3 ms·
When snapshots were the only game, it was all unstable. There's no difference from today. EDIT: sorry, it seems I've mis-read you. I get it now. Unfortunately,
by steveklabnik 11y ago
When snapshots were the only game, it was all unstable. There's no difference from today.
EDIT: sorry, it seems I've mis-read you. I get it now. Unfortunately, we have a strong compatibility story, and so must take strong measures to enforce it. You can use the nightly released on the same night to get a build that's like a particular stable, but allows features.
- mindslight 11y agoSure, that is simply stating a different way. With the resulting implication that if one wants procedural macros, the stable 1.x releases have unfortunately been empty progress. Breakage from forwards-incompatibility isn't my problem, it's potentially having to tackle it every morning rather than in batch mode when there's a new release. edit: If this compatibility story really dictates that releases can't contain beta features, even with a #![unstable_beta_features_1_6] directive, then a separate beta release would be nice. But now it looks like I'm asking for something additional.
- dikaiosune 11y agoThere is a beta release (https://www.rust-lang.org/downloads.html https://www.rust-lang.org/downloads.html), but I don't think it includes unstable features.
- mindslight 11y agoOh yeah, forgot about that. Different purpose than what I'm talking about though. Perhaps mine would be more appropriately named "1.6 with alpha features".
- dikaiosune 11y agoI had the same response when I realized a feature I really wanted was only on nightly (and would probably still be interested). However, in the last month-ish I've been using a fresh nightly every day, multirust has made it super easy to update, and I've only had breaking changes to my dependencies once.
- mindslight 11y agoI only switched to building releases at 1.4 because I thought the nightly treadmill could finally be over. Alas I guess I'll skip installing 1.6 and just switch back to sporadically-updated checkouts whenever some dependency has breaking changes (ooh, such stable). I'm not looking to get the rats on my face and switch to butt-based development ala rustup/multirust/etc.
- kbenson 11y ago> butt-based development Oh, cloud-to-butt, will you ever stop delivering the fun? I hope not.
- Gankro 11y ago1.6 with alpha features is essentially the nightly on the day that 1.6 entered beta. Backports are incredibly rare while the trains are running.
- kbenson 11y agoI think the issue here is the difference between "is" and "essentially". Having a guaranteed matched release with alpha features would apparently be useful, for at least one person. Not being guaranteed could cause some real headaches under certain circumstances.
- pcwalton 11y agoWhat are those circumstances? There are no stability guarantees for unstable features, so I don't see how a nightly build that happens to correspond to a major version has value over any other nightly build. The stability you would get by sticking to "1.6 + features" seems precisely equal to the stability you would get from picking a recent nightly and sticking to that build. It sounds to me like the real issue here is that procedural macros aren't stable, not anything about the release process.
- mindslight 11y agoStabilizing procedural macros would fix this one issue, but IMHO a different release process would have fix the pain as well. My case specifically: Wanting hard-to-stabilize features like procedural macros and no_std. Using external dependencies that work with stable, but when used with unstable they expect you are updating nightly. Not updating nightly due to a wacky idea that people should control tools and not vice-versa. The first time I open a project in a week is not the time I want it to fail compiling due to language/library changes. Maybe my impression is caused by the previous rapid change of nightlies (renaming in libraries, etc), and using month-old snapshots wouldn't be too bad these days. But still after several major releases I'd hope the stable tarballs would have been suitable for general use. Instead of removing not-yet-stable features, they should be put behind a feature gate. With the idea that anybody who explicitly opts-in to unstable features and then complains when they're not forward-compatible can simply be told to pound sand.
- JoshTriplett 11y ago> Unfortunately, we have a strong compatibility story, and so must take strong measures to enforce it. You can use the nightly released on the same night to get a build that's like a particular stable, but allows features. Unfortunately, the crates.io system does not follow the same policy, and crates regularly require feature flags.
- steveklabnik 11y agoYup! The majority of crates do build with stable, but many require features. We've been talking about a way to expose this in the UI so you can tell, but it's slightly more complex than you'd think at first, and nobody has put the time in yet.
- eru 11y agoWould the ghc-like LANGUAGE pragmas work?
- steveklabnik 11y agoWe already have feature flags. But some projects use them to add extra stuff on nightly, but build on stable. That's just one example...
- codingbinary 11y agoYep, exactly that killed my window manager on stable. I _have_ to use nightlies now, because I need the alloc_jemalloc feature.
- steveklabnik 11y agoHow are you using it? jemalloc is already the default.
- codingbinary 11y agoReally? Last time I checked it still didn't work. Will have to try when I get home. The problem was that the config is compiled into a dynamic library and gets passed the config struct. The old alloc method kind of died on me when the dynamic library was trying to alter that struct. So I had to enable jemalloc to keep it from crashing.
- steveklabnik 11y agoHmm. Yeah, I mean, jemalloc is the default allocator, so something must be wonky here. You don't need to explicitly use jemalloc. That's what you get by default.
- codingbinary 11y agoBut that wasn't always the case, right? I think around the time of 1.3/1.4 jemalloc was not the default.
- Gankro 11y agoQuite the opposite; jemalloc has been the default, and using the system allocator has been the "hassle" since well before 1.0. However a few key platforms have jemalloc disabled because it's buggy (deadlocks or worse). I think as of a few days ago it's pretty much universally off on windows. Also whether jemalloc is used depends on how you build your thing -- dylibs use the system allocator (because they're subordinate), static libs (rlibs) inherit from the thing they're linked into, and executables use jemalloc (because they're in control).