On Fri, Jun 28, 2019 at 11:47:46AM +0200, Rhonda D'Vine wrote: > > Bugs that appear in stable have to get fixed in stable. The release > team doesn't block bugfixes for stable, you are exaggerating here - and > data loss is obviously a serious problem, especially when it is just a > one-line change. Your statement makes it sound like the stable release > team is a nutcase -- which it clearly isn't. Please don't imply that. :/
There are *very*, *very*, *very* large number of bug-fixes for e2fsprogs in version 1.43.[3456789]. This isn't actually the first data corruption bug since 1.43.2. And this doesn't include the data corruption bugs found in 1.44.x and 1.45.x that apply to 1.43.2; that would far more effort to track. So there's actually a huge problem hiding here, which is that it's just way too much effort to backport all of the fixes. The problem is that the release team isn't willing to take the maintenance branch updates, but rather wants every single change to be logged as a Debian bug, and individually extracted as a patch, and applied manually to debian/patches. And that's more effort than I as volunteer am willing to do. Red Hat charges a ***huge*** sum of money for crazy enterprise customers who want that kind of loving care of each bug fix being manually curated. And that's because it's a huge amount of work. If someone wants to volunteer to do that work, that's great. But that's what I would really love to find. And if they only want to do it for *this* bug, as opposed to people using resize2fs to shrink file systems, which is actually far more likely to happen --- although, I'm not sure how many customers using Debian Obsolete do that sort of thing --- that's fine. But when there are a huge number of bug fixes, it's very hard to bring the motivation to do what Red Hat charges $$$$$ for its customers to do, when the rest of the open source world just says, update to the latest !@#!@?! version, already! - Ted P.S. This isn't an issue just for e2fsprogs. Even firefox-esr is only supported for less than a year before the answer is, "Sorry, you want the new fixes, you have to take some new features." It's a huge engineering effort to disentangle bug fixes and backport them to aging code bases.