f2b5b0a3b48f9241971c14238040048f8b630811 build: add linker optimization flags to guix (fanquake)
b8b050a8d642e38c81d1e451750c2c8db92fee5e build: add linker optimization flags to gitian descriptors (fanquake)
Pull request description:
This PR adds `-Wl,O2` to our gitian and guix LDFLAGS. This makes the linker perform certain optimisations (and is different from LTO).
Any -O argument will enable optimizations in GNU ld. We can use -O2 here, as this matches our compile flags. Note that this would also enable additional optimizations if using the lld or gold linkers, when compared to -O0.
A nice writeup + diagrams of some of these optimizations is available here: http://lwn.net/Articles/192624/.
#### master
```bash
# bitcoind
Histogram for `.gnu.hash' bucket list length (total of 3 buckets)
Length Number % of total Coverage
0 1 ( 33.3%) 0.0%
1 0 ( 0.0%) 0.0%
2 1 ( 33.3%) 40.0%
3 1 ( 33.3%) 100.0%
```
```bash
# bitcoin-qt
Histogram for `.gnu.hash' bucket list length (total of 3 buckets)
Length Number % of total Coverage
0 0 ( 0.0%) 0.0%
1 1 ( 33.3%) 10.0%
2 0 ( 0.0%) 10.0%
3 0 ( 0.0%) 10.0%
4 1 ( 33.3%) 50.0%
5 1 ( 33.3%) 100.0%
```
#### this PR:
```bash
# bitcoind
Histogram for `.gnu.hash' bucket list length (total of 8 buckets)
Length Number % of total Coverage
0 3 ( 37.5%) 0.0%
1 5 ( 62.5%) 100.0%
```
```bash
# bitcoin-qt
Histogram for `.gnu.hash' bucket list length (total of 19 buckets)
Length Number % of total Coverage
0 9 ( 47.4%) 0.0%
1 10 ( 52.6%) 100.0%
```
#### GNU ld -O
> If level is a numeric values greater than zero ld optimizes the output. This might take significantly longer and therefore probably should only be enabled for the final binary. At the moment this option only affects ELF shared library generation. Future releases of the linker may make more use of this option. Also currently there is no difference in the linker’s behaviour for different non-zero values of this option. Again this may change with future releases.
#### lld -O
> Optimize output file size
ACKs for top commit:
dongcarl:
ACK f2b5b0a3b48f9241971c14238040048f8b630811
laanwj:
ACK f2b5b0a3b48f9241971c14238040048f8b630811
Tree-SHA512: e53f3a4338317dbec65d3a93b57b5a6204aabdf9ac82d99447847a3c8627facc53c58c2cf947376f13edd979fc8129a80f18d9ebeccd191a576c83f1dad5c513
c491368d8cfddf3a5b6d574f10ed67492fcecbed scripts: add MACHO dylib checking to symbol-check.py (fanquake)
76bf97213f4b153dd3ccf1314088a73c4804601d scripts: fix check-symbols & check-security argument passing (fanquake)
Pull request description:
Based on #17857.
This adds dynamic library checks for MACHO executables to symbol-check.py. The script has been modified to function more like `security-check.py`. The error output is now also slightly different. i.e:
```bash
# Linux x86
bitcoin-cli: symbol operator new[](unsigned long) from unsupported version GLIBCXX_3.4
bitcoin-cli: export of symbol vtable for std::basic_ios<char, std::char_traits<char> > not allowed
bitcoin-cli: NEEDED library libstdc++.so.6 is not allowed
bitcoin-cli: failed IMPORTED_SYMBOLS EXPORTED_SYMBOLS LIBRARY_DEPENDENCIES
# RISCV (skips exported symbols checks)
bitcoin-tx: symbol operator new[](unsigned long) from unsupported version GLIBCXX_3.4
bitcoin-tx: NEEDED library libstdc++.so.6 is not allowed
bitcoin-tx: failed IMPORTED_SYMBOLS LIBRARY_DEPENDENCIES
# macOS
Checking macOS dynamic libraries...
libboost_filesystem.dylib is not in ALLOWED_LIBRARIES!
