Commit graph

4094 commits

Author SHA1 Message Date
eldritch horrors ed6b3165ea worker: process timeouts first, and exclusively
check goals for timeouts first, and their activity fds only if no
timeout has occurred. checking for timeouts *after* activity sets
us up for assertion failures by running multiple build completion
notifiers, the first of which will kill/reap the the goal process
and consuming the Pid instance. when the second notifier attempts
to do the same it will core dump with an assertion failure in Pid
and take down not only the single goal, but the entire daemon and
all goals it was building. luckily this is rare in practice since
it requires a build to both finish and time out at the same time.

writing a test for this is not feasible due to how much it relies
on scheduling to actually trigger the underlying bug, but on idle
machines it can usually be triggered by running multiple sleeping
builds with timeout set to the sleep duration and `--keep-going`:

    nix-build --timeout 10 --builders '' --keep-going -E '
      with import <nixpkgs> {};
      builtins.genList
        (i: runCommand "foo-${toString i}" {} "sleep 10")
        100
    '

Change-Id: I394d36b2e5ffb909cf8a19977d569bbdb71cb67b
2024-05-26 20:57:09 +00:00
Artemis Tosini 53e2b0740c
libstore: remove duplicate builder variable
The `builder` local variable and duplicate `args.push_back` are no
longer required since the Darwin sandbox stopped using `sandbox-exec`.
The `drv->isBuiltin` check is not required either, as args are not
accessed when the builder is builtin.

Change-Id: I80b939bbd6f727b01793809921810ff09b579d54
2024-05-25 22:14:55 +00:00
jade dd53bce476 Merge "libstore: parse the buildMode instead of unchecked cast" into main 2024-05-25 17:42:09 +00:00
jade 2a7a824d83 libstore: parse the buildMode instead of unchecked cast
Change-Id: Icf6af7935e8f139bef36b40ad475e973aa48855c
2024-05-24 20:45:05 -06:00
alois31 f047e4357b libstore/build: always enable seccomp filtering and no-new-privileges
Seccomp filtering and the no-new-privileges functionality improve the security
of the sandbox, and have been enabled by default for a long time. In
lix-project/lix#265 it was decided that they
should be enabled unconditionally. Accordingly, remove the allow-new-privileges
(which had weird behavior anyway) and filter-syscall settings, and force the
security features on. Syscall filtering can still be enabled at build time to
support building on architectures libseccomp doesn't support.

Change-Id: Iedbfa18d720ae557dee07a24f69b2520f30119cb
2024-05-24 21:19:29 +00:00
jade 9530b7f2b2 Merge "packaging: rename nixexpr -> lixexpr and so on" into main 2024-05-23 23:19:28 +00:00
jade c97e17144e packaging: rename nixexpr -> lixexpr and so on
This breaks downstreams linking to us on purpose to make sure that if
someone is linking to Lix they're doing it on purpose and crucially not
mixing up Nix and Lix versions in compatibility code.

We still need to fix the internal includes to follow the same schema so
we can drop the single-level include system entirely. However, this
requires a little more effort.

This adds pkg-config for libfetchers and config.h.

Migration path:
expr.hh      -> lix/libexpr/expr.hh
nix/config.h -> lix/config.h

To apply this migration automatically, remove all `<nix/>` from
includes, so: `#include <nix/expr.hh>` -> `#include <expr.hh>`. Then,
the correct paths will be resolved from the tangled mess, and the
clang-tidy automated fix will work.

Then run the following for out of tree projects:

```
lix_root=$HOME/lix
(cd $lix_root/clang-tidy && nix develop -c 'meson setup build && ninja -C build')
run-clang-tidy -checks='-*,lix-fixincludes' -load=$lix_root/clang-tidy/build/liblix-clang-tidy.so -p build/ -fix src
```

Related: lix-project/nix-eval-jobs#5
Fixes: lix-project/lix#279
Change-Id: I7498e903afa6850a731ef8ce77a70da6b2b46966
2024-05-23 16:45:23 -06:00
Artemis Tosini 3de77e6dbd Merge "libutil: Create chmodPath function" into main 2024-05-20 15:13:53 +00:00
Artemis Tosini 5411fbf204
libutil: Create chmodPath function
Move the identical static `chmod_` functions in libstore to
libutil. the function is called `chmodPath` instead of `chmod`
as otherwise it will shadow the standard library chmod in the nix
namespace, which is somewhat confusing.

