because the 0.38 docs lacks any "new in 0.38" docs, so it has to be guessed

The doc is designed for people writing new code against VTE, not for those porting their code from older versions. As such, it solely documents the 2.91 API, and therefore only mentions additions that weren't yet available in the very first 2.91 version (that is, things you need to be aware of if developing against 2.91). In other words, on a backwards incompatible API bump the doc is started over from scratch. With this regard, it's absolutely irrelevant when something appeared and how (if differently) looked like in the previous API. (I do not intend to discuss whether it's good or bad this way, I just stated that this is the reason. I think it's the same with, let's say, GTK+ version 3, you won't see any GTK+-2 stuff or 2->3 porting advice mentioned in GTK+-3's docs.)

No. Unless there's a compelling reason to drop 2.90, we should not drop it.

Con of dropping: Save one step for users who have an old enough system not to have libvte 2.91, however, wish to have the latest Geany with as little hassle as possible. (I never quite understood this goal, I believe those wishing to run the latest apps are expected to keep their base system reasonably up to date as well. Sure it's nice to have some window, but maybe 2 years should be enough.)

Pros of dropping: Cleaner, simpler code, easier maintenance, fewer setups to test. Plus, the aforementioned users, although suffer from the inconvenience of having to install another lib, will get the new features, bufixes, security/privacy fixes (e.g. scrollback buffer no longer written plaintext to disk) of new vte.

As a VTE developer, I see no reason in supporting both versions in parallel, I pretty much find it a waste of the precious (and apparently scarce) developer resources. However, your project, your call.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.