bitcoind: failed DYNAMIC_LIBRARIES
```
Compared to `v0.19.0.1` the macOS allowed dylibs has been slimmed down somewhat:
```diff
src/qt/bitcoin-qt:
/usr/lib/libSystem.B.dylib
-/System/Library/Frameworks/DiskArbitration.framework/Versions/A/DiskArbitration
/System/Library/Frameworks/IOKit.framework/Versions/A/IOKit
/System/Library/Frameworks/Foundation.framework/Versions/C/Foundation
/System/Library/Frameworks/CoreServices.framework/Versions/A/CoreServices
/System/Library/Frameworks/AppKit.framework/Versions/C/AppKit
/System/Library/Frameworks/ApplicationServices.framework/Versions/A/ApplicationServices
/System/Library/Frameworks/CoreFoundation.framework/Versions/A/CoreFoundation
-/System/Library/Frameworks/Security.framework/Versions/A/Security
-/System/Library/Frameworks/SystemConfiguration.framework/Versions/A/SystemConfiguration
/System/Library/Frameworks/CoreGraphics.framework/Versions/A/CoreGraphics
-/System/Library/Frameworks/OpenGL.framework/Versions/A/OpenGL
-/System/Library/Frameworks/AGL.framework/Versions/A/AGL
/System/Library/Frameworks/Carbon.framework/Versions/A/Carbon
/usr/lib/libc++.1.dylib
-/System/Library/Frameworks/CFNetwork.framework/Versions/A/CFNetwork
/System/Library/Frameworks/CoreText.framework/Versions/A/CoreText
/System/Library/Frameworks/ImageIO.framework/Versions/A/ImageIO
/usr/lib/libobjc.A.dylib
```
ACKs for top commit:
laanwj:
ACK c491368d8cfddf3a5b6d574f10ed67492fcecbed
Tree-SHA512: f8624e4964e80b3e0d34e8d3cc33f3107938f3ef7a01c07828f09b902b5ea31a53c50f9be03576e1896ed832cf2c399e03a7943a4f537a1e1c705f3804aed979
683d197970a533690ca1bd4d06d021900e87cb8b Use latest signapple commit (Andrew Chow)
Pull request description:
Update gitian and guix to use the same latest signapple commit.
Also changed guix to use the actual repo. The changes from the fork were incorporated upstream.
ACKs for top commit:
fanquake:
ACK 683d197970a533690ca1bd4d06d021900e87cb8b - sanity checked that the updated package is built:
Tree-SHA512: a4981f8bbe33e6c5654632bc9b9f6f2f1e675741a19ac7296205e370f1e64a747101ecb632e0cc82a0134e4c2e9ce47b3f7b4d8c8f75f0f06dd069c078303759
2c403279e2f0f7c8c27c56d4e7b0573c59571f0a gitian: Remove codesign_allocate and pagestuff from MacOS build (Andrew Chow)
f55eed251488d70d5e2e3a2965a4f8ec0c476853 gitian: use signapple to create the MacOS code signature (Andrew Chow)
95b06d21852b28712db6c710e420a58bdc1a0944 gitian: use signapple to apply the MacOS code signature (Andrew Chow)
42bb1ea363286b088257cabccb686ef1887c1d3b gitian: install signapple in gitian-osx-signer.yml (Andrew Chow)
Pull request description:
The MacOS code signing issues that were encountered during the 0.21.0 release cycle have shown that it is necessary for us to use a code signing tool for which the source code is available and modifiable by us. Given that there appears to not be such a tool available, I have written such a tool, [signapple](https://github.com/achow101/signapple), that we can use. This tool is able to create a valid MacOS code signature, detach it in a way that we were doing previously, and attach it to the unsigned binary. This tool can also verify that the signature is correct.
This PR implements the usage of that tool in the gitian build for the code signed MacOS binary. The code signer will use this tool to create the detached signature. Gitian builders will use this tool to apply the detached signature. The `gitian-osx-signer.yml` descriptor has been modified to install this tool so that the `detached-sig-apply.sh` script can use it. Additionally, the `codesign_allocate` and `pagestuff` tools are no longer necessary so they are no longer added to the tarball used in code signing. Lastly, both the `detached-sig-create.sh` and `detached-sig-apply.sh` scripts are made to be significantly less complex and to not do unexpected things such as unpacking an already unpacked tarball.
The detached code signature that signapple creates is almost identical to that which we were previously creating. The only difference is that the cpu architecture name is included in the extension (e.g. we have `bitcoin-qt.x86_64sign` instead of `bitcoin-qt.sign`). This was done in order to support signing universal binaries which we may want to do in the future. However signapple can still apply existing code signatures as it will accept the `.sign` extension. If it is desired, it can be modified to produce signatures with just the `.sign` extension. However I do not think it is necessary to maintain compatibility with the old process.
ACKs for top commit:
laanwj:
Code review ACK 2c403279e2f0f7c8c27c56d4e7b0573c59571f0a
Tree-SHA512: 2a0e01e9133f8859b9de26e7e8fe1d2610d2cbdee2845e6008b12c083c7e3622cbb2d9b83c50a269e2c3074ab95914a8225d3cd4108017f58b77a62bf10951e0
7c9e821c4e6cb186208ead9c8df616d1f393a49a scripts: add MACHO NOUNDEFS check to security-check.py (fanquake)
4ca92dc6d3f3e487d63286d8871d1829b3d279ff scripts: add MACHO PIE check to security-check.py (fanquake)
Pull request description:
This uses `otool -vh` to print the mach header and look for the `PIE` flag:
```bash
otool -vh src/bitcoind
Mach header
magic cputype cpusubtype caps filetype ncmds sizeofcmds flags
MH_MAGIC_64 X86_64 ALL LIB64 EXECUTE 24 2544 NOUNDEFS DYLDLINK TWOLEVEL WEAK_DEFINES BINDS_TO_WEAK PIE
```
From [`mach-o/loader.h`](https://opensource.apple.com/source/cctools/cctools-927.0.2/include/mach-o/loader.h.auto.html):
```c
#define MH_PIE 0x200000 /* When this bit is set, the OS will
load the main executable at a
random address. Only used in
MH_EXECUTE filetypes. */
```
ACKs for top commit:
laanwj:
code review ACK 7c9e821c4e6cb186208ead9c8df616d1f393a49a
Tree-SHA512: 5ba2f60440d0e31c70371a355c91ca4f723d80f7287d04e2098bf5b11892cc74216ff8f1454603c4db9675d4f7983614843b992b8dcfca0309aadf2aa7ab2e4b
8f7b93047581c67f2133cdb8c7845471de66c30f Drop the leading 0 from the version number (Andrew Chow)
Pull request description:
Removes the leading 0 from the version number. The minor version, which we had been using as the major version, is now the major version. The revision, which we had been using as the minor version, is now the minor version. The revision number is dropped. The build number is promoted to being part of the version number. This also avoids issues where it was accidentally not included in the version number.
