I am encountering an inconsistency when calling my_address() in different contexts and hope someone can help me understand why this is happening.

When I call my_address() within get function, it returns a slice that starts with 0 bits, as shown below:

CS{Cell{004380008d090b360188f0560ee4010388f181ad345579022d85d84b0ac904ead5acb410} bits: 0..267; refs: 0..0}

Conversely, when I call my_address() within an impure function, it returns a slice that starts with 1 bit:

CS{Cell{026fc0004684859b00c4782b07720081c478c0d69a2abc8116c2ec25856482756ad65a0254910843292a22680000000001ab3f0500bd26da1340} bits: 1..268; refs: 0..0}

Despite these differences, when I parse both addresses with parse_std_addr, they return the same workchain and 256-bit integer address:

workchain = 0
integer address = 1993502832782254581058910437300945815969310326214648768594806511355848123808

Therefore, when calculating a child contract address using my_address, it yields different results depending on the context in which it's called. Can anyone explain why this is happening and how I should handle it?