I'm happy to discover Geany 2 with same window but for me there is a major issue: text is blurred on Debian Testing (2.0-1), see below:
Geany 1.38 - GTK 3.24 
Geany 2.0 - GTK 3.24 
Both with `FREETYPE_PROPERTIES="truetype:interpreter-version=35"`.
See https://github.com/geany/geany/issues/3610
Basic advice, try other fonts.
Hum...
DejaVu Sans Book 
DejaVu Sans Mono Book 
DejaVu Serif Book 
Lato Hairline/Thin/Lite/Regular/Medium... 
Monospace Regular  [bigger image](https://github.com/geany/geany/assets/31816829/351527b3-a5f6-4ef4-b7a7-7e7fe...)
So no, there is something wrong. Not sure how to help/debug.
Well, you actually only tried Lato and DejaVu fonts, monospace is an alias, usually of DejaVu. DejaVu is rubbish, I do wish distros wouldn't set it as default, but on my system here I can't see any difference between Geany 1.38 and 2.0 with DejaVu sans mono, and neither looks "blurry" until the point size is very small, and then they both show the _same_ AA effects, mostly on verticals.
It may depend on exactly the versions on your system of the font, of harfbuzz, of Pango, of Freetype, and the screen resolution, yes there are lots of moving parts, and each new Debian changes them (well not your screen, but the software ones).
I can't see anything wrong with your image of Lato, but maybe thats the screen capture and image rendering process.
Download and try [Hack](https://sourcefoundry.org/hack/)
Ok here is a test with Geany 2.0-1 (top) and Geany 1.38 (bottom) wirth Hack font:
Please use the same `View-> Color Scheme` in both and use the same file, they seem to have different line lengths, but is that true or is it a display artifact?
Yes, true, sorry. I restarted Geany 2.0 and 1.38 with default configuration but with Hack font (bellow Pluma 1.26 and Geany 2.0 with Virtual Box, same "wrong" rendering):
---
So I also installed Pluma next to Geany 1.38... but it's strange, I have the right rendering between both:
So I upgrade Geany to 2.0, and the wrong rendering is here:
---
I'm not sure to understand all.
I'm sorry I can't actually see any significant difference between Pluma or the two Geany versions, even if I increase the image size. I wonder if the compression in the PNGs is hiding what you want to show, or if its that my screen resolution interacts with an image taken at your screen resolution to hide the effect?
This is just our old friend `pango_context_set_round_glyph_positions`. Try it with `TRUE` instead of `FALSE` in `scintilla/gtk/PlatGTK.cxx` and each `U` will look the same. While that looks more consistent for the example, it doesn't always look better and the layout is less 'correct' with positioning discrepancies more likely.
The pattern of 5-10 instances being identical then moving to a different pattern could be something like aliasing to fixed-point positions (like an eighth of a pixel) instead of using true floating-point. I've considered doing this in Scintilla to save space.
Scintilla does not place each character individually - instead it asks Cairo and Pango to draw whole segments of text. There is also no possibility of the text being expanded/contracted to fit within the width allocated by Scintilla since the width (or right ordinate) are not passed to Cairo/Pango here.
@elextr: I can't actually see any significant difference
There are layers of image resizing and smoothing getting in the way. Save the image into a file then use a bitmap viewer (like Image Viewer with the 'Smooth images' options turned off in Preferences) with integer magnification to look at the image closely.
Downloading and viewing with all filters off I can see vertical bands in the `U`s, but still no "blurred". But I don't see it the bands on my screen so clearly it depends on the relationship of screen resolution, font, and font size. Somebody who can see it on their system needs to try the suggestion of setting TRUE.
A more visible and annoying positioning artefact is shown in #3738 and that is fixed with TRUE.
This leads to the question, @nyamatongwe what negative effects does setting TRUE potentially cause? Since its hard coded we can only pick one value.
what negative effects does setting TRUE potentially cause?
#3111
github-comments@lists.geany.org