Error in nestest log, or with my logic?

I have started working on my own NES emulator, and after finishing most, if not all, of my CPU core I have started testing it with the nestest but I found a weird error at the following instructions.



Upon further inspection I see that my emulator have the correct values: A = 0, $01 = 0xFF and P = 0x2E. According to the nestest log P = 0xEE after executing BIT, but how is that possible when A = 0 and $01 = 0xFF? 0x00 & 0xFF = 0x00, which means that both the N and V flags should be false at this point. Following the same logic, I would assume that P = 0x2E after executing BIT.

Am I missing something, or is there some error with the nestest log?

The N and V flags are set according the memory value, prior to the AND operation. Reference.

With BIT only the Z flag is from the AND result. N and V are from the operand's fetched value, not the result of the AND.

Edit: Rahsennor already covered it. Sorry for duplicate answer.

Thank you very much. That fixed it. I have been following this resource and it did not mention that at all.

It's good to use multiple references.

This one is a decent high-level overview: http://www.obelisk.me.uk/6502/reference.html
This one details the cycle-by-cycle behavior (there are a few errors though, IIRC): http://nesdev.com/6502_cpu.txt

Do you mind to mention some of these errors? I used such doc to write my 6502 emulator.

I don't think there were any big errors that I found, more like mistakes/typos, like this one for BRK:

...and what's "wrong" with this?

PCH is the hibyte of PC, it doesn't have a "B flag". The note about B flag was supposed to be on another line.

Another error I happened to notice:


(NMI does set the I flag.)