Feature request: CamelCase spell check in code

[janschreiber]

Common behavior for spell checkers that are intended for coding is to turn off checking for everything except comments and strings.
But in quite a lot of software projects, function and variable names consist of plain ordinary English words, often in camel case, such as OnEventReceive( trigger ). Why can't we have those spell-checked, too, only in a novel way?
The idea: If a word is inside a code area, just split it up at the word boundaries marked by the case changes before sending the chunks to the underlying spell-checker (Hunspell or Aspell). In our example, you would send "on", "event", and "received", so, no problem.
If they return for any of the chunks that it is misspelled, mark the function or variable name as misspelled and generate suggestions based on concatenating the suggestions.
For example, if I mistype "OnEventReceive" as "OnEventRecieve", I would expect the following to happen: Aspell suggests "receive" as replacement for "recieve", while not complaining about the other two chunks. As a result, DSpellCheck suggests "OnEventReceive".
I think that could save quite a bit of hassle, in particular when working with a multi-national team.
Obviously, it must be made optional though, because for some projects it won't make sense at all.

[Predelnik]

Well, that's pretty interesting coincidence but I've stared to use intellij IDEA (IDE for java) a few weeks ago and it's spell-checker actually does exactly what you said, separating words by CamelCase and _ in variable and methods names and checking their spelling, I've been thinking that it actually could be possibly be done in my plugin also and not very hard to program. Well the only drawback I see is that since intellij IDEA is very powerful program with refactoring tools so when you are renaming one variable or method inside it, it renames all of it's usages in code while N++ can't do it (maybe it'd be good to add possibility to replace misspelled word in all file though, can't be sure :) ), so the users will face some errors if they forget to rename variable in some place, but even though it's not such a big deal I guess)

Actually I think checking CamelCased words and word separated with _ have a sense even in comments, because people often write variable names in comments.

[janschreiber]

Actually I think checking CamelCased words and word separated with _ have a sense even in comments, because people often write variable names in comments.

That is true. OTH, we must make sure we aren't going overboard here: When writing plain text (as opposed to code) in German, what looks like CamelCase is often a sign of a forgotten space or hyphen, because in German all Nouns start with a capital letter. Something like "HansMueller" in an e-mail is probably a typo. In non-code files, it is probably best to disable CamelCasing by default, but make it an option, because it can be useful for Software documentation and such.

intellij IDEA is very powerful program with refactoring tools so when you are renaming one variable or method inside it, it renames all of it's usages in code while N++ can't do it

Again, this is true. If you open an existing project with a misspelled function name already in use in several files all over the place, it's a bit of a pain. But N++ comes with very powerful find and replace functionalities that will hopefully save the day. And it is always an option for the user to leave everything as it is, if that happens to be the most economic solution.
The use case I was thinking of is something along these lines: The user defines a function or initializes a variable in the first place and is warned about the possible typo before the virus spreads, so to speak. I think it would save a lot of annoyances having to remember the weird spelling of certain variables, in particular in multi-national teams.
(I do the German localization for N++, I know what I'm talking about.) ;-)

[RenniePet]

First off, thank you very much for DSpellCheck - I love it and use it all the time.

Second, this is an old thread, but I would like to cast a vote for implementing this feature, which is relevant for both "camel case" and "Pascal case" identifiers or compound words.

For example, I just noticed that I made a spelling mistake in a document some three years ago - how embarrassing:

1. Write-to-disk support (FileWritter class)

It should, of course, be FileWriter class.

I'm sure Notepad++ is used a lot for documenting programs written in C# and Java and lots of other programming languages where "compound words" are used as identifiers.

Thank you.

[Ben-Voris]

I'd really like to be able to tell DSpellCheck that (for example), argleBargle is a valid word with only that capitalization. This is common with technical documentation where the match must be exact. That is, I should be able to make DSpellCheck valid but have DspellCheck, dspellcheck, dspellCheck, etc. flagged as a misspelling.
This seems related to hunspell/hunspell#468 and #89

[Predelnik]

Yes and like I mentioned in #89 tokenization is completely on side of the plugin so it is not directly related to hunspell issue. It's just a possible improvement on plugin's side.

[Predelnik]

CamelCase word tokenization is implemented in 1.4.0 which is avaialble through plugin manager.
It it currently disabled by default and have to be enabled using this check box:

I noticed that it breaks spell-checking of Scottish surnames if enabled by default but I'm not sure if it's that important, maybe I will make it turned on by default later on.