too long lines were shrinked: x86_64-pc-linux-gnu-g++ -c -pipe -march=native -fno-diagnostics-color -O2 -Wno-deprecated -Wno-deprecated-declarations -Wno-cpp -Wno-inconsistent-missing-override -O0 -g -fno-omit-frame-pointer -D_FILE_OFFSET_BITS=64 -D_REENTRANT -Wall -Wextra -fPIC -DV07_NON_BACKWARD_COMPATIBLE_CHANGE_001 -DV07_NON x86_64-pc-linux-gnu-g++ -c -pipe -march=native -fno-diagnostics-color -O2 -Wno-deprecated -Wno-deprecated-declarations -Wno-cpp -Wno-inconsistent-missing-override -O0 -g -fno-omit-frame-pointer -D_FILE_OFFSET_BITS=64 -D_REENTRANT -Wall -Wextra -fPIC -DV07_NON_BACKWARD_COMPATIBLE_CHANGE_001 -DV07_NON x86_64-pc-linux-gnu-g++ -c -pipe -march=native -fno-diagnostics-color -O2 -Wno-deprecated -Wno-deprecated-declarations -Wno-cpp -Wno-inconsistent-missing-override -O0 -g -fno-omit-frame-pointer -D_FILE_OFFSET_BITS=64 -D_REENTRANT -Wall -Wextra -fPIC -DV07_NON_BACKWARD_COMPATIBLE_CHANGE_001 -DV07_NON x86_64-pc-linux-gnu-g++ -c -pipe -march=native -fno-diagnostics-color -O2 -Wno-deprecated -Wno-deprecated-declarations -Wno-cpp -Wno-inconsistent-missing-override -O0 -g -fno-omit-frame-pointer -D_FILE_OFFSET_BITS=64 -D_REENTRANT -Wall -Wextra -fPIC -DV07_NON_BACKWARD_COMPATIBLE_CHANGE_001 -DV07_NON x86_64-pc-linux-gnu-g++ -c -pipe -march=native -fno-diagnostics-color -O2 -Wno-deprecated -Wno-deprecated-declarations -Wno-cpp -Wno-inconsistent-missing-override -O0 -g -fno-omit-frame-pointer -D_FILE_OFFSET_BITS=64 -D_REENTRANT -Wall -Wextra -fPIC -DV07_NON_BACKWARD_COMPATIBLE_CHANGE_001 -DV07_NON In file included from gui/statistics/Histogram.cpp:23: gui/statistics/Histogram.h:35:27: error: uint32_t was not declared in this scope 35 | const std::vector<uint32_t>& entries() const { return mBins; } | ^~~~~~~~ ------------------------------------------------------------------- This is an unstable amd64 chroot image at a tinderbox (==build bot) name: 17.1_desktop-j4-20230218-040024 ------------------------------------------------------------------- GNUMAKEFLAGS="$GNUMAKEFLAGS --jobserver-style=pipe" gcc-config -l: [1] x86_64-pc-linux-gnu-10 [2] x86_64-pc-linux-gnu-13 * clang/llvm (if any): clang version 15.0.7 Target: x86_64-pc-linux-gnu Thread model: posix InstalledDir: /usr/lib/llvm/15/bin Configuration file: /etc/clang/clang.cfg /usr/lib/llvm/15 15.0.7 Python 3.10.10 Available Ruby profiles: [1] ruby27 (with Rubygems) [2] ruby30 (with Rubygems) * Available Rust versions: [1] rust-bin-1.67.1 * The following VMs are available for generation-2: 1) IcedTea JDK 3.21.0 [icedtea-8] 2) Eclipse Temurin JDK 11.0.18_p10 [openjdk-bin-11] *) Eclipse Temurin JDK 17.0.6_p10 [openjdk-bin-17] 4) Eclipse Temurin JDK 8.362_p09 [openjdk-bin-8] 5) Eclipse Temurin JRE 17.0.6_p10 [openjdk-jre-bin-17] Available Java Virtual Machines: [1] icedtea-8 [2] openjdk-bin-8 [3] openjdk-bin-11 [4] openjdk-bin-17 system-vm [5] openjdk-jre-bin-17 The Glorious Glasgow Haskell Compilation System, version 9.0.2 php cli (if any): HEAD of ::gentoo commit ef501cacacb71a632dd521f79d73d7ee8ece1367 Author: Repository mirror & CI <repomirrorci@gentoo.org> Date: Tue Feb 21 02:16:50 2023 +0000 2023-02-21 02:16:50 UTC emerge -qpvO net-p2p/retroshare [ebuild N ] net-p2p/retroshare-0.6.6-r1 USE="cli gui jsonapi miniupnp service sqlcipher -autologin -libupnp"
Created attachment 853570 [details] emerge-info.txt
Created attachment 853572 [details] emerge-history.txt.bz2
Created attachment 853574 [details] environment
Created attachment 853576 [details] etc.clang.tar.bz2
Created attachment 853578 [details] etc.portage.tar.bz2
Created attachment 853580 [details] logs.tar.bz2
Created attachment 853582 [details] net-p2p:retroshare-0.6.6-r1:20230221-052419.log.bz2
Created attachment 853584 [details] temp.tar.bz2
ping - this should be a trivial fix, just a missing include
also fails to build with gcc 12
Created attachment 874822 [details, diff] gcc-13-build_fix.patch Seems it works. Builds, at least
(In reply to Alex D-Bug from comment #11) > Created attachment 874822 [details, diff] [details, diff] > gcc-13-build_fix.patch > > Seems it works. Builds, at least can confirm the same, it builds with GCC 13
Fixed upstream. Retroshare should be bumped to 0.6.7 for bugfix. Also, cxx-17 patch is also not need with 0.6.7.