Change-Id: I7b5ce379c6c602e3d3a1bbc49dbb70b1ae8f7bad
2024-05-19 22:07:58 +00:00
eldritch horrors 774c56094f libstore: fix old RemoteStore::addToStore serializer
having the serializer write into `*conn` is not legal because we are
in a sinkToSource that will be drained by the remote we're connected
to. writing into `*conn` directly can break the framing protocol. it
is unlikely this code was ever run: to protocol it caters to is from
2016(!) and thoroughly untested in-tree, and since it's been present
since nix 2.17 and the 1.18 protocol broken here is nix 2.0 we might
safely assume that daemons older than nix 2.1 are no longer used now

see also #325 (though that wants <2.3 gone, this is sadly only <2.1)

Change-Id: I9d674c18f6d802f61c5d85dfd9608587b73e70a5
2024-05-19 11:57:55 +00:00
Alyssa Ross 139d31f876 Improve nix-store --delete failure message
On several occasions I've found myself confused when trying to delete
a store path, because I am told it's still alive, but
nix-store --query --roots doesn't show anything.  Let's save future
users this confusion by mentioning that a path might be alive due to
having referrers, not just roots.

(cherry picked from commit 979a019014569eee7d0071605f6ff500b544f6ac)

Upstream-PR: https://github.com/NixOS/nix/pull/10733
Change-Id: I54ae839a85f3de3393493fba27fd40d7d3af0516
2024-05-18 14:49:40 -06:00
Pierre Bourdon d1c8fd3b09 Merge "derived-path: refuse built derived path with a non-derivation base" into main 2024-05-18 07:26:26 +00:00
Pierre Bourdon 5a1824ebe1
derived-path: refuse built derived path with a non-derivation base
Example: /nix/store/dr53sp25hyfsnzjpm8mh3r3y36vrw3ng-neovim-0.9.5^out

This is nonsensical since selecting outputs can only be done for a
buildable derivation, not for a realised store path. The build worker
side of things ends up crashing with an assertion when trying to handle
such malformed paths.

Change-Id: Ia3587c71fe3da5bea45d4e506e1be4dd62291ddf
2024-05-17 02:16:15 +02:00
julia 5b7dcb3005 Allow enabling core dumps from builds for nix & child processes
Fixes lix-project/lix#268

Change-Id: I3f1b0ddf064f891cca8b53229c5c31c74cea3d9f
2024-05-16 17:11:21 -07:00
Artemis Tosini 4b35e6a75e Merge "libstore: Fix sandbox=relaxed" into main 2024-05-12 03:51:19 +00:00
Pierre Bourdon 38d825b21e
filetransfer: fix decompression regression from 121edecf
121edecf65 added a new state field to
carry over content encoding settings from transfer to sink creation, but
never actually set that field.

Change-Id: I714b2efe745561e851b78a4791479b3501db8c72
2024-05-11 14:49:23 +02:00
raito 4ebbd4362f Merge "feat(libstore): print the first line of stdout of SSH in case of failure" into main 2024-05-10 19:33:36 +00:00
raito 8404a1f66d feat(libstore): print the first line of stdout of SSH in case of failure
In case of failure to connect as can be seen in
https://buildbot.lix.systems/#/builders/39/builds/1386/steps/1/logs/stdio

It is difficult to understand what happened, if we enabled the talkative
verbose level, we could learn about the first line SSH sent us.

In practice, this is not workable, we can just make it warn all the
time.

Change-Id: Iaaf56894060a58f2dfc78254bb60b1c43482f9bb
Signed-off-by: Raito Bezarius <raito@lix.systems>
2024-05-10 20:22:47 +02:00
eldritch horrors ceccac835c libutil: remove callback.hh
it's no longer used. it really shouldn't have existed this long since it
was just a mashup of both std::promise and std::packaged_task in a shape
that makes composition unnecessarily difficult. all but a single case of
Callback pattern calls were fully synchronous anyway, and even this sole
outlier was by far not important enough to justify the extra complexity.

Change-Id: I208aec4572bf2501cdbd0f331f27d505fca3a62f
2024-05-10 02:21:11 +02:00
eldritch horrors b66451ae7f libstore: de-callback-ify FileTransfer
also add a few more tests for exception propagation behavior. using
packaged_tasks and futures (which only allow a single call to a few
of their methods) introduces error paths that weren't there before.

