Arch Linux User Repository

Description of build options and defaults. See PKGBUILD for current list.

• _build_save_source (true) – save tarball of patched sources
• _build_repatch (false) – discard previously saved tarball
• _build_pgo (true) – enable profile guided optimization; ~20% better benchmarks, 3× build time
• _build_pgo_reuse (try) – reuse previously generated profile
• _build_pgo_xvfb (true) – use Xvfb for profiling, otherwise, use xwayland-run
• _build_lto (false) – use link-time optimization (LTO); disabling may prevent spurious crashes
• _build_system_libs (true) – use system libraries
• _build_limit_cores (false) – limit parallelization based on memory and core availability

Examples of use:

• _build_pgo=false makepkg -Csr
• extra-x86_64-build -- -- _build_limit_cores=true
• _build_pgo_xvfb=false yay icecat # usage with AUR helpers may vary

Hi,

Build fails since last commit,

Running modification-aware sed: sed s/777/755/; -i toolkit/crashreporter/google-breakpad/Makefile.in
+ /bin/sed 's/777/755/;' -i toolkit/crashreporter/google-breakpad/Makefile.in
++ md5sum toolkit/crashreporter/google-breakpad/Makefile.in
+ '[' 'ee125f2b99c1b0fd66ee6cf3e8295ef7  toolkit/crashreporter/google-breakpad/Makefile.in' = 'efadb4b06811823b5f7ed244ea5a9da3  toolkit/crashreporter/google-breakpad/Makefile.in' ']'
+ /bin/sed 's/chmod a+w/chmod u+w/' -i ./js/src/ctypes/libffi/Makefile.in ./toolkit/crashreporter/google-breakpad/Makefile.in ./toolkit/crashreporter/google-breakpad/src/third_party/glog/Makefile.in
patching file ipc/chromium/src/third_party/libevent/README.mozilla
Reversed (or previously applied) patch detected!  Skipping patch.
1 out of 1 hunk ignored -- saving rejects to file ipc/chromium/src/third_party/libevent/README.mozilla.rej
patching file ipc/chromium/src/third_party/libevent/linux/event2/event-config.h
Reversed (or previously applied) patch detected!  Skipping patch.
1 out of 1 hunk ignored -- saving rejects to file ipc/chromium/src/third_party/libevent/linux/event2/event-config.h.rej
==> ERROR: A failure occurred in prepare().
    Aborting...

@lino_sparks IceCat sources are pulled by makeicecat.sh script, we can patch it, but probably the problem is in the side of yay. With makepkg it works just fine,

This problem doesn't exist. Something is wrong with your setup.

When using an aurhelper(yay) to build, It has problems downloading the firefox tarball, it trys to download from https://ftp.mozilla.org/pub/mozilla.org/firefox/releases/102.4.0esr/source/firefox-102.4.0esr.source.tar.xz

when it should be https://ftp.mozilla.org/pub/firefox/releases/102.4.0esr/source/firefox-102.4.0esr.source.tar.xz The problem persists when building with makepkg

Although downloading the tarball and asc straight from the ftp site, and build with makepkg -si does bypass these problems.

Yeah, built fine w/o any hassle, thanks for cherry pick.

New patched added. It should build now.

Faced error during compile - I checked if this came from original (mozilla esr source) and this isn't made by any gnuzilla patch:

error: expected one of `!` or `::`, found keyword `mod`
6:03.62    --> xxx/icecat/src/gnuzilla-e77a1fa25ac66250e88d5d9901a2ba670edb94cc/output/icecat-102.2.0/third_party/rust/packed_simd_2/src/lib.rs:347:7
6:03.62     |
6:03.62 347 | crate mod llvm {
6:03.62     |       ^^^ expected one of `!` or `::`
6:03.62 error: could not compile `packed_simd_2` due to previous error

Couple of steps led to upstream bug: https://bugzilla.mozilla.org/show_bug.cgi?id=1783784

"Approved for 102.3esr. Thanks for the fix-up."

TL;DR

https://firefox-source-docs.mozilla.org/writing-rust-code/update-policy.html

Firefox 102     Rust 1.60.0     1.59.0  2022 April 7    2022 May 26     2022 June 28

third_party/rust/packed_simd_2/README.MD has some info:

# The Crates.io Version Can No Longer Be Updated!

The original maintainer is out of contact, and the new maintainers (the Portable SIMD Project Group) do not have the appropriate cra
tes.io permissions to issue updates.

We are aware that the version available on crates.io is currently broken, and will not build.

If you need to continue to use the crate, we have published a "next version" under an alternative name.

Adjust your `[dependencies]` section of `Cargo.toml` to be the following:
```toml
packed_simd = { version = "0.3.6", package = "packed_simd_2" }

I also noticed that ever since glibc was upgraded on Aug 19, 2022 there were problems with some apps compiling or running. An example is MakeMKV and Ungoogled Chromium.

Here are lines 39 - 63 of
gnuzilla-e77a1fa25ac66250e88d5d9901a2ba670edb94cc/output/icecat-102.2.0

# In automation, manage an sccache daemon. The starting of the server
# needs to be in a make file so sccache inherits the jobserver.
ifdef MOZBUILD_MANAGE_SCCACHE_DAEMON
SCCACHE_STOP = $(MOZBUILD_MANAGE_SCCACHE_DAEMON) --stop-server

# When a command fails, make is going to abort, but we need to terminate the
# sccache server, otherwise it will prevent make itself from terminating
# because it would still be running and holding a jobserver token.
# However, we also need to preserve the command's exit code, thus the
# gymnastics.
SCCACHE_STOP_ON_FAILURE = || (x=$$?; $(SCCACHE_STOP) || true; exit $$x)

build::
    # Terminate any sccache server that might still be around.
    -$(SCCACHE_STOP) > /dev/null 2>&1
    # Start a new server, ensuring it gets the jobserver file descriptors
    # from make (but don't use the + prefix when make -n is used, so that
    # the command doesn't run in that case)
    mkdir -p $(UPLOAD_PATH)
    $(if $(findstring n,$(filter-out --%, $(MAKEFLAGS))),,+)env SCCACHE_LOG=sccache=debug SCCACHE_ERROR_LOG=$(UPLOAD_PATH)/sccache.log $(MOZBUILD_MANAGE_SCCACHE_DAEMON) --start-server
endif

### Build it
build::
    +$(MOZ_MAKE) $(SCCACHE_STOP_ON_FAILURE)

Here are the lines 355 -361 of
gnuzilla-e77a1fa25ac66250e88d5d9901a2ba670edb94cc/output/icecat-102.2.0/config/rules.mk

# The root makefile doesn't want to do a plain export/libs, because
# of the tiers and because of libxul. Suppress the default rules in favor
# of something else. Makefiles which use this var *must* provide a sensible
# default rule before including rules.mk
default all::
    $(foreach tier,$(TIERS),$(call SUBMAKE,$(tier)))

I hope that this helps.

@rev.cressy I noticed I had rust 1.62.0... I'll try with 1.63, perhaps it need some patch.

I did a full system upgrade and still icecat build breaks at the same point. I usually do a full system upgrade every week.