Live Expressions Uses Wrong Variable Address

[Kier]

Am using 1.16.0.

This picture tells the whole story:

Am I doing something wrong or is Live Expressions broken?

[Kraal]

Hello,

I am suspecting an alignment request for live expressions. But after reading that https://community.st.com/t5/stm32cubeide-mcus/live-expressions-showing-unexpected-value/td-p/119158 I am not so sure anymore.

Best regards.

[Kier]

@Rim LANDOLSI Please could you comment on this problem? Is it under investigation?

[Rim LANDOLSI]

Hello @Kier and welcome to the community,

 

I tried to reproduce the problem with a different variable type, but I couldn't reproduce it. Here is an example in the screenshot.

To further investigate the issue, could you share the variable type definition you used for this project?

Thanks,

Rim.

[Kraal]

@Rim LANDOLSIin your example the address of system_type is aligned on 32-bit boundaries.

We don't know what is the size of system_type in the original post, but since it was put at an odd address by the linker, it should be 8 bits. @Kier can you give us more info on the variable in use ? Or better yet, make a small project that reproduce the issue ?

 

Best regards,

Kraal

[Kier]

Thanks all.

The variable type in question was an enum:

typedef enum
{
 SYS_TYPE_TP1,
 SYS_TYPE_TP2,
 SYS_TYPE_MAX_NUM,
 SYS_TYPE_UNDEF = 0x7F,
} SystemType_t;

I'm assuming short enums are used so the underlying type is a char.

SystemType_t system_type = SYS_TYPE_UNDEF;

Unfortunately, I can't recreate the problem now. The address (and data) in the same in both Variables and Live Expressions:

 

[unsigned_char_array]

Try deleting all the live expressions and add them again when you are on a breakpoint where the expression variables are in scope. I've had issues with adding a second expression or adding one of a variable out of scope.