Change-Id: I42ca5236f156fefec17df972f6e9be45989cf805
2024-05-10 02:21:11 +02:00
eldritch horrors 28a98d152c libstore: de-callback-ify Store::queryRealisation
Change-Id: I8d74745c519518f163f51dfaa39063836f17599e
2024-05-09 23:18:05 +02:00
eldritch horrors 17965bf11c libstore: un-callback-ify Store::queryRealisationUncached
Change-Id: I4a328f46eaac3bb8b19ddc091306de83348be9cf
2024-05-09 23:18:05 +02:00
eldritch horrors 2f4a1dd6e0 libstore: de-callback-ify Store::queryPathInfoUncached
Change-Id: I23a156aaff5328f67ca16ccd85c0ea1711b21e35
2024-05-09 23:18:05 +02:00
eldritch horrors c77bd88259 libstore: de-callback-ify BinaryCacheStore::getFile
Change-Id: I36b3eb9f645aa04058151e7b2353e15e6f29057b
2024-05-09 23:18:05 +02:00
eldritch horrors 1a002d1a11 libstore: de-callback-ify CA realisation substitution
this is the *only* real user of file transfer download completion
callbacks, and a pretty spurious user at that (seeing how nothing
here is even turned on by default and indeed a dependency of path
substitution which *isn't* async, and concurrency-limited). it'll
be a real pain to keep this around, and realistically it would be
a lot better to overhaul substitution in general to be *actually*
async. that requires a proper async framework footing though, and
we don't have anything of the sort, but it's also blocking *that*

Change-Id: I1bf671f217c654a67377087607bf608728cbfc83
2024-05-09 23:18:05 +02:00
Théophane Hufschmitt adea821d87
libstore: Fix sandbox=relaxed
The fix for the Darwin vulnerability in ecdbc3b207
also broke setting `__sandboxProfile` when `sandbox=relaxed` or
`sandbox=false`. This cppnix change fixes `sandbox=relaxed` and
adds a suitable test.

Co-Authored-By: Artemis Tosini <lix@artem.ist>
Co-Authored-By: Eelco Dolstra <edolstra@gmail.com>
Change-Id: I40190f44f3e1d61846df1c7b89677c20a1488522
2024-05-08 19:31:43 +00:00
Qyriad b9be46fb31 remove the autoconf+Make buildsystem
We're not using it anymore. Any leftover bugs in the Meson buildsystem
are now just bugs.

Closes #249.

Change-Id: I0465a0c37ae819f94d40e7829f5bff046aa63d73
2024-05-07 17:04:30 -06:00
Nikodem Rabuliński e8a603fb2f Merge changes Icf26010a,Ib6161567 into main
* changes:
  Always initialize curl in parent process on darwin
  Fix failing darwin tests
2024-05-07 21:26:24 +00:00
Nikodem Rabuliński f894cce79b
Always initialize curl in parent process on darwin
Because of an objc quirk[1], calling curl_global_init for the first time
after fork() will always result in a crash.
Up until now the solution has been to set
OBJC_DISABLE_INITIALIZE_FORK_SAFETY for every nix process to ignore
that error.
This is less than ideal because we were setting it in package.nix,
which meant that running nix tests locally would fail because
that variable was not set.
Instead of working around that error we address it at the core -
by calling curl_global_init inside initLibStore, which should mean
curl will already have been initialized by the time we try to do so in
a forked process.

[1] 01edf1705f/runtime/objc-initialize.mm (L614-L636)

Change-Id: Icf26010a8be655127cc130efb9c77b603a6660d0
2024-05-07 20:43:17 +02:00
eldritch horrors 964ac8b0e8 libutil: de-callback-ify computeClosure
only two users of this function exist. only one used it in a way that
even bears resemblance to asynchronicity, and even that one didn't do
it right. fully async and parallel computation would have only worked
if any getEdgesAsync never calls the continuation it receives itself,
only from more derived callbacks running on other threads. calling it
directly would cause the decoupling promise to be awaited immediately
*on the original thread*, completely negating all nice async effects.

Change-Id: I0aa640950cf327533a32dee410105efdabb448df
2024-05-07 14:35:20 +00:00
eldritch horrors 230860dbb8 libstore: limit CA realisation info substitution concurrency
this seems to be an oversight, considering that regular substitutions
are concurrency-limited. while not particularly necessary at present,
once we've removed the `Callback` based interfaces it will be needed.

