With the release of binutils/ld 2.36, ld swapped to much improved
default settings when producing windows binaries with mingw-w64. One of
these changes was to stop stripping the .reloc section from binaries,
which is required for working ASLR.
.reloc section stripping is something we've accounted for previously,
see #18702. The related upstream discussion is in this thread:
https://sourceware.org/bugzilla/show_bug.cgi?id=19011.
When we switched to using a newer Guix time-machine in #23778, we begun
using binutils 2.37 to produce releases. Since then, our windows
installer (produced with makensis) has not functioned correctly when run on
a Windows system with the "Force randomization for images (Mandatory ASLR)"
option enabled. Note that all of our other release binaries, which all
contain .reloc sections, function fine under the same option, so it
cannot be just the presence of a .reloc section that is the issue.
For now, restore makensis to it's pre-binutils-2.36 behaviour, which
fixes the produced installer. The underlying issue can be further
investigated in future.
Github-Pull: #25788
Rebased-From: 7a0b129c41
c73415bc10 build: Fix xargs warnings for Guix builds (Hennadii Stepanov)
Pull request description:
On master (e3ce019667) there are warnings in `./contrib/guix/guix-build` logs:
```
xargs: warning: options --max-args and --replace/-I/-i are mutually exclusive, ignoring previous --max-args value
```
This PR fixes such warnings.
ACKs for top commit:
prusnak:
utACK c73415b
Tree-SHA512: a7b55f59afbb19b78f795cea64acacf29903cfcd5fd7c37a771b073c1f2ff54555a26f3d00c1c73a8ef588396217ddf598e32b2ae961559042cc051b0aad162a
======================================================================
ERROR: test_revocation_mode_soft (tests.test_validate.ValidateTests)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/tmp/guix-build-python-certvalidator-0.1-1.e5bdb4b.drv-0/source/tests/test_validate.py", line 85, in test_revocation_mode_soft
validate_path(context, path)
File "/tmp/guix-build-python-certvalidator-0.1-1.e5bdb4b.drv-0/source/tests/../certvalidator/validate.py", line 50, in validate_path
return _validate_path(validation_context, path)
File "/tmp/guix-build-python-certvalidator-0.1-1.e5bdb4b.drv-0/source/tests/../certvalidator/validate.py", line 358, in _validate_path
raise PathValidationError(pretty_message(
certvalidator.errors.PathValidationError: The path could not be validated because the end-entity certificate expired 2022-01-14 12:00:00Z
2f356a0ca8 scripted-diff: Drop Darwin version for better maintainability (Hennadii Stepanov)
Pull request description:
After this PR, any macOS tools version bumping in the future will touch fewer files in the repo.
Pointing a Darwin version for the `--host` system does not matter for the following reasons:
- in terms of the resulted binaries, we should only care about the minimum supported macOS version which is a separated parameter in our build system.
- in terms of the build system itself, the usage of the `$(host)` variable is self-consistent enough. Btw `$(host_os)` value already has the version dropped:
```
$ make -C depends --no-print-directory print-host_os HOST=x86_64-apple-darwin19
host_os=darwin
```
ACKs for top commit:
gruve-p:
ACK 2f356a0ca8
promag:
ACK 2f356a0ca8.
fanquake:
ACK 2f356a0ca8
Tree-SHA512: 374896ab0ba02b0d8b4b21431fe963bd213b0d09586e0898c13a4c5fa294c1b693f1b2c92880c245c4157c14217b4825b36522f461930477f4d2a727086ebb2a
a3f61676e8 test: Make more shell scripts verifiable by the `shellcheck` tool (Hennadii Stepanov)
Pull request description:
Some shell scripts from `contrib/guix` and `contrib/shell` are not verifiable by the `shellcheck` tool for the following reasons:
- they have no extension (see 4eccf063b2 from bitcoin/bitcoin#21375)
- they have the `.bash` extension while `.sh` is expected
This PR adds these scripts to the input for the `shellcheck` tool, and it fixes discovered `shellcheck` warnings.