Created attachment 888476 [details] ebuild, eapi 7
(In reply to jy6x2b32pie9 from comment #14) > Created attachment 888476 [details] > ebuild, eapi 7 Hello, bug#927207 is about net-p2p/retroshare version bump, there's also a new 'plugins' use flag: https://bugs.gentoo.org/attachment.cgi?id=887848
Created attachment 888770 [details, diff] net-p2p/retroshare/files: retroshare-0.6.6-fix-uint32_t-compilation.patch (In reply to Alex D-Bug from comment #11) > Created attachment 874822 [details, diff] [details, diff] > gcc-13-build_fix.patch > > Seems it works. Builds, at least Hello Alex D-Bug, Since you provided a solution, could you please open a https://github.com/gentoo/gentoo PR? Thanks ;) I'm including an attachment that adopts your solution combining some upstream patches: - https://github.com/RetroShare/RetroShare/commit/e1934fd9b03cd52c556eb06d94fb5d68b649592e - https://github.com/RetroShare/RetroShare/commit/15128192003722a3686c1b6f27132ba064cd1705 $ diff retroshare-0.6.6-r1.ebuild retroshare-0.6.6-r2.ebuild 15c15 < KEYWORDS="amd64 x86" --- > KEYWORDS="~amd64 ~x86" 56c56,59 < PATCHES=( "${FILESDIR}/${P}-fix-cxx17-compilation.patch" ) --- > PATCHES=( > "${FILESDIR}/${P}-fix-cxx17-compilation.patch" > "${FILESDIR}/${P}-fix-uint32_t-compilation.patch" > ) Also, the compiler suggests a solution to the uint32_t compilation error too: In file included from gui/statistics/Histogram.cpp:23: gui/statistics/Histogram.h:35:27: error: ‘uint32_t’ was not declared in this scope 35 | const std::vector<uint32_t>& entries() const { return mBins; } | ^~~~~~~~ gui/statistics/Histogram.h:23:1: note: ‘uint32_t’ is defined in header ‘<cstdint>’; did you forget to ‘#include <cstdint>’? 22 | #include <iostream> +++ |+#include <cstdint>
Had a look at he gcc-13-build_fix.patch and the net-p2p/retroshare/files: retroshare-0.6.6-fix-uint32_t-compilation.patch. Also tested both and can confirm both fix the issue. When comparing them I first thought the compiler as well as the retroshare-0.6.6-fix-uint32_t-compilation.patch mustg be right. When looking at the other patch through I noted that the retroshare-0.6.6-fix-uint32_t-compilation.patch casts the value to uint32_t just to store it as long int (causing another cast) while the gcc-13-build_fix.patch casts directly to long int. My guess now is that this is a upstream issue: Either they want to cast to a architecture dependend fixed word size (uint32_t is guaranteed to have a wordsize of 32 bit) or they want a wordsize depending on the target architecture (long int). Maybe they ported away from fixed word size and this cast to uint32_t was just an oversight or maybe they really need a fixed word size here no matter what architecture and the second cast to long int is a mistake (the LoC shold look like this: uint32_t bin = (uint32_t)floor((val - mStart)/(mEnd - mStart) * mBins.size()); ) but nobody noticed as they didn't compile for any architectures with a wordsize of long int smaller than 32 bit (are there even any such architectures?). rl;dr: Just my two cents but while both patches work on at least amd64 and probably a lot of other archs, too, I'm unsure which of the patches does the right thing in the eyes of the retroshare devs, hence I think it would be a good idea if one of the patch authors could take the patches upstream and simply ask them.