The CLIENT_VERSION remains the same format as previous as previously, as the Major version was 0 so it never actually got included in it.
The user agent string formatter is updated to follow this new versioning.
***
Honestly I'm just tired of all of the people asking for "1.0" that maybe this'll shut them up. Skip the whole 1.0 thing and go straight to version 22.0!
Also, this means that the terminology we commonly use lines up with how the variables are named. So major versions are actually bumping the major version number, etc.
ACKs for top commit:
jnewbery:
Code review ACK 8f7b930475
MarcoFalke:
review ACK 8f7b93047581c67f2133cdb8c7845471de66c30f 🎻
Tree-SHA512: b5c3fae14d4c0a9c0ab3b1db7c949ecc0ac3537646306b13d98dd0efc17c489cdd16d43f0a24aaa28e9c4a92ea360500e05480a335b03f9fb308010cdd93a436
2ecaf214331b506ebfac4f4922241744357d652b gitian: remove execstack workaround for ricv64 & powerpc64le (fanquake)
5baff2b31840bdbc465f55b875aa6e9480288215 build: use focal in gitian descriptors (fanquake)
Pull request description:
This PR changes the gitian descriptors to use Ubuntu Focal (20.04), over Bionic (18.04), moving from GCC 7.5 to GCC 8.4 for native Linux builds, mingw-w64 GCC 7.3 to mingw-w64 GCC 9.3 for Windows builds, while continuing to use GCC 8.4 for all cross builds and Clang 8.0.0 for macOS builds.
It also drops the `-Wl,-z,noexecstack` workaround we've been using for the riscv64 and powerpc64le hosts, as it's no-longer needed. One new package is installed in the osx build, `libtinfo5`, as libtinfo5.so is required by our downloaded Clang 8.
A bump to Focal will at least be required if we want to update to a newer Qt (5.15, #19716) for 22.0, as we need a newer version of [`g++-mingw-w64`](https://packages.ubuntu.com/focal/g++-mingw-w64-x86-64) and the [`mingw-w64`](https://mingw-w64.org/doku.php) headers. This can still be done while continuing to use GCC 8.4 for Linux builds (see below), however the newer `g++-mingw-w64` will be based off of GCC 9.3.
**Some considerations**
GCC 9 is affected by #20005 "memcmp with constants that contain zero bytes are broken in GCC", and the newer `g++-mingw-w64` will be based off of GCC 9.3.
The `--no-*` variants of the Windows linker flags (i.e `--no-dynamicbase`) we use to [test our `security-check.py` script](16b784d953/contrib/devtools/test-security-check.py (L53)) are not patched into the mingw binutils in Focal (they have been re-added in Groovy (20.10)). This isn't currently an issue, however, we might add a call to `test-security-check` for Guix (#20980), and if we wanted to do the same for gitian, it would not work. Note how it's quite "easy" for us to apply the `--no-*` variant patch to our Guix build; it would be quite a bit harder to do in Gitian.
