it happens very often, I did seriously underestimate the required effort... By Friday, I must have considered to aim for an incomplete conversion goal to not end up being paralyzed working on something that has not that much ROI now the most critical part has been taken care of. This morning I did fallback to the previous tasks that I was working on before being stroke by my latest idea which includes doing the necessary required for the new upgrade.
As I was returning on those tasks, I have been interrupted by a bad result from the latest trade my system did. I am glad that I did as it did allow me to discovre a new bug! This bug was not a software design bug, it was a mathematical problem. There is something that I was computing wrongly and this was resulting into bad trading. I have been extremely lucky to stumble into it today. This issue has been in the system for a very long time without never been noticed before.
In the next upgrade a fix will be included. The next upgrade is going to be something. It will include all the work that I have put into the code in the last week and there is a lot...
I have been working the whole week with excellent productivity. One of the thing that I have been working on is upgrading my logging subsystem with a new C++20 feature that I have heard about by pure luck. I did stumble on the idea when I did watch a video presenting the feature.
Converting my whole codebase to use the new system is a huge time commitment. I started by making a small proof of concept and it has been conclusive. With the new feature in place I have seen my best timing ever recorded....
The objective is to eventually migrate totally to the new system but there is no urgency to reach that goal. The proof of concept did migrate the most critical performance wise part. I just want to migrate the whole system for the sake of consistency but the low hanging benefit has already been collected. I guess that I could refactor 1 module per day, I could be done in about 1 month or so...
I have not updated my server for at least a week. I am overdue for an upgrade but before doing so, I wanted address at least 2 pending issues that would make an upgrade worthwhile...
My latest discovery just kept me busy the whole week. I did work on putting in place the proof of concept from Monday to Wednesday... On deployment, as soon as I have found the good result I was getting with the new system, I wanted to migrate more of my code to use the new system. Initially, I was thinking that a single day would be enough but as it happens very often, I did seriously underestimate the required effort... By Friday, I must have considered to aim for an incomplete conversion goal to not end up being paralyzed working on something that has not that much ROI now the most critical part has been taken care of. This morning I did fallback to the previous tasks that I was working on before being stroke by my latest idea which includes doing the necessary required for the new upgrade.
As I was returning on those tasks, I have been interrupted by a bad result from the latest trade my system did. I am glad that I did as it did allow me to discovre a new bug! This bug was not a software design bug, it was a mathematical problem. There is something that I was computing wrongly and this was resulting into bad trading. I have been extremely lucky to stumble into it today. This issue has been in the system for a very long time without never been noticed before.
In the next upgrade a fix will be included. The next upgrade is going to be something. It will include all the work that I have put into the code in the last week and there is a lot...
