Replace UrlEncode library with local implementation.

[robertlipe]

Description

Removed UrlEncode third-party dependency.

• Add include/UrlEncode.h proto and new (better) body in src/network.cpp.
  Uses RFC 1738 style percent-encoding.
• Guarded the urlEncode function and its usage in src/deviceconfig.cpp with #if ENABLE_WIFI. It may be possible to do better later, but we agree for now that without WIFI, there is no URL.
• Removed unused includes of UrlEncode.h from PatternStocks.h and PatternSubscribers.h.

Contributing requirements

• I read the contribution guidelines in CONTRIBUTING.md.
• I understand the BlinkenPerBit metric, and maximized it in this PR.
• I selected main as the target branch.
• All code herein is subjected to the license terms in COPYING.txt.

[rbergen]

This may elicit a "yeah, right" from you, but I saw this one coming when we were discussing #788. :)

I have a couple of comments. I expect their phrasing will show I feel (much) more strongly about one than the other.

[rbergen]

@robertlipe Just quickly checking if you're planning to get back to this?

[robertlipe]

Yes. Distracted with other projects... Some of which are ND, but in different trees. The rationale was to minimize and localize the change. That's why I kept/made the header. We have a ton of audio/networking code that isn't "correctly" conditionalized so I didn't go on a recursive chase for that. I minimized the blast radius. But I understood that's what it takes, so I'll come back to this. As for the "yeah, right", thing, I probably lead you to that. My distaste of the platformio approach is not exactly a secret. I have another, more extreme version of this that brings in ALL (?) the libs (and has a python thingy that checks them out/updates them). That lets us actually FIX some of the abandoned and misguided code we're depending on and reduces our dependencies on Platformio's buggy, slow build system, which is two things that I'm not a fan of in any software. It makes builds less slow because we're not waiting for 200+network fetches, making 50 copies of everything, and filling our disks with redundant copies. Trying to get patches accepted upstream into abandoned projects just isn't working and I really don't want to become a general purpose fork of those packages; I just want them integrated sensibly. This one was just easier to cold replace. Im testing the waters. 😎 But, like so many things in my ND world right now, it's piled up behind other things. The Arduino3/P4 thing may depend on it... I actually took a break from that very recently and developed two new fun effects from scratch. I have a few more in the idea incubator, too. I also have a note at home that there was something else I recently half-proposed that you'd pre-approved in principle, so it's on my near-term list, too. Oh, and some doc. I've not forgotten about the project. Far from it.

…

On Sat, Jan 10, 2026, 3:00 PM Rutger van Bergen ***@***.***> wrote: *rbergen* left a comment (PlummersSoftwareLLC/NightDriverStrip#811) <#811 (comment)> @robertlipe <https://github.com/robertlipe> Just quickly checking if you're planning to get back to this? — Reply to this email directly, view it on GitHub <#811 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACCSD3Y4KS7DZBB6IG6KTPD4GFR5VAVCNFSM6AAAAACQWIG6NOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTOMZTGU2DOOBWGU> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[rbergen]

No problem, I was just checking in, in the context of regular PR housekeeping.

The rationale was to minimize and localize the change. That's why I
kept/made the header. We have a ton of audio/networking code that isn't
"correctly" conditionalized so I didn't go on a recursive chase for that. I
minimized the blast radius.

But I understood that's what it takes, so I'll come back to this.

Well actually, I understand the header, or a header separate from any others we have. It's networking-related code in our thinking, but pulling in network.h wherever URL encoding may be need feels like overkill. Maybe a utilities.h (with accompanying .cpp) would make more "generic" sense in our specific context, particularly considering your ideas about a possible future for other dependencies.

In any case, the other thing (the current scope of #if ENABLE_WIFI) I feel far more strongly about.

I also have a note at home that there was something else I recently
half-proposed that you'd pre-approved in principle, so it's on my near-term
list, too. Oh, and some doc.

Haha, I always love it when people make vague references to things I've pre-approved in principle. ;)

I've not forgotten about the project. Far from it.

I didn't think that for a second. My question was strictly scoped to this PR.

[rbergen]

One minor case of PR creep, which I've noted for the record. LGTM otherwise.

[rbergen]

I'll pretend I didn't see this feature being snuck in with the UrlEncode refactor. ;)

[robertlipe]

Thank you for the integration.

Moving uptime, unchanged, from a file that was networking (where it used to live because debug stuff was entangled with network stuff) to a file that is home for debugging was not a new feature and was not sneaky. It was a very deliberate change as part of making ENABLE_WIFI (which clearly otherwise applies to networking.cc) honest, as it required that method to be rehomed so it could still be available even in builds without ENABLE_WIFI, as this is useful over the console port.

I landed DoUptime back in 2023 in #402 because when running long times for leaks or stability testing, I want to see that it's been running a long time and didn't crash or reboot minutes before I walked back into the room. Even then we noted it was an awkward fit

But agreed that this PR morphed from "don't download and install a 20 line function 50 times" to make ENABLE_WIFI (and related networking #defines) actually bracket when they claim to bracket. That was me taking your request to its conclusion.

[rbergen]

...which I appreciate. In any case, the comment was mainly about the addition of the "uptime" command to the CLI command set. To which you may say "are you serious?", to which I would reply "no, that's why I approved and merged."

I think the PR adds a number of sanity checks/extra safeguards to a complicated area of the codebase (anything networking), which is unequivocally a good thing to do. So, thank you for it!

[robertlipe]

But uptime wasn't added. It was added in 2023. It was moved from one file to another.

Are we talking past each other?

But, yeah, this definitely ballooned, but leaves it better than we found it.