Do not move diff marks to next line if the user deletes a line about compareplus HOT 30 CLOSED

I'll test that ASAP. Thanks for reporting.

BR

from compareplus.

Thank you Pavel. I appreciate that.

BR

from compareplus.

Confirmed. I'll look into this shortly.

BR

from compareplus.

Fixed in master branch.

BR

from compareplus.

Hello Pavel,

That's a progress. I really appreciate it.
It still needs some tuning up. :)

Could you please try the modified files?

Delete the first not-equal line.
Result: the marker from the next line is removed too.

Similar issues in other cases.

Thank you very much.

BR

Test.zip

from compareplus.

Hello Yaron,

This is fixed, thanks.

BR

from compareplus.

Hello Pavel,

Thank you so much. I really appreciate it.

** I don't expect Compare Plugin to dynamically re-compare, but it would be good to fix this issue if possible.

When I opened this issue I wasn't sure whether Compare or NPP moved the mark to the next line.
Testing Bookmarks, I understand it's NPP.
And if so, this request may well be in the realm of dynamically re-comparing (where do we draw the line?).

Anyway, the current implementation has some major improvements but also some regressions.
Could you please compare the prefs files (prefBackup as current; lines 61-62)?

I haven't checked the relevant code. How about undoing the line-delete, remove its mark and redo the deletion?
That is if if you think it should be implemented at all.

Also, our discussion in #29 might affect this as well.
I would suggest to revert the changes and return to this issue later.
What do you think?

Best regards.

from compareplus.

Hello Yaron,

I'm a bit confused now.
Let's leave aside the dynamic re-compare idea (which is great by the way but way too complex). Just let's leave it for "dessert" ;)

What possible connection might exist between the user deleting some lines from the compared file and the compare process itself? This has nothing to do with re-compare. The line markings were simply left behind while the line is gone (deleted by the user) and the next line markings were moved up together with its corresponding line which resulted in several different markings for the moved lines.
So I just make sure the deleted line/s markings go away with the line/s itself/themselves.

How about undoing the line-delete, remove its mark and redo the deletion?

The current implementation is functionally the same but more optimal. It first removes the markings and then moves to deleting the files. No need for undo/redo.

Anyway, the current implementation has some major improvements but also some regressions.
Could you please compare the prefs files (prefBackup as current; lines 61-62)?

I tried that and I didn't see any problem. Files compare just fine both ways.
Am I missing something?
I admit I haven't tried comparing them with old version of the plugin - please don't make me do so, it's rather painful :)

Please clarify further the problem as I don't see one, thanks!

BR

from compareplus.

Hello Pavel,

Thank you for a delicious meal.
Judging by the appetizers and the main course, the dessert would be great too.
You'll end up with three Michelin stars. :)

What possible connection might exist between the user deleting some lines...

I had in mind the "Two Equal Files" concept. Implementing it, there shouldn't be any empty lines with no marks.
So it was just a vague idea that this might affect the issue in question.

The current implementation is functionally the same but more optimal. It first removes the markings and then moves to deleting the files. No need for undo/redo.

Great!

I tried that and I didn't see any problem.

Please see the screenshots.
http://s24.postimg.org/4sxyqejn9/Old.png
http://s24.postimg.org/wgf9i8u7p/New.png

Just compare. No deletions.

Could some other recent modifications be the cause?

Best regards.

from compareplus.

Hello Yaron,

Could some other recent modifications be the cause?

Sure, it should have something to do with my modifications on the blank lines handling but the weird thing is that I cannot reproduce this issue on my side.
Are you using the latest master branch commit - e0b9c81 ?
Please try again, thanks.

BR

from compareplus.

Hello Pavel,

Thank you very much.

Yes, I'm using the latest master branch commit.
@xylographe, can you please try it?

BR

from compareplus.

@Yaron10
Could you please elaborate?
Which file in the left (main) view?
Delete line 3 in the file on the left or the file on the right?

from compareplus.

