Builder offers plenty of settings to run apps with a specific context, but one I feel it missing is running apps with a pseudolocale. @hadess created gettext-pseudolocale a few years ago (see also the related blog post), and while I fond it pretty hard to read, I like the idea of having some pseudolocale to help detect internationalization issues early. A pseudolocale wich widens labels is welcome too as English is fairly compact, which often leads to too wide windows, poorly ellipsized labels, etc. in other languages.
3 Likes
The “Bent” string replacement is not really supposed to be readable. It’s barely readable enough to detect problems. I would happily take a patch that adds the “squares” mangling:
Or maybe you want to try the undocumented easter egg by setting the mode to “malkovich”:
2 Likes
Android’s “English (XA)” claims to produce pretty readable, accented, characters:
Oh the square mangling seems nice! It pretty clearly discriminate internationalized strings from non-internationalized ones, it remains very readable, it slightly widens the text, and it maybe doesn’t work too bad with the screen readers (at least the rounded bubbles one work, so hopefully the squares may work too).
Since yesterday I’m contributing to gettext-pseudolocale to help make it robust enough for us to rely on it.
1 Like
We now have some bugs fixed, regression tests, and a “squares” mode. I believe gettext-pseudolocale is now reliable enough to start using it, unless maybe @hadess thinks otherwise.
“not” or “now”?
I think it might be generally useful to have as a helper tool in a Flatpak SDK, whether that’s GNOME’s or Freedesktop’s, and I can make releases as needed if that’s a requirement.
1 Like
“now”, sorry for the typo.
So… is this something we want? Is it worth adding gettext-pseudolocale to the GNOME or FDO SDK and adding the feature to Builder? Just to know if it’s something I should work on or not. 