Arch Linux User Repository

Binary version of this package is available on my repository on OBS and language packs are available as separete waterfox-classic-i18n-* packages.

Source files at: https://build.opensuse.org/package/show/home:hawkeye116477:waterfox/waterfox-classic-kpe.

[home_hawkeye116477_waterfox_Arch]
Server = https://download.opensuse.org/repositories/home:/hawkeye116477:/waterfox/Arch/$arch

If you got 404, then temporarily try to replace download.opensuse.org with downloadcontent.opensuse.org. That should enforce direct download without mirror.

key=$(curl -fsSL https://download.opensuse.org/repositories/home:hawkeye116477:waterfox/Arch/$(uname -m)/home_hawkeye116477_waterfox_Arch.key)
fingerprint=$(gpg --quiet --with-colons --import-options show-only --import --fingerprint <<< "${key}" | awk -F: '$1 == "fpr" { print $10 }')

pacman-key --init
pacman-key --add - <<< "${key}"
pacman-key --lsign-key "${fingerprint}"

well it goes through the [compile] portion fine, it gets to the [misc] portion, it buils some tests apparently, and gets to

waterfox-kde/src/Waterfox/build/unix/elfhack/test-ctors.c and fails:

9:40.49 Stack dump: 9:40.49 1. <eof> parser at end of file 9:40.49 2. Code generation 9:40.49 3. Running pass 'Function Pass Manager' on module 'waterfox-kde/src/Waterfox/build/unix/elfhack/test-ctors.c'. 9:40.49 4. Running pass 'X86 DAG->DAG Instruction Selection' on function '@test' 9:40.52 Stack dump: 9:40.52 1. <eof> parser at end of file 9:40.53 2. Code generation 9:40.53 3. Running pass 'Function Pass Manager' on module '/home/trumpop/waterfox-kde/src/Waterfox/build/unix/elfhack/test-array.c'. 9:40.53 4. Running pass 'X86 DAG->DAG Instruction Selection' on function '@test' 9:40.55 #0 0x00007f47d08b2b7b llvm::sys::PrintStackTrace(llvm::raw_ostream&) (/usr/bin/../lib/libLLVM-9.so+0xb08b7b) 9:40.55 #1 0x00007f47d08b0a44 llvm::sys::RunSignalHandlers() (/usr/bin/../lib/libLLVM-9.so+0xb06a44) 9:40.55 #2 0x00007f47d08b0bd6 (/usr/bin/../lib/libLLVM-9.so+0xb06bd6)

with a bunch of other messages followed by

9:40.55 #20 0x00007f47cf4d21d8 clang::EmitBackendOutput(clang::DiagnosticsEngine&, clang::HeaderSearchOptions const&, clang::CodeGenOptions const&, clang::TargetOptions const&, clang::LangOptions const&, llvm::DataLayout const&, llvm::Module*, clang::BackendAction, std::unique_ptr<llvm::raw_pwrite_stream, std::default_delete<llvm::raw_pwrite_stream> >) (/usr/bin/../lib/libclangCodeGen.so.9+0xc51d8) 9:40.55 clang-9: error: unable to execute command: Segmentation fault 9:40.55 clang-9: error: clang frontend command failed due to signal (use -v to see invocation) 9:40.55 clang version 9.0.0 (tags/RELEASE_900/final) 9:40.55 Target: x86_64-pc-linux-gnu 9:40.55 Thread model: posix 9:40.55 InstalledDir: /usr/bin 9:40.55 clang-9: note: diagnostic msg: PLEASE submit a bug report to and include the crash backtrace, preprocessed source, and associated run script.

9:40.58 clang-9: note: diagnostic msg: PLEASE submit a bug report to and include the crash backtrace, preprocessed source, and associated run script. 9:40.60 clang-9: note: diagnostic msg: 9:40.60 **** 9:40.60 9:40.60 PLEASE ATTACH THE FOLLOWING FILES TO THE BUG REPORT: 9:40.60 Preprocessed source(s) and associated run script(s) are located at: 9:40.60 clang-9: note: diagnostic msg: /tmp/test-ctors-c12089.c 9:40.60 clang-9: note: diagnostic msg: /tmp/test-ctors-c12089.sh 9:40.60 clang-9: note: diagnostic msg: 9:40.60 9:40.60 *** 9:40.60 make[5]: [waterfox-kde/src/Waterfox/config/rules.mk:791: test-ctors.o] Error 254 9:40.60 make[5]: Waiting for unfinished jobs.... 9:40.63 clang-9: note: diagnostic msg: 9:40.63 ***** 9:40.63 9:40.63 PLEASE ATTACH THE FOLLOWING FILES TO THE BUG REPORT: 9:40.63 Preprocessed source(s) and associated run script(s) are located at: 9:40.63 clang-9: note: diagnostic msg: /tmp/test-array-bfb319.c 9:40.63 clang-9: note: diagnostic msg: /tmp/test-array-bfb319.sh 9:40.63 clang-9: note: diagnostic msg:

