Hacker News Reader

Comment by intelijstupide | Hacker News Reader

1. I upgraded a huge application from Vite 2 to Vite 3 with literally no change.

2. Nobody forces you to upgrade, that's how semver works.

1) If it didn't break backwards compat why did it bump the major version?

2) Constantly pushing major versions in order to ignore backwards compatibility and pass that problem on to consumers might not break semver but you're still wasting my time by expecting me to clean up your mess if I choose to depend on your library. (Which I won't)

Or are you suggesting vite will backport security patches indefinitely?

It broke backwards compatibility in the sense that the minimum NodeJS version was incremented.

Re 2.: unless they plan to simultaneously support multiple major versions, they kind of do. Maybe not right away, but once the main branch of development is based on the new major version, that's the highest priority moving forward. All older versions might get backported bug fixes, but that's it.

op is just bad mouthing by implying that this release magically broke their app overnight.

I'm not bad mouthing anyone. I wasn't even specifically targeting vite. I just simply don't care for the people in the industry that think they have ownership over my time. If they want me to use their shit, they need to make it worth while.

The speed at which you release breaking change is the speed in which I run in the opposite direction.

Look at all the arguments in this thread "you don't have to" "its optional" "its just a small change" "it hardly even needed fixing".

Do they not realise that me even having to look at this cost me time and money? Do they not realise that when you multiply that out to all your devs at their day jobs and then multiply it across the entire stack of software all pulling these stupid stunts that it adds up to a huge waste of time for everyone involved except the snowflake that thought their breaking change was worth it?

"Utter failure" seems to be way overstated given the changes:

* Requirement on slightly newer versions of browsers

* Some deprecated functions removed

* New approaches with old ones still supported via "legacy mode" flags.

* A few settings changed their defaults

All in all, seems pretty reasonable for a major version bump.

Breaking changes (https://vitejs.dev/guide/migration.html) are minimal and unlikely to affect most applications, and when they do the issues should be trivial to fix.

Major version bump !== Your app breaks

According to semantic versioning, a major version indicates breaking changes, which is always bad from a users perspective.

Most projects don't really use semver anymore though, which is why this association seems to be going away

It doesn't mean that anything will break. You deprecate old ways long before you remove them from your tool so that by the time you do the breaking change only a tiny slither of people are affected by it.

Well, because like with many frameworks, they haven't thought it through yet ;)

because maintaining backwards compatibillity is an antipattern, its producing so much needless global complexity in our tools for everyone forever, just to save some effort when upgrading ...

Reply on news.ycombinator.com