I don't have the time to read the whole thread right now (and I'm not sure I
ever will), but I think this is important, so here are my comments to the
original proposal (to which I am totally opposed!):
Stephen Gallagher wrote:
Right now, we have a policy that essentially forbids source code
from
being bundled into a package. In technical terms, this means
essentially that the packaging policies mandate that any code that
appears more than once in the repository must be turned into a shared
library and dynamically linked into any package that requires it.
And that's a good thing.
I would like to propose that the no-bundled-libraries policy be
amended as follows: "Any package that has an existing mechanism to
link against a shared system library and functions correctly when
doing so must link against that library and not bundle it internally.
Any package whose upstream releases cannot link against a shared
system library (or are incompatible with the version in Fedora) may
bundle that library (without requiring a special exemption) but MUST
add Provides: bundled(<libname>) = <version> in the spec file for each
known bundled library.(This will allow us to track down the bundling
when we need to). Package maintainers should continue attempt to
engage upstream to support linking against shared system libraries
wherever possible, due to the advantages it provides the package
maintainer."
This will lead to package maintainers not even TRYING to unbundle libraries.
But that is really what the packager is for. Instead, what we see right now
is packagers just running to FPC for an exception, and if they don't get it,
the package gets threatened of removal from Fedora if it is already in, or
gets stuck under review forever otherwise. With your proposed policy, it
would just bundle everything with no attempt whatsoever at fixing that. We
need more proactive downstream unbundling. We have to realize that upstream
will not care and that is is our job to fix their mess.
If the library is incompatible, there are generally 2 approaches to solving
the problem: patch the application to not need the incompatible library
additions, or patch the library to add them. Which one is more workable is a
case by case decision. (In particular, we definitely need library
maintainers in Fedora to be more flexible when it comes to adding non-
upstream patches to make applications work. "Upstream, upstream, upstream"
does not always work, it should not be a dogma.) If the application simply
needs a too new version of the library, then often this is a recent change
in the application, so simply reverting the commits that need the new
library is an effective solution. (In some cases, that also works for non-
upstream library changes: revert the application commit that started using
them.)
The reason for this proposal is relatively simple: we know the
advantages to unbundling, particularly with security and resource-
usage. However, the world's developer community largely *does not
care*. We fought the good fight, we tried to bring people around to
seeing our reasoning and we failed.
See above. Upstream will not care. Some of them never did. (This is not a
new problem.) It is OUR job to unbundle it in downstream patches. If
upstream accepts the patches, all the better. If they don't, we need to keep
the downstream patches.
The point of software is to provide a service to an end-user. Users
don't run software because it has good packaging policies, they run
software because it meets a need that they have. If they can't get
that software from Fedora, they *will* get it from another source (or
use a different OS that doesn't get in their way). I'll take a moment
to remind people that two of Fedora's Four Foundations are "Features"
and "First". We want Fedora to be the most feature-complete
distribution available and we want to get there before anyone else
does. I would say that holding to our no-bundling policy actively
defeats our efforts on that score.
That really means our packagers are failing to do their job.
== Advantages to bundling ==
* Guarantees that the application is running with the same set of
code that upstream tested. (Fewer Fedora-specific bugs means less
burden on the maintainer)
But multiplies the number of library versions in a given Fedora release by
the number of applications! In addition to the obvious space and security
issues, this is also a debugging/testing nightmare!
* Simplifies packaging of updates. (Fedora maintainer does not need
to keep tabs on unbundling patches to keep in sync for new versions)
Not really. The maintainer then has to keep track of things such as new
BuildRequires (and file list changes if there are more files than just one
static library that's linked directly into the binary and not installed) for
each and every bundled library.
Running into FTBFS bugs (which can sometimes happen on package updates, but
most commonly, it happens when something else in the distribution gets
updated) is also all the more likely the more libraries are bundled.
* Increases the available pool of software that can be packaged
substantially (many modern languages such as Ruby and Go are
realistically only functional with allowable bundling)
We need proper packaging guidelines that address that properly. The Go
guidelines (distributing libraries in source form) are a joke! Go software
should only be allowed with gcc-go shared libraries, and all libraries
should then be packaged separately as shared libraries.
We need to ignore the language's own wannabe package management entirely and
patch the build systems (hopefully centrally-installed ones – if it's
autotools-style copypasta, then we need to patch each and every package,
sadly) to work with system packages.
And for languages like Java where libraries get bundled as binary blobs, I
think allowing that kind of bundling would be totally unacceptable and would
go against our commitment to freedom.
* Did I mention the reduction in maintainer burden yet?
It's less burden initially, but fighting with updated build dependencies for
the bundled libraries, FTBFS issues in the bundled libraries, etc. may well
turn out to be MORE work. And unbundling delivers much better quality
packages (disk space consumption, RAM consumption, timeliness of bug and
security fixes), for which doing more work (if it is actually more work) is
warranted.
Kevin Kofler