Gitian Builds @ 2ecaf214331b506ebfac4f4922241744357d652b
#### Linux
```bash
8882ea78486fbae4fac574b9089eb1107c6372d0dd7dfcda4f0f930576f9d6c1 bitcoin-2ecaf214331b-aarch64-linux-gnu-debug.tar.gz
50a9e30943b4eee5163edff3331241e745ff32a2c4463c21a6fdc5986e2d0383 bitcoin-2ecaf214331b-aarch64-linux-gnu.tar.gz
ec4e55a447fddf033fee33cd5f22bfeda3c3612f059194bcf6238859f7989d7a bitcoin-2ecaf214331b-arm-linux-gnueabihf-debug.tar.gz
444fe1b3b933c00bcbd4a9d86888cff3b61c1215b1debccd2843e842d1224777 bitcoin-2ecaf214331b-arm-linux-gnueabihf.tar.gz
88e486ff465980dc1a4aab9687d142ec6f727ed2c52cf539f69db2877dee83b2 bitcoin-2ecaf214331b-powerpc64-linux-gnu-debug.tar.gz
66144ac264c65cada9d86446e6026c85b04fb88198b8f41b42840f6031db3e6c bitcoin-2ecaf214331b-powerpc64-linux-gnu.tar.gz
34bcc13d78d929d575e34e77a6672f23ca7ea23230b28ec2eed563889352ba86 bitcoin-2ecaf214331b-powerpc64le-linux-gnu-debug.tar.gz
b4c5f959664f3063df4330edfe343c17120eb6b556ee1c15c4aeb2c1c54ffd49 bitcoin-2ecaf214331b-powerpc64le-linux-gnu.tar.gz
918fa72ab6f6ebce4e9663c93f72fe26651c260477cbb54749f7eb61438b5cc1 bitcoin-2ecaf214331b-riscv64-linux-gnu-debug.tar.gz
f704f9f8c053ffe37d854e2e81e0f4c0614c435dad7f5d82518c681b73a76ae6 bitcoin-2ecaf214331b-riscv64-linux-gnu.tar.gz
b59e3a62f1df9d79f30e916b3c9655f654036fe3a420040c53acc8dd9f4162c5 bitcoin-2ecaf214331b-x86_64-linux-gnu-debug.tar.gz
a4dc9ca877cc97544e65db11be38406d16f15d74fcdcd2318bb92474729bc60d bitcoin-2ecaf214331b-x86_64-linux-gnu.tar.gz
b40ba2d5da498330ade92a4ccebcceb1452b94c8ffeacb336f87e93b5c88d8af src/bitcoin-2ecaf214331b.tar.gz
af6ebc91147778e4e6705eade62608dde4d6e60522d79087fa9129bdb7c01199 bitcoin-core-linux-22-res.yml
```
#### Windows
```bash
121a3970a6911cb8c453b2ce37d03f6cbb43333e29db8fa516c68563fb367f43 bitcoin-2ecaf214331b-win-unsigned.tar.gz
6294e9efebe935092f9ba119dc60ad4094f18b51c4181324e54d3057524d6101 bitcoin-2ecaf214331b-win64-debug.zip
5b5a236b63e67f5f6c07ad9aa716aa7b72fb63722c96798b332c6d164738f9cf bitcoin-2ecaf214331b-win64-setup-unsigned.exe
c1fa5894c5e02a201637567c80b9bde9024f44673dcd06fd4d489c1709179279 bitcoin-2ecaf214331b-win64.zip
b40ba2d5da498330ade92a4ccebcceb1452b94c8ffeacb336f87e93b5c88d8af src/bitcoin-2ecaf214331b.tar.gz
665fd7eb61aed368150db58a254f15fb5efb51a4efa5abcc52571cb7a1a5de22 bitcoin-core-win-22-res.yml
```
#### macOS
```bash
6a1deae7662aa782baa82a42590f862c6bcdc4f4e38daa9b8c2a9eed1fbb5397 bitcoin-2ecaf214331b-osx-unsigned.dmg
1ee843266e84928a4323fa255c833528c2617a2c9fd2f98fb26ba19bbfc1227b bitcoin-2ecaf214331b-osx-unsigned.tar.gz
097b64dadc167d8e5b733421bf1541a40760ad952990f7cf3f35adc6ae2616d0 bitcoin-2ecaf214331b-osx64.tar.gz
b40ba2d5da498330ade92a4ccebcceb1452b94c8ffeacb336f87e93b5c88d8af src/bitcoin-2ecaf214331b.tar.gz
6e378fb543928e40c7119b96be6ff773d38506a9a888f8b02c7f1b8a0801a80e bitcoin-core-osx-22-res.yml
```
ACKs for top commit:
laanwj:
Build script changes review ACK 2ecaf214331b506ebfac4f4922241744357d652b
Tree-SHA512: 975d5830b787d2e08988f43cbc6e839294171c1d94c8219636308b05f9b77041421612ae67be24a631674670cfc9c2d96d8177f2b3158a78fc3deea19631febf
faa2f06f5eaf8578873495f44603ee74d7a1abf4 scripted-diff: [build] Ensure source tarball has leading directory name (MarcoFalke)
Pull request description:
This has been fixed in 0.20, so it needs to be fixed on master as well to avoid a regression
#18945
ACKs for top commit:
laanwj:
ACK faa2f06f5eaf8578873495f44603ee74d7a1abf4
hebasto:
ACK faa2f06f5eaf8578873495f44603ee74d7a1abf4, tested gitian builds only.
promag:
ACK faa2f06f5eaf8578873495f44603ee74d7a1abf4.
Tree-SHA512: e3b025c29c45b025002abc35262bb5d771f6cbd807f1c256c477c243685e93cd43ad9f642b38e3cf218590912abe6ea0ddfec3bfbef36f99080aad74ed6cc0af
33a84e8f405ed6dd8885419cef305b4e6c7a428a build: Update and sort package list in gitian-linux.yml (Hennadii Stepanov)
95051682bedc2ef1076af5ede5bd56ca243279e3 build: Drop old hack which is unneeded now (Hennadii Stepanov)
Pull request description:
The hack was aimed to fix an issue in Ubuntu Trusty 14.04 (see #8188).
The current hack implementation was added in #8315.
On master (8db23349fe9b512e6801d59d17052c5a7a1c64df) this hack is effectively noop, and it is no longer needed.
I see this PR as a step to removing `libfaketime` from gitian builds.
