Multiline comment: what is considered the comment content?

[genius257]

So long story short: i am writing a AutoIt3 parser and want some feedback from other AutoIt3 users on this.

The following seems to be a valid multi-line comment block:

#csA
B
#ceC

What would you consider the content of the comment block?

1:

A
B
C

2:

A
B

3:

B

Or maybe something else?

Edited April 1, 2023 by genius257

[Danp2]

I would assume that it would behave the same as #region, which treats any text after the keyword as a comment --

Quote

There is an additional folding keyword #region...#endregion to permit really large blocks of code to be compressed.  Any text following these special keywords is considered a comment:

 

[pixelsearch]

I always use the following syntax to make it clearer :

#cs
1st line of comment
2nd line
...
#ce

e.g. nothing written on the #cs and #ce lines, to get a neat display.
Just my 2cts...

[SOLVE-SMART]

I personally write the first comment line in the second line of the multiline comment (like @pixelsearch). But what I expect is the comment content AB, like I voted above. This would match several other programming language implementations of multiline comments. Each charater between the keywords/signs is a part of the comment content.

Thanks @genius257 for your work 👍 .

Best regards
Sven

Edited April 1, 2023 by SOLVE-SMART

[OJBakker]

I usually keep the directives on separate lines.

My preference would be to consider the example as invalid code.

#csA

and 

#ceC

should be reported as invalid directives.

using comment on directive lines can be done with existing syntax without ambiguity.

#cs ;A
B
#ce ;C

or implicit comment (not my preference)

#cs A
B
#ce C

I have voted for AB because that seems to be the current implementation but my preference would be it would be just  B.

I prefer to see comment on a directives line not as part of the comment. It has to major functions

1: remark why did I comment this code

2: add a marker to both begin comment and end comment to pair both, especially used with nested comments or large blocks of code.

 

[genius257]

Hi all Thanks for the answers so far!

So i think i will go with the AB solution.

 

I agree that most if not all people will use multi-line comments, where the text content always starts at line two in the block (me included).

It does make the most sense.

But from a parser perspective, all content after the #cs, and before the #ce that works in the native AutoIt3 executable should be included.

 

16 hours ago, OJBakker said:

My preference would be to consider the example as invalid code.

I can see why you think that, bout my current parser needs to be a one to one implementation of the parser rules AutoIt version 3.3.14.5 uses.

I apologize that i did not make clear what my current target AutoIt version was.

I do plan to make parsers for future AutoIt3 versions also, but choose 3.3.14.5 as my starting point. If they "patch" the multi-line syntax in a later version, that specific parser implementation will of course reflect that

 

I thank you all! You have saved me days of time wasted on this "simple" implementation choice

[Jos]

Just a little off-topic remark and only FYI: I see that there is an issue in the current au3lexer and it doesn't do correct syntax highlighting for these:

On 4/1/2023 at 10:48 AM, OJBakker said:

My preference would be to consider the example as invalid code.

#csA

and 

#ceC

should be reported as invalid directives.

The lexer should not recognize them as Start/Stop of commentblocks and should simply be colored as some PREPROCESSOR statement.  Will be fixed.

Edited April 2, 2023 by Jos