ACKs for top commit:
dongcarl:
Code Review ACK a3f61676e8, this is a good robustness improvement for our shell scripts.
jamesob:
crACK a3f61676e8
Tree-SHA512: 6703f5369d9c04c1a174491f381afa5ec2cc4d37321c1b93615abcdde4dfd3caae82868b699c25b72132d8c8c6f2e9cf24d38eb180ed4d0f0584d8c282e58935
From what I can see the only platform this drops support for is CentOS
7. CentOS 7 reached the end of it's "full update" support at the end of
2020. It does receive maintenance updates until 2024, however I don't
think supporting glibc 2.17 until 2024 is realistic. Note that anyone
wanting to self-compile and target a glibc 2.17 runtime could build with
--disable-threadlocal.
glibc 2.18 was released in August 2013.
https://sourceware.org/legacy-ml/libc-alpha/2013-08/msg00160.html
0f95247246 Integrate univalue into our buildsystem (Cory Fields)
9b49ed656f Squashed 'src/univalue/' changes from 98fadc0909..a44caf65fe (fanquake)
Pull request description:
This PR more tightly integrates building Univalue into our build system. This follows the same approach we use for [LevelDB](https://github.com/bitcoin-core/leveldb/), ([`Makefile.leveldb.include`](https://github.com/bitcoin/bitcoin/blob/master/src/Makefile.leveldb.include)), and [CRC32C](https://github.com/bitcoin-core/crc32c) ([`Makefile.crc32c.include`](https://github.com/bitcoin/bitcoin/blob/master/src/Makefile.crc32c.include)), and will be the same approach we use for [minisketch](https://github.com/sipa/minisketch); see #23114.
This approach yields a number of benefits, including:
* Faster configuration due to one less subconfigure being run during `./configure` i.e 22s with this PR vs 26s
* Faster autoconf i.e 13s with this PR vs 17s
* Improved caching
* No more issues with compiler flags i.e https://github.com/bitcoin/bitcoin/pull/12467
* More direct control means we can build exactly the objects we want
There might be one argument against making this change, which is that builders should have the option to use "proper shared/system libraries". However, I think that falls down for a few reasons. The first being that we already don't support building with a number of system libraries (secp256k1, leveldb, crc32c); some for good reason. Univalue is really the odd one out at the moment.
Note that the only fork of Core I'm aware of, that actively patches in support for using system libs, also explicitly marks them as ["DANGEROUS"](a886811721/configure.ac (L1430)) and ["NOT SUPPORTED"](a886811721/configure.ac (L1312)). So it would seem they exist more to satisfy a distro requirement, as opposed to something that anyone should, or would actually use in practice.
PRs like #22412 highlight the "issue" with us operating with our own Univalue fork, where we actively fix bugs, and make improvements, when upstream (https://github.com/jgarzik/univalue) may not be taking those improvements, and by all accounts, is not currently actively maintained. Bitcoin Core should not be hamstrung into not being able to fix bugs in a library, and/or have to litter our source with "workarounds", i.e #22412, for bugs we've already fixed, based on the fact that an upstream project is not actively being maintained. Allowing builders to use system libs is really only exacerbating this problem, with little benefit to our project. Bitcoin Core is not quite like your average piece of distro packaged software.
There is the potential for us to give the same treatment to libsecp256k1, however it seems doing that is currently less straightforward.
ACKs for top commit:
dongcarl:
ACK 0f95247246 less my comment above, always nice to have an include-able `sources.mk` which makes integration easier.
theuni:
ACK 0f95247246. Thanks fanquake for keeping this going.
Tree-SHA512: a7f2e41ee7cba06ae72388638e86b264eca1b9a8b81c15d1d7b45df960c88c3b91578b4ade020f8cc61d75cf8d16914575f9a78fa4cef9c12be63504ed804b99
I used Guix's values for the powerpc64(le) dynamic linkers, and the
/lib-prefix seems to be a Guix-ism rather than standard. The standard
path for the linker-loaders start with /lib64.
I've taken the new loader values from SYSDEP_KNOWN_INTERPRETER_NAMES in
glibc's sysdeps/unix/sysv/linux/powerpc/ldconfig.h file.
