fix: make read() backward compatible again

[avinxshKD]

read() now returns plain data again, and read_with_status() is the explicit tuple API; tests were switched to the explicit path where status is asserted.
Needed because the tuple return broke existing Python studies and templates.

Fixes #547

[avinxshKD]

@pradeeban pls have a look

[pradeeban]

Merging. I like your code, PR title, and captions. Very systematic with no bloat.

[avinxshKD]

@pradeeban Thank you so much, (one of my fav compliments :p)

btw what’s your take on atomic writes + integrity framing for the Python/C++/Java runtimes? That’s a core transport gap.

[pradeeban]

Are you asking in the context of concore? I thought we already have that support for at least python (file-sharing, local), Verilog, C++, and Mediator versions, as in all-or-nothing.

[avinxshKD]

Hey @pradeeban I meant file‑based runtimes. I don’t see atomic/verified writes in the current Python/C++/Java code, they write straight to the target file. If I’m missing on something, could you point me to it? I’ll drop this if it already exists.

[pradeeban]

ok, I was referring to atomic in a holistic sense. The entire transaction across the two processes succeeds, or neither of them succeeds (this is a critical aspect of concore protocol. After all, you don't want the transaction to be in an unstable state). The file-sharing and mediator-based approaches compare the files and achieve this (probably in an ancient way). Not as in atomic/verified writes in the programming constructs.

Maybe we can come back to this again. We don't want to overdo either (because, performance. Every program in a control system can run hundreds of times).

[avinxshKD]

@pradeeban Got it, thanks for clarifying. I’ll hold the atomic/verified write idea for now.
If we revisit it, I can sketch a low‑overhead option that doesn’t hit perf.

Also If there’s any doc or note on the “holistic atomic” behavior, I’d really appreciate a link when you have a chance.

[pradeeban]

The IEEE Access paper is the best bet on concore protocol.

https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9739711

See if it adds anything.