Bug: UTF-8 encoding in Github repo, issue #315

We’re having a problem with displaying unicode characters when the users’ PC isn’t set up in a certain way. The UI incorrectly displays non-Latin characters. This may be because github is not recognizing some files as UTF-8. See this discussion thread here:

image

Here’s a .gitattributes file example to define how to encode c# files from Source of .gitattributes - Api.Samples - dRofus Bitbucket

C++ source code

*.rc text *.rc2 text *.cpp text diff=cpp working-tree-encoding=UTF-8 *.c text diff=cpp working-tree-encoding=UTF-8 *.h text diff=cpp working-tree-encoding=UTF-8 *.hpp text diff=cpp working-tree-encoding=UTF-8

Github issue #315 here: BUG: UTF-8 encoding issue · Issue #315 · FashionFreedom/Seamly2D · GitHub

Susan,

I believe this is exactly the problem… Like I mentioned I have my Atom editor set at UTF-8 and the app(s) compile fine without the added chars on the unicode chars. If I change this then the files(s) are not interpreted correctly and the extraneous char(s) show up.

UTF

In other words if I compile the exact same code Github just compiled, I get the correct chars, while Github does not. I believe that indeed there it could be something incorrect in the .gitattributes file that’s causing Github to not recognize the files as UTF-8.

Can we just create a test branch and adjust the gitattributes to see what works?

Hey I made the .gitattributes update. Building test release now, in about an hour @douglas @grace & @pneumarian please download & test from here:

Here in Ubuntu Linux the problem was not. & the fix didn’t break anything obvious.

:unicorn:

Did not fix the Windows install.

Just to be clear… are you saying there is no problem with the Ubuntu distribution?

Yes, I have never experienced this bug in Ubuntu Linux, & the fix didn’t introduce new issues. (which obvious phrasing I wish I’d been able to think of previously.)

Hmmm… Well that would seem to indicate there’s something in the Windows build chain on Github that’s not handling the source files as UTF-8? I don’t have the issue with my local builds, and for the heck of it I checked a recent Valentina Gitlab build and there is no issue there either.

Hi @Pneumarian, please test this build:

@Sue

Just built 315 b - Fixed here. :slight_smile:

Super fantastic…!

Nothing broken over here. @Grace? How’s it look for you?

Sorry, only seeing this now. Having power & water issues again.

image

The % is still looking the same:

image

I’m still wondering if it’s not the MS Visual C++, since this only started when I had to download and install that along with the 1st build.

Something else that happens from time to time that I can’t quite put my finger on, is that it seems to almost hang while doing certain tasks, just long enough to let me think ‘What now?’ but then it goes on. I haven’t been able to quite predict which tasks yet.

@grace what are your language & label preferences?

@grace Are you seeing character problems with point names?

Sorry about the water and electricy problems. Hope you’re okay! Stay safe!

Hi @Sue GUI language is English Label language is American English Font is now Arial - I tested a few others but the remains "Å "

No, the points are fine.

image

Thank you :slight_smile:

So to recap the current state of this issue:

  • Character encoding in the canvas is working.
  • Character encoding in the GUI executable is not working.

This works: Encoding characters external to the executable, eg point names read from the XML pattern file:
image

This works: Encoding characters external to the executable, eg file names read from the operating system:
image

This doesn’t work: Encoding some internal characters within the executable. The degree symbol ° displays as  °
image

Are there other instances where the character display isn’t working?

And is this really an encoding problem, or is ° in the code?

On my Linux laptop, the angle symbol displays correctly. image