fix: Race condition fix for issue #257

[milesburton]

feat(license): migrate to MIT license

• Replace GNU Lesser General Public License with MIT License in both header and implementation files
• Update copyright year to 2024
• Add full MIT license text and conditions

feat(error-handling): add retry mechanism for temperature readings

• Add retryCount parameter to getTemp() with default value of 0
• Add retryCount parameter to getTempC() with default value of 0
• Implement retry logic in temperature reading functions

fix(device-search): improve address retrieval reliability

• Add index validation in getAddress() method
• Enhance error handling for device search
• Improve efficiency of device enumeration

docs(api): update function documentation

• Add documentation for new retry parameters
• Update method signatures in header file
• Clarify temperature reading behavior

test(temperature): enhance temperature reading reliability

• Add retry mechanism to handle intermittent sensor failures
• Implement progressive retry logic
• Add timeout protection for device communication

refactor(device-handling): improve device management

• Optimize device count verification
• Enhance device address validation
• Improve error handling in device communication

[RobTillaart]

@milesburton

minor note:
Add retryCount parameter to getTempF() with default value of 0 Fahrenheit too, to be consistent.

[milesburton]

@milesburton

minor note: Add retryCount parameter to getTempF() with default value of 0 Fahrenheit too, to be consistent.

Thanks Rob, let me make that change

[milesburton]

@RobTillaart @tdwilson Can you guys sanity check this?

[RobTillaart]

@milesburton

Probably later this week, I have some catching up to with some other projects.

[milesburton]

Hey @RobTillaart, thanks for your getting back to me. I’ve been digging through my project cabinets but, unfortunately, I can’t find the test rig I built for exactly this purpose. Tim also got back to me, although it may be a few weeks before he can do any testing on his end.

In the meantime, I’m considering building a small test rig that would automatically sanity-test changes like these. Given all the Arduino libraries you’ve worked on, do you happen to have a similar setup already? I’d appreciate any insight or suggestions you might have.

[RobTillaart]

@milesburton

The .github/workflows has three actions defined

• Arduino-lint - checks the library.properties file and checks the examples, this is also used by the library manager of Arduino.
• Json-check - not used but could check the library.json file used by platformIO
• Arduino-CI - does two things (1) runs unit test and (2) compiles all examples

The detailed output of these actions can be seen here (and as green ticks e.g. in a PR)

Note:
The unit tests are disabled as there is a problem with finding CRC from onewire. See test folder
The compilation of the examples works, the last successful one was 3 days ago.
See - https://github.com/milesburton/Arduino-Temperature-Control-Library/actions/runs/12612151488/job/35148582797

The boards used for compilation are defined in the file .arduino-ci.yml (note the . at the begin)
I have some additional code in my libraries this file to also compile for the rpipico

Compiling the examples at least learns that the code of the library is syntactically correct and compiles.
Given the broad range of the examples most code paths are touched.
Type mismatches and several type of bugs are caught.

Compilation only is however not a unit test.

---

Finally in my A1301 repo / develop branch (my playground) I have a first try of the platformIO test runner.
Again this is also compile only.

[RobTillaart]

@milesburton
Looks to me like the PR has a corrupted / very stripped version of the DallasTemperature.h file
40 lines versus 346 in the main branch.

.cpp file idem.

This will never be mergeable

[milesburton]

@RobTillaart It's been an interesting evening working through this! I tried mocking out the CRC, but that didn't quite go as planned. I also attempted to skip the tests, but unfortunately, that didn’t resolve the issue. I explored several options for installing the AVR library directly, but it seems we may need to take a different approach.

At this point, I think the best course of action is to "dog food" the change and directly assess whether it causes any issues. If you uncomment the tests in the dev container, they work as expected, so I believe we're in a good place.

If anyone with a working setup could test this out, it would help us move forward with merging. In the meantime, I’ll create a branch and continue troubleshooting tomorrow to ensure everything works correctly.