that's all. I wondered if some march or mtune settings are causing this? I have a ryzen 3700x (zen2) so maybe the cpu is conflicting there? i believe clang/llvm 9 have native support for ryzen (march=zenver2) so i switched those, and it fails at the exact same point. both test-ctors.c and the following file waterfox-kde/src/Waterfox/build/unix/elfhack/test-array.c fail with the same errors, both in the same folder waterfox-kde/src/Waterfox/obj-x86_64-pc-linux-gnu/build/unix/elfhack

and like i said, downgrading clang/llvm to previous versions, it builds just fine

@toxygen Hmm, on OBS was fine. What error message displayed?

man, i dont know what's up with either my system or with the new rust/llvm/clang, i couldnt build the new version with the latest versions of those, had to downgrade to clang/llvm 8.0.1 and the previous version of rust, and it built successfully

with the latest versions, the build itself went fine, but the "misc" portion failed for me with some llvm error followed by a compiler error.

so far working ok with using the previous versions to build it though

@toxygen Cssparser needs updating, but seems that's not that easy, cuz it's required by servo, which also needs some fixes to use new version of cssparser. That's rather not worth to make that fixes, cuz Servo/Stylo can't work good for now, so you can just disable stylo and that will fix that error also :-)

cant build because of changes in rust 1.38 and a cssparser error: (file: src/Waterfox/third_party/rust/cssparser/src/parser.rs) 3:20.30 error[E0506]: cannot assign to self.input.cached_token because it is borrowed 3:20.43 error: Could not compile cssparser.

per https://bugzilla.mozilla.org/show_bug.cgi?id=1585099 "Rust was updated to 1.38 in OpenBSD, and as a side effect thunderbird 68 & firefox-esr 68 now fail to build, choking on cssparser - see https://marc.info/?l=openbsd-ports&m=156984549605237&w=2 for the original report/bugfix."

this seems to affect arch/artix waterfox-kde as well, at least the error and location are the same i attempted some patching but got nowhere (as per comment "backporting https://github.com/servo/rust-cssparser/pull/245/commits/3c98d22c5de3b696bf1fde2b6c90069812312aa6 on top of third-party/rust/cssparser/src/parser.rs was enough to fix the build with rust 1.38" on that bugzilla) but i got nowhere

also looks like downgrading rust to 1.37-2 allows build to complete, with a warning instead of a build error at the same file/location (src/Waterfox/third_party/rust/cssparser/src/parser.rs)

Binary version of this package is available on my repository on OBS and language packs are available as separete waterfox-classic-i18n-* packages.

Source files at: https://build.opensuse.org/package/show/home:hawkeye116477:waterfox/waterfox-classic-kpe.

[home_hawkeye116477_waterfox_Arch]
Server = https://download.opensuse.org/repositories/home:/hawkeye116477:/waterfox/Arch/$arch

If you got 404, then temporarily try to replace download.opensuse.org with downloadcontent.opensuse.org. That should enforce direct download without mirror.

key=$(curl -fsSL https://download.opensuse.org/repositories/home:hawkeye116477:waterfox/Arch/$(uname -m)/home_hawkeye116477_waterfox_Arch.key)
fingerprint=$(gpg --quiet --with-colons --import-options show-only --import --fingerprint <<< "${key}" | awk -F: '$1 == "fpr" { print $10 }')

pacman-key --init
pacman-key --add - <<< "${key}"
pacman-key --lsign-key "${fingerprint}"

Rust-simd should work again now, thanks to my patch based on <https://bugzilla.mozilla.org/show_bug.cgi?id=1372230> and <https://bugzilla.mozilla.org/show_bug.cgi?id=1521249> :-)

you need to add "git" to makedepends, also 'libevent'

It worked on my tablet, but my desktop PC is failing to verify waterfox.1.

By chance did you change it recently, or is it a screwup on my end somehow?

EDIT: Nvm, fixed it by skipping waterfox.1's SHA256 verification in the pkgbuild manually using Pacaur's --edit flag. I mean, it's a manual, in plaintext, I can't imagine that ever being a risk.