ACKs for top commit:
dongcarl:
tACK 33a84e8f405e
laanwj:
Code review ACK 33a84e8f405ed6dd8885419cef305b4e6c7a428a
Tree-SHA512: 90036c555a500649ccc3d108bf11f09a9cfd2c92c0b598f7e0c0df63a713ae7abaf78f350b68c025470619c967223f45f6a235ad37a6ce1d1a0341ed34963ba0
bfe1ba2f5b36056e0c41edf8206b93d3d83098df rel-builds: Specify core.abbrev for git-rev-parse (Carl Dong)
27e63e01cce368d67092de8f0c736927d6f6aa69 build: Accomodate makensis v2.x (Carl Dong)
1f2c39a30e0f82046c7aecddfda3eb99cb536816 guix: Remove logical cores requirement (Carl Dong)
a4f6ffa71e335d4b2a6bf525b7f416968f9cd9f7 lint: Also enable source statements for non-gitian (Carl Dong)
d256f91cb1b0d6ff5170106b99b0266cbe51f5a2 rel-builds: Directly deploy win installer to OUTDIR (Carl Dong)
fa791da02f9684e3fd554b687fb692ae6a23d65a nsis: Specify OutFile path only once (Carl Dong)
14701604d0904bc5bbf1c67de08f8ee6d3215523 guix: Expose GIT_COMMON_DIR in container as readonly (Carl Dong)
f5a6ac4f48b18f93050d77bcb23f9cf45ec34647 guix: Make source tarball using git-archive (Carl Dong)
395c1137f630dc495ffb2752a23bc1dfd470ee53 gitian: Limit sourced script to just assignments (Carl Dong)
Pull request description:
Based on: #18556
Related: https://github.com/bitcoin/bitcoin/pull/17595#discussion_r399728721
ACKs for top commit:
fanquake:
ACK bfe1ba2f5b36056e0c41edf8206b93d3d83098df - I agree with Carl, and am going to merge this. I'd like for Linux Guix builds to be working again, and we can rebase #18818.
Tree-SHA512: c87ada7e3de17ca0b692a91029b86573442ded5780fc081c214773f6b374a0cdbeaf6f6898c36669c2e247ee32aa7f82defb1180f8decac52c65f0c140f18674
2aa48edec0101f8a77a2189244fc62722ff7a123 refactor: Drop unused ${WRAP_DIR}/${HOST} directory (Hennadii Stepanov)
1362be044724bb49d785ca2e296a3b43343c1690 build: Drop make dist in gitian builds (Hennadii Stepanov)
Pull request description:
After the merge of #18331, the packaged source tarball is created by `git archive`, but the binaries are built from another one which is made by `make dist`.
With this PR the only source tarball, created by `git archive`, is used both for binaries building and for packaging to users.
Close#16588.
Close#18547.
As a good side-effect, #18349 becomes redundant.
**Change in behavior**
The following variables 1b151e3ffc/configure.ac (L2-L6)
are no longer used for naming of directories and tarballs.
Instead of them the gitian descriptors use a git tag (if available) or a commit hash.
---
Also a small refactor commit picked from #18404.
ACKs for top commit:
dongcarl:
ACK 2aa48edec0101f8a77a2189244fc62722ff7a123
MarcoFalke:
ACK 2aa48edec0101f8a77a2189244fc62722ff7a123
fanquake:
ACK 2aa48edec0101f8a77a2189244fc62722ff7a123 - I've had a quick look over this, and don't want to block merging if this actually gets as closer to finally having this all sorted out. Obviously we've still got #18741, and after speaking to Carl this morning, there will likely be even more changes after that (not Guix specific).
Tree-SHA512: d3b16f87e48d1790a3264940c28acd5d881bfd10f3ce94fb0c8a6af76d8039289d01e0cd4972adac49ae24362857251f6c1e5e09e3e9fbf636c10708b4015a7c
3b3b93174a216961f935b63b03732fbc2821ff92 nsis: Write to correct filename in first place (Carl Dong)
Pull request description:
Per MarcoFalke's suggestion here https://github.com/bitcoin/bitcoin/pull/17029#discussion_r333216722
ACKs for top commit:
MarcoFalke:
unsigned ACK 3b3b93174a216961f935b63b03732fbc2821ff92, makes sense to name it that way because it will raise the "unsinged" error in Windows
Tree-SHA512: da72aae438505e162d0b3cd27d873b7ad8176178bb459a738e61b6e2ad0fa739d905b3109fab641bb1a3950fe59ad526c5568d12cf48a305166cdb7db6686543
e4d366788bc2e8dce8e6ca572fce08d913d15d6b build: Drop needless EXTRA_DIST content (Hennadii Stepanov)
6c4da59f5b5b3c40526d38965d4ffa7fd59f2ebc build: Drop SOURCEDIST reordering (Hennadii Stepanov)
5e6b8b391243016cb06e9e107c2e6a13a744b31e build: Use git archive as source tarball (Hennadii Stepanov)
Pull request description:
This PR:
- is an alternative to #17104
- closes#16734
- closes#6753
The idea is clear described by some developers:
- [MarcoFalke](https://github.com/bitcoin/bitcoin/pull/17097#issuecomment-540691850):
> This whole concept of explicitly listing each and every file manually (or with a fragile wildcard) is an obvious sisyphean task. I'd say all we need to do is run git archive and be done with it forever, see #16734, #6753, #11530 ...
