This week (June 2–8) marked a big shift — not just in terms of clock signals and data bits (hello shift registers!), but also a personal shift from the Community Bonding Period to the real deal: the Coding Phase of GSoC.
The coding period kicked off and I dove straight into implementing shift registers — SISO, SIPO, PISO, PIPO. Classic stuff. The first question was: do I create separate components for each? Or just one smart, configurable one?
Took a inspiration from Logisim, discussed with mentor and went with the second one. Just one “Shifter” component with properties like parallelLoad: yes/no. Cleaner UI, less clutter, and feels more user-friendly.
Then came the design phase. I built the component UI using the Canvas API. Control signals (Reset, Clock, Shift/Load) go on the top, and inputs/outputs on the bottom. As you increase the number of stages, the bottom part grows — dynamic and neat.
Initially had it in vertical layout, but then switched to horizontal because it just makes more sense when visualizing right shifts.
With the UI structure done, I started implementing the core logic in the resolve() method. This method detects rising clock edges and handles:
- Resetting the register
- Parallel loading the input data
- Shifting values serially to the right
Everything seemed fine — until it came....
🐛 The Unexpected Issue
Strangely, I noticed that the resolve() method wasn’t being called at all during simulation. I checked the logic, rewrote parts of it, even added logs — nothing. Eventually, I rolled back and tried running the same code outside the versioned folder (v0), and it worked perfectly.
That gave me a strong hint: the issue was somehow tied to the versioning structure. I’m still not sure why this happens, so I decided to take it up with my mentor Aman Asrani. Meanwhile, to keep the flow going, I completed the implementation without using the versioned folder.
Towards the end of the week, I tested edge cases, tweaked a few last-minute design elements (like defaulting to horizontal layout), and handed over the logic for review. I’m now waiting on the versioning issue to be resolved, after which I’ll raise a proper PR for this new component.
Looking forward to Week 2 — and hopefully resolving the versioning quirk! 🚀
Top comments (0)