All except the first are "block" characters from the same unicode set, so they are almost guaranteed to be the same width, even in proportional-width fonts (I'm using oxygen-font). However, the first character is an underscore, which isn't related to the others, and often has a different width. This means that the overall width of the bar graph fluctuates depending on how many "low" values are currently displayed.
I think it looks better if we just drop the underscore from the list (I think the per-CPU bar graph already uses these characters, without the underscore, so it's slightly more consistent too).
ammgws
commented
4 years ago
The bar graph used for up/down net graph uses these characters to represent different levels:
https://github.com/greshake/i3status-rust/blob/d258f6614efd790c2969b6db4eb6128287df4a3c/src/widgets/graph.rs#L52
All except the first are "block" characters from the same unicode set, so they are almost guaranteed to be the same width, even in proportional-width fonts (I'm using oxygen-font). However, the first character is an underscore, which isn't related to the others, and often has a different width. This means that the overall width of the bar graph fluctuates depending on how many "low" values are currently displayed.
I think it looks better if we just drop the underscore from the list (I think the per-CPU bar graph already uses these characters, without the underscore, so it's slightly more consistent too).