The library doesn't read from the correct location in EEPROM

[amowry]

Subject of the issue

With v, 3.0.0, the library seems to read from the wrong location in EEPROM.

Your workbench

I'm using a custom NRF52840 board with the Adafruit ItsyBitsy NRF52840 bootloader and board package. The EEPROM is an ST M24128-BFMH6TG.

The previous of the library (2.0.1) works fine with this setup.

Steps to reproduce

Using the attached code and library v. 2.01, the value "10" is read back from address 1 in EEPROM, as would be expected.
However, with v. 3.0.0, a value of "255" is returned, indicating that the library is reading from some other address than 1.

Sparkfun_external_EEPROM_library_test.txt

[toogooda]

Yes please roll back V3 it is very broken waist so much time then tried previous library and all is working just get 255 for reads

[nseidle]

Agreed. I believe I've rolled back to a stable v1.0.14. Please check for me.

[amowry]

v1.0.14 seems to work well for me. Thanks for your help!

[nseidle]

Glad to hear it's working for you. Sorry for introducing a bad version!

[amowry]

No problem at all-- thanks for your support!

[pauljeffress]

Hi @nseidle I'm using PlatformIO and it seemed to give me v3.0.0 and not the rolled back v1.0.14, until I manually forced it to. I'm not sure if that is a you or a PlatformIO issue?

[nseidle]

Sorry, I don't know how to down-force PlatformIO.

[amowry]

I've actually noticed the same thing on Arduino IDE-- it still says that 3.0.0 is the most recent version. Should that be showing 3.1 as an option? I've manually selected 1.0.14 and haven't been letting it update.

[pauljeffress]

@nseidle @amowry yeah I think perhaps the only solution is to create v3.1 with the known working code from 1.0.14. Because as far as I know both Arduino IDE and PlatformIO just look at GitHub, so if anyone installs for the first time or updates the library today, they will get the broken v3.0.

[nseidle]

A fine hack to the problem. New v3.1.0 released. Please let me know if you see it listed correctly.

[amowry]

The 3.1.0 update appears in Arduino IDE now and seems to work fine. Thanks!

[amowry]

Both 3.1.1 and 3.1.2 seemed to break this again-- it's reading all 255 values from the EEPROM.

[nseidle]

Sorry for breaking things again, again.

I've pushed v3.2.0. Please have a test.

I had a 'ah ha' moment, let's see if it's good...

myMem.setMemoryType(128); // Valid types: 0, 1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048

I've added a setMemoryType() to make initialization (no auto-detect) more straight forward.

I've tested this against the 24xx128 an everything works as expected in Ex7. Please let me know how to replicate your setup.

[pauljeffress]

@nseidle 3.2.0 is failing for me, I rolled back to 3.1.0 and its was all good.
I have a AT24C32 (32kx8bit i.e. 4096 Bytes), with addressing set to 0x57, not the default 0x50.
I'm using a SAMD51 MCU.
I have a slight variation of your test code that I am using for the test with diff library versions.
Note, the memory size that's detected is always wrong, I think it always has been.
Results with 3.1.0

17:50:20.424 -> Qwiic EEPROM example
17:50:20.424 -> Memory detected!
17:50:20.424 -> Mem size in bytes: 65536
17:50:20.424 -> I read: 200
17:50:20.424 -> I read: -366
17:50:20.424 -> I read: -7.35

Results with 3.2.0

17:38:17.950 -> Qwiic EEPROM example
17:38:17.950 -> Memory detected!
17:38:17.950 -> Mem size in bytes: 65536
17:38:17.950 -> I read: 0
17:38:17.950 -> I read: 0
17:38:17.950 -> I read: 0.00

Results with 1.0.14

17:52:44.395 -> Qwiic EEPROM example
17:52:44.395 -> Memory detected!
17:52:44.395 -> Mem size in bytes: 65536
17:52:44.395 -> I read: 200
17:52:44.395 -> I read: -366
17:52:44.395 -> I read: -7.35