For future reference, loader path values can also be found on glibc's
website: https://sourceware.org/glibc/wiki/ABIList?action=recall&rev=16
This addresses issues like the one in #12467, where some of our compiler flags
end up being dropped during the subconfigure of Univalue. Specifically, we're
still using the compiler-default c++ version rather than forcing c++17.
We can drop the need subconfigure completely in favor of a tighter build
integration, where the sources are listed separately from the build recipes,
so that they may be included directly by upstream projects. This is
similar to the way leveldb build integration works in Core.
Core benefits of this approach include:
- Better caching (for ex. ccache and autoconf)
- No need for a slow subconfigure
- Faster autoconf
- No more missing compile flags
- Compile only the objects needed
There are no benefits to Univalue itself that I can think of. These changes
should be a no-op there, and to downstreams as well until they take advantage
of the new sources.mk.
This also removes the option to use an external univalue to avoid similar ABI
issues with mystery binaries.
Co-authored-by: fanquake <fanquake@gmail.com>
a43b8e9555 build: set OSX_MIN_VERSION to 10.15 (fanquake)
Pull request description:
Taken out of #20744, as splitting up some of the build changes was mentioned [here](https://github.com/bitcoin/bitcoin/pull/22937#discussion_r707303172).
This is required to use `std::filesystem` on macOS, as support for it only landed in the libc++.dylib shipped with 10.15. So if we want to move to using `std::filesystem` for `23.0`, this bump is required.
See also: https://developer.apple.com/documentation/xcode-release-notes/xcode-11-release-notes
> Clang now supports the C++17 \<filesystem\> library for iOS 13, macOS 10.15, watchOS 6, and tvOS 13.
macOS 10.15 was released in October 2019. macOS OS's seem to have a life of about 3 years, so it's possible that 10.14 will become officially unsupported by the end of 2021 and prior to the release of 23.0.
Guix builds:
```bash
bash-5.1# find guix-build-$(git rev-parse --short=12 HEAD)/output/ -type f -print0 | env LC_ALL=C sort -z | xargs -r0 sha256sum
abc8b749be65f1339dcdf44bd1ed6ade2533b8e3b5030ad1dde0ae0cede78136 guix-build-a43b8e955558/output/dist-archive/bitcoin-a43b8e955558.tar.gz
1edcc301eb4c02f3baa379beb8d4c78e661abc24a293813bc9d900cf7255b790 guix-build-a43b8e955558/output/x86_64-apple-darwin19/SHA256SUMS.part
e9dbb5594a664519da778dde9ed861c3f0f631525672e17a67eeda599f16ff44 guix-build-a43b8e955558/output/x86_64-apple-darwin19/bitcoin-a43b8e955558-osx-unsigned.dmg
11b23a17c630dddc7594c25625eea3de42db50f355733b9ce9ade2d8eba3a8f3 guix-build-a43b8e955558/output/x86_64-apple-darwin19/bitcoin-a43b8e955558-osx-unsigned.tar.gz
257ba64a327927f94d9aa0a68da3a2695cf880b3ed1a0113c5a966dcc426eb5e guix-build-a43b8e955558/output/x86_64-apple-darwin19/bitcoin-a43b8e955558-osx64.tar.gz
```
ACKs for top commit:
hebasto:
ACK a43b8e9555
jarolrod:
ACK a43b8e9
Tree-SHA512: 9ac77be7cb56c068578860a3b2b8b7487c9e18b71b14aedd77a9c663f5d4bb19756d551770c02ddd12f1797beea5757b261588e7b67fb53509bb998ee8022369
Previously, if the builder exported $VERSION in their environment (as
past Gitian-building docs told them to), but their HEAD does not
actually point to v$VERSION, their build outputs will differ from those
of other builders.
This is because the contrib/guix/guix-* scripts only ever act on the
current git worktree, and does not try to check out $VERSION if $VERSION
is set in the environment.
Setting $VERSION only makes the scripts pretend like the current
worktree is $VERSION.
This problem was seen in jonatack's attestation for all.SHA256SUMS,
where only his bitcoin-22.0rc3-osx-signed.dmg differed from everyone
else's.