- [laanwj](https://github.com/bitcoin/bitcoin/pull/17097#issuecomment-540706025):
> I agree, I've never been a fan of it. I don't think we have any files in the git repository we don't want to ship in the source tarball.
---
The suggested changes have a downside which is pointed by [**luke-jr**](https://github.com/bitcoin/bitcoin/pull/17104#issuecomment-540828045):
> ... but the distfile needs to include autogen-generated files.
This means that a user is not able to run `./configure && make` right away. One must run `./autogen.sh` at first.
Here are opinions about mandatory use of `./autogen.sh`:
- [ryanofsky](https://github.com/bitcoin/bitcoin/issues/16734#issuecomment-534139356):
> It's probably ok to require autogen. I think historically configure scripts were supposed to work on obscure unix systems that would just have a generic shell + make tool + c compiler, and not necessarily need gnu packages like m4 which are needed for autogen.
- [laanwj](https://github.com/bitcoin/bitcoin/issues/16734#issuecomment-540729483):
> I also think it's fine to require autogen. What is one dependency more, if you're building from source.
---
~Also this PR provides Windows users with ZIP archives of the sources. Additionally the commit ID is stored in these ZIP files as a file comment:~
---
Note for reviewers: please verify is `git archive` output deterministic?
ACKs for top commit:
MarcoFalke:
re-ACK e4d366788bc2e8dce8e6ca572fce08d913d15d6b, only change is adding two dots in a the path 🛳
laanwj:
ACK e4d366788bc2e8dce8e6ca572fce08d913d15d6b
Tree-SHA512: d1153d3ca4a580696019b92be3555ab004d197d9a2146aacff9d3150eb7093b7d40eebd6eea12d861d93ff62d62b68706e04e64dbe5ea796ff6757486e462193
fae9084ac5b10f94bdee54853d307838c4254e9c build: Skip i686 build by default in guix and gitian (MarcoFalke)
fa55a2554c2661b8f2a759044d5ac85c9979d9ca depends: Remove reference to win32 (MarcoFalke)
Pull request description:
Closes#17504
Now that we no longer provide downloads for i686 on our website (https://bitcoincore.org/en/download/), there is no need to build them by default.
i686 can still be built in depends (tested by ci/travis) and in guix/gitian by setting the appropriate `HOSTS`.
ACKs for top commit:
practicalswift:
ACK fae9084ac5b10f94bdee54853d307838c4254e9c -- patch looks correct
dongcarl:
ACK fae9084ac5b10f94bdee54853d307838c4254e9c patch looks correct
laanwj:
Code review ACK fae9084ac5b10f94bdee54853d307838c4254e9c
hebasto:
ACK fae9084ac5b10f94bdee54853d307838c4254e9c, I have reviewed the code and it looks OK, I agree it can be merged.
Tree-SHA512: b000c19a2cd2a596a52028fa298c4022c24cfdfc1bdb3795a90916d0a00a32e4dd22278db93790b6a11724e08ea8451f4f05c77bc40d1664518e11a8c82d6e29
fae75306bac4c82dd07a1b85ce5dfb020e052fe8 scripted-diff: Set gitian arch back to amd64 (MarcoFalke)
Pull request description:
This was required to allow gitian builds on non-amd64 architecture, however, it seems to break the current builds (with lxc), see https://github.com/bitcoin/bitcoin/pull/17409#issuecomment-554099626
Also, the gititan builds wouldn't be deterministic across arches anyway, see #17468
So instead of wasting more time on this, revert the change and hope that guix allows to compile on non-amd64 architectures.
Top commit has no ACKs.
Tree-SHA512: 801e9a30ae1b0882ef45d5eb3a3cf80f3ace3b99db046069dbd95b6162119e977e3cf3134287d1ac5d09483906206acc71e1ac34d6b74dbc533d46aaf73f5cc2
333362991c5219e6d9ede5fa54328bcbb0277cf2 doc: Explain $LIB in LD_PRELOAD in gitian descriptors (MarcoFalke)
fab9850ef4a76e4f09278909a2865e5175661653 scripted-diff: Avoid hardcoded libfaketime dir in gitian (MarcoFalke)
Pull request description:
Without this gitian prints warnings for me:
```
ERROR: ld.so: object '/usr/lib/x86_64-linux-gnu/faketime/libfaketime.so.1' from LD_PRELOAD cannot be preloaded (cannot open shared object file): ignored.