Change-Id: Ide2d08169fcc24752cbd07a1d33fb8482f7034f5
2024-05-07 14:35:20 +00:00
Qyriad 005b2b61e6 fix fallback chroot store creation after b247ef72d
When /nix/var (or, more precisely, NIX_STATE_DIR) does not exist at all,
Lix falls back to creating an adhoc chroot store in XDG_DATA_HOME.

b247ef72d[1] changed the way Store classes are initialized, and in the
migration, a `params2` was accidentally changed to `params`. This commit
restores the correct behavior, and in lieu of a single *character* fix,
this commit also changes the variable name to something more reasonable.

Fixes #274.

[1]: b247ef72dc

n.b., this code might deserve some more looking at anyway. this fallback
store creation throws away *all* Store params passed to
openFromNonUri() in favor of an entirely new set which only contains
the `root` param, which may or may not be the correct behavior

Change-Id: Ibea559b88a50e6d6e75a1f87d9d7816cabb2a8f3
2024-05-06 20:54:21 -06:00
eldritch horrors f75d0752ce filetransfer: correctly abort empty transfers
returning 0 from the callback for errors signals successful transfer if
the source returned no data even though the exception we've just caught
clearly disagrees. while this is not all that important (since the only
viable cause of such errors will be dataCallback, and the sole instance
of it being used already takes care of exceptions) we can just do this.

Change-Id: I2bb150eff447121d82e8e3aa4e00057c40523ac6
2024-05-06 21:18:23 +02:00
eldritch horrors 121edecf65 filetransfer: extract decompressor creation
this will be necessary if we want download() to return a source instead
of consuming a sink, which will in turn be needed to remove coroutines.

Change-Id: I34ec241e9bbc5d32fbcd243b244e29c3757533aa
2024-05-06 20:19:03 +02:00
jade 4fd32351f3 Merge "Warn on untrusted client settings being ignored" into main 2024-05-05 19:35:35 +00:00
eldritch horrors 6b08138929 filetransfer: abort transfer on receiver exception
not doing this will cause transfers that had their readers disappear to
linger. with lingering transfers the curl thread can't shut down, which
will cause nix itself to not shut down until the transfer finishes some
other way (most likely network timeouts). also add a new test for this.

Change-Id: Id2401b3ac85731c824db05918d4079125be25b57
2024-05-05 18:09:31 +00:00
jade 2e8f9ac944 Warn on untrusted client settings being ignored
These are such a footgun and trip people up a lot. Let's make Lix louder
about this.

Related: lix-project/lix#261
Change-Id: I6a8d57c9817caaa6b0cbf886c615dda51038f628
2024-05-04 21:30:26 -07:00
jade 9909a175bf Fix /etc/group having desynced IDs from the actual UID in the sandbox
This was found when `logrotate.conf` failed to build in a NixOS system
with:

    /nix/store/26zdl4pyw5qazppj8if5lm8bjzxlc07l-coreutils-9.3/bin/id: cannot find name for group ID 30000

This was surprising because it seemed to mean that /etc/group was busted
in the sandbox. Indeed it was:

    root0:
    nixbld:!💯
    nogroup65534:

We diagnosed this to sandboxUid() being called before
usingUserNamespace() was called, in setting up /etc/group inside the
sandbox. This code desperately needs refactoring.

We also moved the /etc/group code to be with the /etc/passwd code, but
honestly this code is all spaghetti'd all over the place and needs some
more serious tidying than we did here.

We also moved some checks to be earlier to improve locality with where
the things they are checking come from.

Change-Id: Ie29798771f3593c46ec313a32960fa955054aceb
2024-05-04 17:36:50 -07:00
Maximilian Bosch 79d0ae6670 Merge "libstore/local-derivation-goal: prohibit creating setuid/setgid binaries" into main 2024-05-04 07:26:15 +00:00
Maximilian Bosch 045ee37438 libstore/local-derivation-goal: prohibit creating setuid/setgid binaries
With Linux kernel >=6.6 & glibc 2.39 a `fchmodat2(2)` is available that
isn't filtered away by the libseccomp sandbox.

Being able to use this to bypass that restriction has surprising results
for some builds such as lxc[1]:

> With kernel ≥6.6 and glibc 2.39, lxc's install phase uses fchmodat2,
> which slips through 9b88e52846/src/libstore/build/local-derivation-goal.cc (L1650-L1663).
> The fixupPhase then uses fchmodat, which fails.
> With older kernel or glibc, setting the suid bit fails in the
> install phase, which is not treated as fatal, and then the
> fixup phase does not try to set it again.

