Checks for integer overflow in function tf.convert_to_tensor() ---->

[Sarvesh-Kesharwani]

i added a conditional check if input value to "tf.convert_to_tensor(a, tf.int32)" is not exceeding int32 limit, if dtype is int32.

closes issue #58994

[Sarvesh-Kesharwani]

guys please guide me further as this is my first PR.
i'll be grateful for your support. thankyou

[mihaimaruseac]

I think the overflow fixes have to be in C++, not in Python

[Sarvesh-Kesharwani]

hi, mihaimaruseac
thanx for replying, i was stuck findivg what i am doing wrong.
i guess i will be able to make some progress now.
thanx

[edwardyehuang]

I think the overflow fixes have to be in C++, not in Python

Agree. The example below is also dangerous.

import tensorflow as tf

a = 2147483647
a = tf.convert_to_tensor(a, tf.int32)

print(a) # tf.Tensor(2147483647, shape=(), dtype=int32)

b = a + tf.constant(1, tf.int32)

print(b) # tf.Tensor(-2147483648, shape=(), dtype=int32)

[gbaned]

@Sarvesh-Kesharwani Can you please resolve conflicts? Thank you!

[Sarvesh-Kesharwani]

@Sarvesh-Kesharwani Can you please resolve conflicts? Thank you!

sure

[mihaimaruseac]

This PR is now just adding a single empty line

[Sarvesh-Kesharwani]

This PR is now just adding a single empty line

I regret to inform you that I am currently unable to dedicate enough time to familiarize myself with the required code base for this PR. As a result, I will have to close this PR.