M318 and M351 checks

[kossnikita]

Message Number	Type	Message Text
M318	WARNING	'LEVEL' 'NAME' <'TAG'> is equal to <name>
M351	WARNING	Peripheral 'TYPE' 'NAME' is equal to Peripheral name.

Doesn't add if it's equal to name:
Peripheral:

• displayName
• description
• alternatePeripheral
• groupName
• headerStructName
  Register:
• displayName
• description
• alternateRegister

I can't say that I'm happy with this method. This may be unexpected for the user.
But this is a very simple way to remove displayName duplicate name in my case. Patching this manually seems terrible.

[burrbull]

I can't say that I'm happy with this method.

Silently remove tags is very bad idea.
And this is not a place to patch.

[kossnikita]

Can we do this as an option in the configuration, which can be activated by an argument in svd-tools?

[burrbull]

Are those displayNames present in initial SVD or produced by patches?

Can we do this as an option in the configuration, which can be activated by an argument in svd-tools?

svdtools patch or svdtools convert?

[kossnikita]

Are those displayNames present in initial SVD or produced by patches?

In the original SVD.

Good question. I think that this should only be in the patch, since this is a significant content change.

[burrbull]

If we are talking about svdtools patch and initially present displayNames it should be formed as rule. Similar to _clear_fields: https://github.com/stm32-rs/stm32-rs/blob/ea6c0312dfc79e74bf3cbbcb735a3c32328dcfb2/devices/stm32g071.yaml#L6

If it is incorrectly produced by patch, it is a bug.

[burrbull]

Actually I don't understand what sense of displayName and why it should be different from name.
Possibly it is something C-relared.

[kossnikita]

Similar to _clear_fields

Another undocumented feature... 😅
I will take care about it someday.

I think displayName should be used in debugger viewer for some cases. And it should be different because it's equal to name by default. This makes my svds much smaller. I can't imagine any others reasons. It's only warning in any case.

[kossnikita]

I'm not sure I understand you. Are you proposing to introduce a new feature for clearing displayName similar to _clear_fields? Or does such a feature already exist?

[burrbull]

New feature. For already present displayNames.

More investigation for other cases.