```
```
$ ls /usr/lib/aarch64-linux-gnu/faketime/libfaketime.so.1
/usr/lib/aarch64-linux-gnu/faketime/libfaketime.so.1
```
ACKs for top commit:
laanwj:
ACK 333362991c5219e6d9ede5fa54328bcbb0277cf2
Tree-SHA512: 3e7c4471b69c2ae38c29d0cc0db8b9eae0912085299d7f5ac67eeb4b6a2fdc7eb23d806eeeae0b0c2da22d6d1ba82513cab23652876b97aada9928b2c7d38e7e
9d1f971c7554bd5eb80792a1b9a692e0145ed516 gitian: Put things in the right place to begin with (Carl Dong)
71949a97a7d050d4cd511731062a9ffe471ac438 gitian: Eliminate rename dependency (Carl Dong)
999a9a5f5b4165dbdcb032c8d58c0ac1085ddcd1 gitian: Smaller diff with gitian-linux.yml (Carl Dong)
c4a3c25ba11b20871699e570fc7b98a7b8472503 gitian: Fix README inclusion in archives (Carl Dong)
93cb974980c578c459c15fd28e565c9722a1627d gitian: Use split-debug.sh for Win builds (Carl Dong)
Pull request description:
It would seem that our `gitian-win.yml` has not been keeping up with `gitian-linux.yml`, this PR:
1. Minimizes the diff size between `gitian-{win,linux}.yml`
2. Eliminates the `rename` dependency
ACKs for top commit:
laanwj:
ACK 9d1f971c7554bd5eb80792a1b9a692e0145ed516
Tree-SHA512: 84ed47c685e12d0064c02811907ae3d0fd3c47db8773d497dcc38f0defbfb3040fd82899fb026cf355f229b906d05a1c8038a95642bb90d044afbc2e0b239af2
993aa414d3 scripted-diff: gitian: Use REFERENCE_DATETIME directly. (Carl Dong)
Pull request description:
Fixes regression introduced by #16141.
```
-BEGIN VERIFY SCRIPT-
sed -i 's#\$REFERENCE_DATE\\\\\\ \$REFERENCE_TIME#\$REFERENCE_DATETIME#g' contrib/gitian-descriptors/*
-END VERIFY SCRIPT-
```
-----
Note that this could have been fixed by escaping properly, but using `REFERENCE_DATETIME` directly is simpler.
Future note: `REFERENCE_{DATE{,DATETIME},TIME}` is a bit ridiculous. At the very _least_ gitian should use epoch, as it is the most parse-able, and preferably set SOURCE_DATE_EPOCH.
ACKs for commit 993aa4:
Tree-SHA512: 8457e5fffde66e1d2b846547b6807416b884c171f63569f76dfefd498d2a58ad6f9eb93931eb6cfc7ff38c6b460b0c488ca87d1a68bc630c48f365a74b6ee163
* lint: Skip shell linting if gawk is not installed
* lint: Skip Gitian descriptor scripts checking if jq is not installed
* ci: Install gawk and jq
`yq` requires `jq`
* Fix shellcheck warnings
bd3f5a90ecd6de40516141b23b0861dbba0b31b6 build: remove mingw linker workaround from win gitian descriptor (fanquake)
Pull request description:
This workaround was added as part of the switch to gitian building using Ubuntu 14.04 (#6900).
However, it should no longer be required, as we have switched to Bionic (#13171) and that
has a far newer version of binutils.
Original discussion: https://github.com/bitcoin/bitcoin/pull/6900
binutils patch: https://sourceware.org/bugzilla/show_bug.cgi?id=16192
ACKs for top commit:
MarcoFalke:
ACK bd3f5a90ecd6de40516141b23b0861dbba0b31b6
theuni:
ACK bd3f5a90ecd6de40516141b23b0861dbba0b31b6
laanwj:
ACK bd3f5a90ecd6de40516141b23b0861dbba0b31b6
Tree-SHA512: 01a5789994decf8cdedf7aaa0a449d2100a77e2e6b422d6b9dd5a4ac3e2e0b538c3d43aae4a1c3713614782f3c6b09d8d8bb21c20e86ce3c1734183dedd02d0c
b8705a091565d4373e0e8ad4aa764cfd906708db Set 'distro' explicitly (Hennadii Stepanov)
Pull request description:
The [gitian-builder](https://github.com/devrandom/gitian-builder) implicitly uses `ubuntu` as a default distro.
[bin/gbuild#L237](81edd2fc8e/bin/gbuild (L237)):
```ruby
distro = build_desc["distro"] || "ubuntu"
```
This PR sets a gitian building distro explicitly in description files.
Tree-SHA512: d2a692047f3466a5c637433610854d1100fe68a78fb03b4a81c70911fa14b0228d9cf25bcf115930aab9cc8c5063dacaf79bcd151f1c3f6fee6419389aefcb8b
c966ff14c77870378847b9e6063b9671739ddc1f gitian: fixed SC2001 regex (willyk)
Pull request description:
Currently the gitian-win-signer.yml produces OUTFILE names without `-unsigned` stripped out
This is due to regex having an`%` in front of it
```
$ INFILE="bitcoin-0.19.0-win64-setup-unsigned.exe"
$ echo "${INFILE/%-unsigned}"
bitcoin-0.19.0-win64-setup-unsigned.exe
$ echo "${INFILE/-unsigned}"
bitcoin-0.19.0-win64-setup.exe
```
Fixes#17361
ACKs for top commit:
laanwj:
ACK c966ff14c77870378847b9e6063b9671739ddc1f
hebasto:
ACK c966ff14c77870378847b9e6063b9671739ddc1f
Tree-SHA512: 954547f9dfa4cab4def5f284d4837c21f0e6fed7454a04e83e6b1b7d3fd3f9661ea657047f0d8162f6591909d32ef2f72e801b2f3a44cbb1131ac344cb913a69
eafd2593674b6890418a50eadf69425071b824e4 build: Add NX workaround for RV64 (Wladimir J. van der Laan)
f6e42256fe8900ed90f9644906b77f97e1600952 build: Allow export of environ symbols (Wladimir J. van der Laan)
Pull request description:
This export was introduced in #17270 which added
```
//! Necessary on some platforms
extern char** environ;
```
This should (finally) make the gitian build pass again (fix issue #17525.).