Here is my deduced sequence of events:
1. Aug 27th: He guix-builds 22.0rc3 and uploads his attestations up to
guix.sigs
2. Aug 30th, sometime after POSIX time 1630310848: he pulls the latest
changes from master in the same worktree where he guix-built 22.0rc3
and ends up at 7be143a960
3. Aug 30th, sometime before POSIX time 1630315907: With his worktree
still on 7be143a960, he guix-codesigns. Normally, this would result
in outputs going in guix-build-7be143a960e2, but he had
VERSION=22.0rc3 in his environment, so the guix-* scripts pretended
like he was building 22.0rc3, and used 22.0rc3's guix-build directory
to locate un-codesigned outputs and dump codesigned ones.
However, our SOURCE_DATE_EPOCH defaults to the POSIX time of HEAD
(7be143a960), which made all timestamps in the resulting codesigned
DMG 1630310848, 7be143a960e2's POSIX timestamp. This differs from the
POSIX timestamp of 22.0rc3, which is 1630348517. Note that the
windows codesigning procedure does not consider SOURCE_DATE_EPOCH.
We resolve this by only allowing VERSION overrides via the FORCE_VERSION
environment variable.
The SHA256SUMS file can be used in a sha256sum -c command to verify
downloaded binaries. However users are likely to download just a single
file and not place this file in the correct directory relative to the
SHA256SUMS file for the simple verification command to work. By not
including the directory name in the SHA256SUMS file, it will be easier
for users to verify downloaded binaries.
Co-authored-by: Carl Dong <contact@carldong.me>
90b3e482e9 release: Release with separate SHA256SUMS and sig files (Carl Dong)
Pull request description:
This allows us to:
- remove the rfc4880 EOL hacks, and
- release with a SHA256SUMS.asc file that's a combination of all signer signatures
ACKs for top commit:
achow101:
ACK 90b3e482e9
laanwj:
Concept and code review ACK 90b3e482e9
Tree-SHA512: 5d5086063d303aa0cbd590e5fdf2ae8f555e25f4e43bf67545e33384449b990e94834c711622530ad0eb3dcc83f52746884a5081dadb0acff8dd799cfadafac7
9b313dfef1 guix: Ensure EPOCH_SOURCE_DATE does not include GPG information (Andrew Chow)
43225f0a2a guix: Remove extra \r from all.SHA256SUMS line ending (Andrew Chow)
d080c27066 guix, doc: Add a note that codesigners need to rebuild after tagging (Andrew Chow)
4a466388a0 guix: Allow changing the base manifest in guix-verify (Andrew Chow)
33455c7696 guix: Make all.SHA256SUMS rather than codesigned.SHA256SUMS (Andrew Chow)
Pull request description:
`guix-verify` expects `all.SHA256SUMS` but `guix-attest` produces `codesigned.SHA256SUMS`. Since `all.SHA256SUMS` makes more sense (as the file contains all the sha256sums, not just the codesigned ones), `guix-attest` has been changed to output a file of that name.
As a quality of life improvement, `guix-verify` can take `SIGNER` and use the signer's manifest as the base to compare against. This makes it easier to compare a single person's attestations with everyone else's and can make it more obvious when one builder is clearly mismatching with everyone else.
Lastly `release-process.md` is updated with a note about a gotcha that can cause a mismatch in the codesigned attestation.
ACKs for top commit:
fanquake:
ACK 9b313dfef1
Tree-SHA512: 0d60627def38288dbd3059ad1e72cad224f9205da11b1a561c082ef28250a074df5cc5f2797c91a7be027bc486a3fda3319c2e496a8724e5b539337236c6f990
If the user has set log.showSignature=true in their git config, then the
git log will always output GPG signature information. Since git log is
used to set EPOCH_SOURCE_DATE, this will mistakenly have GPG signature
information in it which causes issues for the build. To avoid this
issue, we override the config and force log.showSignature=false.
guix-attest mistakenly added an extra \r to the line endings in
all.SHA256SUMS, causing guix-verify to erroneously fail.
Co-Authored-By: Carl Dong <contact@carldong.me>