Please note that there are still ways to bypass this sandbox[2] and this is
mostly a fix for the breaking builds.

This change works by creating a syscall filter for the `fchmodat2`
syscall (number 452 on most systems). The problem is that glibc 2.39
is needed to have the correct syscall number available via
`__NR_fchmodat2` / `__SNR_fchmodat2`, but this flake is still on
nixpkgs 23.11. To have this change everywhere and not dependent on the
glibc this package is built against, I added a header
"fchmodat2-compat.hh" that sets the syscall number based on the
architecture. On most platforms its 452 according to glibc with a few
exceptions:

    $ rg --pcre2 'define __NR_fchmodat2 (?!452)'
    sysdeps/unix/sysv/linux/x86_64/x32/arch-syscall.h
    58:#define __NR_fchmodat2 1073742276

    sysdeps/unix/sysv/linux/mips/mips64/n32/arch-syscall.h
    67:#define __NR_fchmodat2 6452

    sysdeps/unix/sysv/linux/mips/mips64/n64/arch-syscall.h
    62:#define __NR_fchmodat2 5452

    sysdeps/unix/sysv/linux/mips/mips32/arch-syscall.h
    70:#define __NR_fchmodat2 4452

    sysdeps/unix/sysv/linux/alpha/arch-syscall.h
    59:#define __NR_fchmodat2 562

I added a small regression-test to the setuid integration-test that
attempts to set the suid bit on a file using the fchmodat2 syscall.
I confirmed that the test fails without the change in
local-derivation-goal.

Additionally, we require libseccomp 2.5.5 or greater now: as it turns
out, libseccomp maintains an internal syscall table and
validates each rule against it. This means that when using libseccomp
2.5.4 or older, one may pass `452` as syscall number against it, but
since it doesn't exist in the internal structure, `libseccomp` will refuse
to create a filter for that. This happens with nixpkgs-23.11, i.e. on
stable NixOS and when building Lix against the project's flake.

To work around that

* a backport of libseccomp 2.5.5 on upstream nixpkgs has been
  scheduled[3].

* the package now uses libseccomp 2.5.5 on its own already. This is to
  provide a quick fix since the correct fix for 23.11 is still a staging cycle
  away.

We still need the compat header though since `SCMP_SYS(fchmodat2)`
internally transforms this into `__SNR_fchmodat2` which points to
`__NR_fchmodat2` from glibc 2.39, so it wouldn't build on glibc 2.38.
The updated syscall table from libseccomp 2.5.5 is NOT used for that
step, but used later, so we need both, our compat header and their
syscall table 🤷

Relevant PRs in CppNix:

* https://github.com/NixOS/nix/pull/10591
* https://github.com/NixOS/nix/pull/10501

[1] https://github.com/NixOS/nixpkgs/issues/300635#issuecomment-2031073804
[2] https://github.com/NixOS/nixpkgs/issues/300635#issuecomment-2030844251
[3] https://github.com/NixOS/nixpkgs/pull/306070

(cherry picked from commit ba6804518772e6afb403dd55478365d4b863c854)
Change-Id: I6921ab5a363188c6bff617750d00bb517276b7fe
2024-05-03 16:29:06 +02:00
Artemis Tosini ecdbc3b207 libstore: check additionalSandboxProfile
Currently LocalDerivationGoal allows setting `__sandboxProfile`
to add sandbox parameters on Darwin when `sandbox=true`.
This was only supposed to have an effect when `sandbox=relaxed`

Change-Id: Ide44ee82d7e4d6b545285eab26547e7014817d3f
2024-05-03 00:59:18 +00:00
Robert Hensing b7ce11c97d Disallow store path names that are . or .. (plus opt. -)
As discussed in the maintainer meeting on 2024-01-29.

Mainly this is to avoid a situation where the name is parsed and
treated as a file name, mostly to protect users.
.-* and ..-* are also considered invalid because they might strip
on that separator to remove versions. Doesn't really work, but that's
what we decided, and I won't argue with it, because .-* probably
doesn't seem to have a real world application anyway.
We do still permit a 1-character name that's just "-", which still
poses a similar risk in such a situation. We can't start disallowing
trailing -, because a non-zero number of users will need it and we've
seen how annoying and painful such a change is.

What matters most is preventing a situation where . or .. can be
injected, and to just get this done.