Hello xylographe,

With the latest commit I get wrong results regardless of any deletions.
http://s24.postimg.org/4sxyqejn9/Old.png
http://s24.postimg.org/wgf9i8u7p/New.png

prefsBackup in the top/left view. Just compare.

Thank you.
BR

from compareplus.

@pnedev
Something does in fact go wrong. Using the two files from Yaron, prefs Backup.js and prefs.js, comparing in the same way as in the images (prefs.js in the bottom (sub) view) : when I position the caret on blank line 62 (in the top view), then press Del followed by Ctrl-Z (undo) and Alt-D (re-compare) an empty line (not a blank/gap line) has been inserted, marked with a red minus.

After "Clear Compare" the empty line (line 61 in the top view) is still there, although NPP has not marked the file as modified in the tab bar. A subsequent "Reload" of prefs Backup.js makes empty line 61 disappear.

Can you reproduce that, Pavel?

from compareplus.

@xylographe,

But what do you see when you just compare the files and scroll to line 61?

from compareplus.

Hello guys,

I'll leave that for another time, I need some rest.
Can you meanwhile re-test with latest master? Thanks.

BR

from compareplus.

@xylographe ,
After a quick test I was able to reproduce the blank line appearing. I'll investigate later.

BR

from compareplus.

Hello Pavel,

The Compare Plugin seems to require a team of developers.
I've been wondering for quite a while how you can juggle with all these issues at the same time.
Allow me to express again my sincere gratitude and appreciation.
Please slow down. And please have a good rest this weekend. :)

I'd rather continue discussing this issue next week.

Have a pleasant and refreshing weekend.

Yaron

from compareplus.

@Yaron10

But what do you see when you just compare the files and scroll to line 61?

Nothing unusual as you can see in NPP-CP-prefs-001
Using CP f228177

@pnedev

Can you meanwhile re-test with latest master?

Same steps as above lead to same result in CP f228177

from compareplus.

@xylographe,

I've downloaded a fresh portable (7z) NPP 6.9.1 and got the same wrong results.
(I just put ComparePlugin.dll in the plugins folder; no other changes).
Could you please try that?
Could you also upload your built DLL?

I hope Pavel is lying somewhere in the sun and won't be reading any of these messages until next week. :)

Thank you.
Have a nice weekend.

from compareplus.

ComparePlugin-f228177 (built with VS 2015)
I'm using normal, downloaded 32-bit NPP 6.9.1 on Win7SP1[64]

from compareplus.

Thank you. I appreciate that.

It works properly with your build.
I'm using VS 2013 Express. Deleted everything and rebuilt it. Never had a problem before.
Any idea? If not, I'll wait until Pavel returns from his vacation. :)

BR

from compareplus.

Oops, I typed "(built with VS 2015)" but I meant "(built with VS 2013)"

My idea: Upgrade to VS Community 2013 perhaps?

from compareplus.

Thanks again.
I'll wait for Pavel. He might have another idea.

BR

from compareplus.

Hello guys,

@xylographe ,
The issue with the appearing line is a tough one.
This is because now I'm manually deleting the mark of the line and when you undo, the line is automatically back but I do not restore its mark. At that moment N++ knows the file is not changed but when you clear the compare results the extra blank is not deleted because the marking is missing and you finally get a file that appears saved but with an extra line. @Yaron10 was right to some extent that this issue is on the border line with a dynamic re-compare.
I'll think about fixing that, I have an idea already.

@Yaron10 ,
It's very strange that VS 2013 Express is giving different results. I have no idea now what might be wrong but it is a compiler issue for sure. I'll try to analyze it further.

BR

from compareplus.

Hello Pavel,

Thank you. I appreciate it.

I think the compiler issue is a question for Claudia.

BR

from compareplus.

Hello Yaron,

OK, that's good. Claudia may have some idea about that or think of something we miss.

BR

from compareplus.

Hello @xylographe ,

The particular issue you reported is fixed with the latest commit, thanks.

BR

from compareplus.

Hello Pavel,

That's a major improvement.
Thank you.

BR

from compareplus.

Thank you, Pavel.
Fix confirmed. (CP 81dbed4)

from compareplus.