Built on top of #17538 which should be merged first.
Top commit has no ACKs.
Tree-SHA512: 5c2054d52d0957aec3dc945b76d8e219187d22dc03889e7a88fb76049bf8e4a3e9f4da00dd1e9dd0351211f8e70d1a1b8ad7244f0348dab698e9d14b9d0c0bd4
fa193dc8e6f3b96fa2dba2f1c1668f7720fed320 doc: Remove win32 from the release process (MarcoFalke)
faf666f8148eeb305a9c4f78459aff2c7268016b Remove Windows 32 bit build (MarcoFalke)
Pull request description:
The Windows 32 bit build has been removed from https://bitcoincore.org/en/download/, so unless there are complaints, we don't need to build it even
ACKs for commit fa193d:
fanquake:
utACK fa193dc8e6
Tree-SHA512: d6f2976a2e0c407698f720b00ac23ec4056626de4eff8621f4c5581120af0460afd1bdef72329cc0e7d92afca48d94ae5fce6777cb36bfabb60b8034ff08fd88
530d02addbfea01ab24a2acd17af456a1e7b798a build: pass -fno-ident in Windows gitian descriptor (fanquake)
Pull request description:
`-fno-ident` prevents compilers from emitting compiler name and version number information that can needlessly bloat binaries.
For example, in the `v0.19.0.1` Windows release binaries, there are > 1000 GCC compiler version strings embedded:
```bash
# GCC: (GNU) 7.3-posix 20180312... & GCC: (GNU) 6.3.0 20170415.......
strings bitcoind.exe | rg GCC | wc -l
1021
```
They end up collected in the end of the`.rdata` section, and cannot be removed by `strip`. i.e:
```bash
objdump --section=.rdata --full-contents bitcoind.exe
...
cfcc00 00000000 00000000 00000000 00000000 ................
cfcc10 00000000 00000000 00000000 00000000 ................
cfcc20 4743433a 2028474e 55292036 2e332e30 GCC: (GNU) 6.3.0
cfcc30 20323031 37303431 35000000 00000000 20170415.......
cfcc40 4743433a 2028474e 55292037 2e332d70 GCC: (GNU) 7.3-p
cfcc50 6f736978 20323031 38303331 32000000 osix 20180312...
cfcc60 4743433a 2028474e 55292037 2e332d70 GCC: (GNU) 7.3-p
cfcc70 6f736978 20323031 38303331 32000000 osix 20180312...
```
The flag is available for [Clang](https://clang.llvm.org/docs/ClangCommandLineReference.html#cmdoption-clang-qn) and [GCC](https://gcc.gnu.org/onlinedocs/gcc/Code-Gen-Options.html#index-fno-ident).
Relevant code in [GCC](https://github.com/gcc-mirror/gcc/blob/master/gcc/toplev.c#L565-L578):
```c
/* Attach a special .ident directive to the end of the file to identify
the version of GCC which compiled this code. The format of the .ident
string is patterned after the ones produced by native SVR4 compilers. */
if (!flag_no_ident)
{
const char *pkg_version = "(GNU) ";
char *ident_str;
if (strcmp ("(GCC) ", pkgversion_string))
pkg_version = pkgversion_string;
ident_str = ACONCAT (("GCC: ", pkg_version, version_string, NULL));
targetm.asm_out.output_ident (ident_str);
}
```
ACKs for top commit:
practicalswift:
ACK 530d02addbfea01ab24a2acd17af456a1e7b798a
laanwj:
ACK 530d02addbfea01ab24a2acd17af456a1e7b798a
Tree-SHA512: b3b28f43ec483dee28d1df8548fe72425bf00e750701825c256395f6aa7b23256eb27609b51779b86aed108b6eaa3912181a9d8282e23eebf9cee7784f9fabe0
32da92bdf6bb55d6d312b0f85797d439cc942db5 gitian: Improve error handling (Wladimir J. van der Laan)
Pull request description:
Improve error handling in gitian builds:
- Set fail-on-error and pipefail flag, this causes a command to fail when either of the pipe stages fails, not only when the last of the stages fails, so this improves error detection.
- Also use `xargs` instead of `find -exec`, because `find` will not propagate errors in the executed command, but `xargs` will.
This will avoid some issues like #15541 where non-determinism is silently introduced due to errors caused by environment conditions (such as lack of disk space in that case).
Tree-SHA512: d5d3f22ce2d04a75e5c25e935744327c3adc704c2d303133f2918113573a564dff3d3243d5569a2b93ee7eb0e97f8e1b1ba81767e966af9015ea711a14091035
