Extend concurrent_009 to comment in between multiline condition

[Anselmo95]

Good morning,

I would like if possible to keep the indentation of comments in line with the multi line condition set by rule concurrent_009.

output <= '1' when input = "0000" or (input = "1111" and
                                       -- comment_one
                                       input2) = "0101" else
          sig_a or sig_b when input = "0001" and
                              input = "1001" else
          sig_c and sig_d when input = "0010" else
          '0';

At the moment vsg is reporting ERROR: a.vhd(2)comment_010 -- Indent level 0 on the example above.

[jeremiah-c-leary]

Evening @Anselmo95 ,

VSG makes a distinction between indenting and alignment. Where indenting is based on hierarchy and is a set number of spaces for each level. Alignment is the relative column locations between elements in the code. In your example, the output target is at an indent, while the rest of the statement is an alignment.

I had not considered applying alignment to comments within a statement, but it makes sense since the comment is relative to the code in the statement.

I do not believe this will be difficult, but I will need to investigate the implementation.

Thanks,

--Jeremy

[Anselmo95]

Evening @jeremiah-c-leary,

Thanks for the explanation about the difference between indentation and alignment!

I have been looking at the code to see if I can implement this myself but I'm progressing quite slowly. Anyway I'll open a PR if I manage to do it.

Have a nice day!

[jeremiah-c-leary]

Morning @Anselmo95 ,

I have been thinking about this and might have an idea on how to implement it. So the indentation is set during phase 4 in vsg/rule_list.py lines 147 to 148:

146             # Update indents before checking indent
147             if phase == 4:
148                 self.oVhdlFile.set_token_indent()

This function sets the indent attribute of the token using the set_indent method on the base token clase in vsg/parser.py lines 46 to 47:

46     def set_indent(self, iIndent):                                                                                                                                                                            47         self.indent = iIndent                                                                                                                                                                                 

So maybe what is missing is an alignment attribute on the token class. This attribute would be set during alignment rules. I believe with both the indent and alignment attributes we could properly align the comments.

Using your code example:

|<-indent->|<--alignment------------------------->|
            output <= '1' when input = "0000" or (input = "1111" and
                                                   -- comment_one
                                                   input2) = "0101" else
                      sig_a or sig_b when input = "0001" and
                                          input = "1001" else
                      sig_c and sig_d when input = "0010" else
                      '0';

This would allow for indents to allow smart tabs or spaces and the alignment to be only spaces.

The base rule vsg/rules/token_indent.py would need to be updated to check for the alignment attribute. This will probably be the majority of the effort.

The last step would be to update vsg/rules/multiline_conditional_alignment.py. I want to say you can just subtract the iFirstLine from iIndent somewhere in that the rule. This file is quite large though, but there are a lot of similarities in all the functions. Maybe we could just add a function before the sort on line 225 that set the alignment attribute of comments? Since another rule would actually check the indent and alignment of comments.

Regards,

--Jeremy