| T O P I C R E V I E W |
| Kaiserludi |
Posted - Sep 12 2017 : 1:24:35 PM
Hi.
VA_X.dll file version 10.9.2114.0 built 2016.10.15
language: C++
Your documentation at https://docs.wholetomato.com/default.asp?W199 states the following:
"Commented Code
Visual Assist does not underline what it believes is commented code. This includes symbols defined in any context.
Words within comments that appear in PascalCase, camelCase or hungarian_notation are not underlined. Visual Assist assumes the words refer to symbols."
However apparently Visual Assist is way to generous in what it considers to be commented code, resulting in the waste majority of comments not getting spell-checked, because Visual Assist wrongly treats the mas commented code.
The spelling error in the following comment gets correctly detected by Visual Assist:
/* donn't*/
However the same error gets ignored in the following comment:
/* donn't Test::test */
As a result the spell-check is effectively deactivated for API reference documentation type of comments like Doxygen documentation, in which the doc of one symbol frequently refers to the doc of another symbol by mentioning that symbol by its fully or partly qualified name (not just the variable or function name, but including the class and namespaces where necessary).
Unfortunately these kinds of comments are by far the ones for which the spell-checking is the most important, as those comments tend to be long and as their texts (and spelling errors within it) will be shown to any user who is reading the API documentation. |
| 4 L A T E S T R E P L I E S (Newest First) |
| accord |
Posted - Sep 19 2017 : 07:02:30 AM
For now, you can use manual spellcheck (VAssistX | Tools | Spell check) as a workaround. |
| accord |
Posted - Sep 18 2017 : 11:03:19 AM
I've put in a feature request to add option to disable spell checker's smart code detection in comments:
case=111146
Thank you for the feedback. |
| Kaiserludi |
Posted - Sep 14 2017 : 09:51:49 AM
Hi accord.
I agree that deciding if a comment is a "real" comment or just commented out code is often far from trivial.
So actually I was about to suggest the same after reading the original part of your post that you then afterwards already suggested yourself in the edit:
It would be great if one could switch this feature on and off.
That way one could even on default leave it the same as it is now and switch it off temporarily when working on comments / in-code documentation. Once one has fixed the actual spelling mistakes one could switch the option back on to silence any false positives that might get wrongly marked as spelling errors during off-mode.
The real inconvenience by the way is that one symbol name is enough to turn off spell-checking for the rest of the line, which can mean for multiple sentences and not for just 2-3 words per symbol name.
When writing or proof-reading comments for doc generation that may be 10-50 lines and a couple hundred words long while containing 5-20 symbol names, I think seeing 10 or so false positives is preferable over having the spell check turned of for 400 out of 500 of the words that are not symbol names, especially if after one is done fixing the actual spelling errors one can just get rid of the false-positives by switching the check back on. |
| accord |
Posted - Sep 13 2017 : 1:52:43 PM
Unfortunately it's not always clear what people mean to be a definition and what they mean to be a code. A mistyped word in your example can be the type or a returning type. It's "don't" in your case which is not a valid symbol name because of the special character, but in other cases, it can be difficult to tell the difference, especially since commented code can be incomplete.
The only "real" solution I can think of right now is introducing a switch to not check whether the comment is a code? So everything would be underlined that's you would like to. Since symbol names defined outside comments are already recognised by the spell checker, thus not underlined, maybe the result wouldn't be that bad? |
|
|
