[TC 11.57rc5] Bad encoding checking in "Compare by content"

[yyang]

"Compare by content" sometimes fails to detect the file encoding despite the value of CompareUtf8Detect. My OS is Windows 10 Simplified Chinese Edition.

To reproduce:

1. Save "好的" into good.txt, and "坏的" into bad.txt, both in UTF-8.
2. Select only good.txt and start "Compare by content".
3. Enter the file path of bad.txt manually, and click the "Compare" button. Both files will displayed in the ANSI codepage, not UTF-8.

This issue also happens when both file paths are entered manually.

[petermad]

Confirmed as described. Both files saved with BOM

The same applies when saved as Unicode files

[ghisler(Author)]

Not a bug, this is by design: The encoding check is only performed when the compare tool is initially opened. It's not done when changing the names after opening the tool, otherwise it would not be possible to manually change the encoding.

Moderator message from: ghisler(Author) » 2026-04-28, 07:51 UTC

Moved to will not be changed

[yyang]

Thanks for the reply! Maybe it'll be better if the ANSI<->ANSI button dynamically displays the current (potentially detected) encodings whenever either or both file paths are reset. But setting the encodings manually with this button is good enough.

[ghisler(Author)]

You can select the two files before opening the compare tool, either both in the same panel, or one in each panel. This way they will be detected automatically.

[yyang]

Yes, I'm familiar with this usage.

[white]

Wouldn't it be better if it also starts with Auto-detect when the tool is started with less than 2 files and the comparing isn't done yet?

Then the button looks like this:

Code: Select all

[  Auto<->Auto  ]

And the dialog looks like this:

Code: Select all

Choose Encoding

Left:                        Right:
(*) Auto-detect              (*) Auto-detect
( ) ANSI (plain text)        ( ) ANSI (plain text)
( ) Unicode (UTF-16)         ( ) Unicode (UTF-16)
( ) Unicode MacOS            ( ) Unicode MacOS
( ) UTF-8                    ( ) UTF-8 
                            
     [    OK    ]  [   Cancel   ]  [   Help   ]

If it is set to Auto-detect, and the Compare button is clicked, encoding is determined and set to the determined encoding for left and right.

[ghisler(Author)]

That's an interesting idea, but it wouldn't solve all problems, e.g. when the user changed one of the file names later to a file with different encoding type.

[white]

That's an interesting idea, but it wouldn't solve all problems, e.g. when the user changed one of the file names later to a file with different encoding type.

Why is that a problem? When he changes the filename, the user can set encoding to Auto-detect before hitting the Compare button? Or when he changes the filename the encoding could be set to auto-detect automatically, and if the user wants to he can choose something different before hitting the Compare button.