(cherry picked from commit f1b4663805a9dbcb1ace64ec110092d17c9155e0)
Change-Id: I900a8509933cee662f888c3c76fa8986b0058839
2024-05-02 19:34:38 +02:00
Robert Hensing 3a058dc4b3 Revert "StorePath: reject names starting with '.'"
This reverts commit 24bda0c7b381e1a017023c6f7cb9661fae8560bd.

(cherry picked from commit 9ddd0f2af8fd95e1380027a70d0aa650ea2fd5e4)
Change-Id: Ideb547e2a8ac911cf39d58d3e0c1553867bdd776
2024-05-02 18:53:51 +02:00
puck 0c831765bd Run all derivation builders inside the sandbox on macOS
This replaces the external sandbox-exec call with direct calls into
libsandbox. This API is technically deprecated and is missing some
prototypes, but all major browsers depend on it, so it is unlikely to
materially change without warning.

This commit also ensures the netrc file is only written if the
derivation is in fact meant to be able to access the internet.

This change commits a sin of not actually actively declaring its
dependency on macOS's libsandbox.dylib; this is due to the dylib
cache in macOS making that explicit dependency unnecessary. In the
future this might become a problem, so this commit marks our sins.

Co-authored-by: Artemis Tosini <lix@artem.ist>
Co-authored-by: Lunaphied <lunaphied@lunaphied.me>
Change-Id: Ia302141a53ce7b0327c1aad86a117b6645fe1189
2024-04-27 14:44:15 -06:00
puck 9229e87347 Fix progress bar on copyPaths
This variable should not be shared between activities.

Change-Id: I4eee89bc7acb320a3972dc3a55bfb087d3a9eb3a
2024-04-27 18:03:15 +00:00
Ilya K 9462c01c3e libstore/ssh: shut
This is just logspam, and we have NIX_SSHOPTS for people that want the logspam.

Change-Id: Ieff71473686f0661f9c53c212f8952dd2c9565c3
2024-04-27 12:05:17 +03:00
Maximilian Bosch 8773439a85 Merge "ssh-ng: Set log-fd for ssh to 4 by default" into main 2024-04-26 18:30:33 +00:00
Artemis Tosini 789aa39576 Merge "gc: Find roots using libproc on Darwin" into main 2024-04-26 17:26:45 +00:00
Maximilian Bosch 104448e75d ssh-ng: Set log-fd for ssh to 4 by default
That's expected by `build-remote` and makes sure that errors are
correctly forwarded to the user. For instance, let's say that the
host-key of `example.org` is unknown and

    nix-build ../nixpkgs -A hello -j0 --builders 'ssh-ng://example.org'

is issued, then you get the following output:

    cannot build on 'ssh-ng://example.org?&': error: failed to start SSH connection to 'example.org'
    Failed to find a machine for remote build!
    derivation: yh46gakxq3kchrbihwxvpn5bmadcw90b-hello-2.12.1.drv
    required (system, features): (x86_64-linux, [])
    2 available machines:
    [...]

The relevant information (`Host key verification failed`) ends up in the
daemon's log, but that's not very obvious considering that the daemon
isn't very chatty normally.

This can be fixed - the same way as its done for legacy-ssh - by passing
fd 4 to the SSH wrapper. Now you'd get the following error:

    cannot build on 'ssh-ng://example.org': error: failed to start SSH connection to 'example.org': Host key verification failed.
    Failed to find a machine for remote build!
    [...]

...and now it's clear what's wrong.

Please note that this is won't end up in the derivation's log.

For previous discussion about this change see
https://github.com/NixOS/nix/pull/7659.

Change-Id: I5790856dbf58e53ea3e63238b015ea06c347cf92
2024-04-26 19:04:06 +02:00
eldritch horrors a1ad4e52a6 filetransfer: don't decompress in curl wrapper itself
only decompress the response once all data has been received (in the
fully buffered case), or at least outside of the curl wrapper itself
(in the receive-to-sink case). unfortunately this means we will have
to duplicate decompression logic for these two cases for time being,
but once the curl wrapper has been rewritten to return a real future
or Source we can deduplicate this logic again. the curl wrapper will
have to turn into a proper Source first and use decompression source
logic which also does not currently exist—only decompression *sinks*

Change-Id: I66bc692f07d9b9e69fe10689ee73a2de8d65e35c
2024-04-26 15:26:37 +00:00