NPPM_GETCURRENTVIEW vs NPPM_GETCURRENTSCINTILLA

Coises

Does NPPM_GETCURRENTVIEW supersede NPPM_GETCURRENTSCINTILLA with a simpler API (the older message being retained for compatibility), or are there cases where these two might not yield the same result? If they are different, what is the distinction between them?

Looking at the code (here and here), they test different things, even though it seems like those two tests would be expected to yield the same result. Likewise, the descriptions don’t use exactly the same words (“Retrieves the current Scintilla view” vs “Retrieves the current used view”), though they sound like they mean the same thing.

rdipardo

@Coises said:

Does NPPM_GETCURRENTVIEW supersede NPPM_GETCURRENTSCINTILLA with a simpler API (the older message being retained for compatibility) ?

My impression is that you’ve discovered another example of what I’ve said elsewhere — namely, that the plugin API is not so much a coherent system as an append-only list of features accumulated over many years. From that perspective, the best answer you’re likely to get will come from the revision history.

Apart from cosmetic improvements to the documentation comments (--ignore-rev bb278b26f), it does appear that NPPM_GETCURRENTSCINTILLA came first, although exactly when has been lost in the version control system’s translation from subversion to git (note the leading ?s):

$ git blame -L/NPPM_GETCURRENTSCINTILLA/,+6 --ignore-rev bb278b26f -- PowerEditor/src/MISC/PluginsManager/Notepad_plus_msgs.h

?4dd3b257e (Don Ho         2009-04-24 23:34:47 +0000 55)        #define NPPM_GETCURRENTSCINTILLA (NPPMSG + 4)
?9cbd03c30 (LEONARDO SILVA 2022-02-23 13:51:58 -0300 56)        // BOOL NPPM_GETCURRENTSCINTILLA(0, int* iScintillaView)
?4dd3b257e (Don Ho         2009-04-24 23:34:47 +0000 57)        // Get current Scintilla view.
?a87e89ea9 (Don HO         2018-11-07 23:30:59 +0100 58)        // wParam: 0 (not used)
?0f8d5724a (Shridhar Kumar 2022-04-26 22:32:24 -0400 59)        // lParam[out]: iScintillaView could be 0 (Main View) or 1 (Sub View)
?4dd3b257e (Don Ho         2009-04-24 23:34:47 +0000 60)        // return TRUE

The relatively more recent NPPM_GETCURRENTVIEW API has existed without substantial revision since 2013.

$ git blame -L/NPPM_GETCURRENTVIEW/,+6 --ignore-rev bb278b26f -- PowerEditor/src/MISC/PluginsManager/Notepad_plus_msgs.h

d6ea29c577 (Don Ho 2013-03-03 01:04:09 +0000 697)       #define NPPM_GETCURRENTVIEW    (NPPMSG + 88)
?d6ea29c57 (Don Ho 2013-03-03 01:04:09 +0000 698)       // int NPPM_GETCURRENTVIEW(0, 0)
?775d204ef (Don HO 2018-12-04 13:38:25 +0100 699)       // Get the current used view.
?7874f8dab (Don HO 2020-11-19 04:29:50 +0100 700)       // wParam: 0 (not used)
?7874f8dab (Don HO 2020-11-19 04:29:50 +0100 701)       // lParam: 0 (not used)
d6ea29c577 (Don Ho 2013-03-03 01:04:09 +0000 702)       // Return: current edit view of Notepad++. Only 2 possible values: 0 = Main, 1 = Secondary

Make of that what you will. I would point only out that the documented return value of NPPM_GETCURRENTSCINTILLA is not supported by the actual implementation; the FALSE-returning code path contradicts the statement that this API “Returns always True [sic]”.

Coises

@rdipardo said:

I would point only out that the documented return value of NPPM_GETCURRENTSCINTILLA is not supported by the actual implementation; the FALSE-returning code path contradicts the statement that this API “Returns always True [sic]”.

Technically so, though the FALSE value only happens when the caller sends an invalid parameter (LPARAM null, when it should be a pointer to an int). There’s also the additional result value of -1, which is documented, but which a comment in the code says “cannot happen”; the equivalent doesn’t exist in NPPM_GETCURRENTVIEW.

What troubled me, though, was that the two calls base their responses on different things. I made some attempt to search the code. As best I could tell, the connection between Notepad_plus::_pEditView and Notepad_plus::_activeView is… haphazard. Even though they sound like they represent the same information in different forms (an integer that represents the main or secondary view as 0 or 1, and a pointer to one of two ScintillaEditView structures that represent the main and secondary views), I can’t “prove” that they always change together. So I cannot be certain that they do not represent different information, used for different purposes.

If NPPM_GETCURRENTVIEW were meant merely to supersede the older, clumsier-to-use NPPM_GETCURRENTSCINTILLA, one would expect it to have been based on the same code (if for no other reason, for clarity). I couldn’t find any explanation of why this call was created or what problem it was meant to solve. So I was hoping perhaps someone would know and remember the rationale for the newer call and/or the purposes of the two different variables on which the calls are based, and could tell me either “they’re the same thing, use whichever you like” or “here is the situation in which they can return different results.”