max/nix-super
1
0
Fork 0
mirror of https://github.com/privatevoid-net/nix-super.git synced 2024-11-23 14:36:16 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions
nix-super/Makefile

122 lines
3.3 KiB
Makefile
Raw Normal View History

Build Windows DLLs with `-Wl,--export-all-symbols` This is not the most elegant, but will match the SOs in exporting everything for now. Later we can refine what is public/private to clean up the interface.
2024-01-09 19:41:42 +02:00
# External build directory support
Use `buildprefix` in a few more places `installcheck` doesn't yet work, but the rest of the build can now happen mostly inside a separate build directory. Progress on #9342 Co-authored-by: Valentin Gagarin <valentin.gagarin@tweag.io>
2023-11-25 07:33:21 +02:00
include mk/build-dir.mk
-include $(buildprefix)Makefile.config
clean-files += $(buildprefix)Makefile.config
Use positive source filtering for the standalone functional tests job Additionally this skipping of the building is reimplemented to be a bit more robust and use the same idioms as the functionality for skipping the tests. In particular, it will now work even if the source files exist, so we can do this during development too.
2023-10-06 18:57:31 +03:00
Build Windows DLLs with `-Wl,--export-all-symbols` This is not the most elegant, but will match the SOs in exporting everything for now. Later we can refine what is public/private to clean up the interface.
2024-01-09 19:41:42 +02:00
# List makefiles
Use positive source filtering for the standalone functional tests job Additionally this skipping of the building is reimplemented to be a bit more robust and use the same idioms as the functionality for skipping the tests. In particular, it will now work even if the source files exist, so we can do this during development too.
2023-10-06 18:57:31 +03:00
ifeq ($(ENABLE_BUILD), yes)
Update Makefile variable names
2014-02-01 13:23:14 +02:00
makefiles = \
Fix Perl bindings
2019-11-07 13:18:37 +02:00
mk/precompiled-headers.mk \
Fix "make dist"
2014-02-01 15:28:31 +02:00
local.mk \
Rename Makefile -> local.mk
2014-01-30 13:11:06 +02:00
src/libutil/local.mk \
src/libstore/local.mk \
Move fetchers from libstore to libfetchers
2020-03-30 15:03:28 +03:00
src/libfetchers/local.mk \
Rename Makefile -> local.mk
2014-01-30 13:11:06 +02:00
src/libmain/local.mk \
src/libexpr/local.mk \
Move command plugin interface to libnixcmd
2021-01-26 13:22:24 +02:00
src/libcmd/local.mk \
Start of new Nix command-line interface
2016-02-09 22:28:29 +02:00
src/nix/local.mk \
resolve-system-dependencies: implement in C++
2016-08-13 21:36:22 +03:00
src/resolve-system-dependencies/local.mk \
Rename Makefile -> local.mk
2014-01-30 13:11:06 +02:00
scripts/local.mk \
nix: Implement basic bash completion
2020-05-10 21:32:21 +03:00
misc/bash/local.mk \
Add a fish completion script This is only rudimentary support as allowed by `NIX_GET_COMPLETIONS`. In the future, we could use complete’s `--wraps` argument to autocomplete arguments for programs after `nix shell -c`.
2021-06-02 01:45:03 +03:00
misc/fish/local.mk \
Install zsh completion script
2021-06-01 19:44:03 +03:00
misc/zsh/local.mk \
Install systemd units
2014-04-07 12:18:54 +03:00
misc/systemd/local.mk \
Add a launchd configuration file to run nix-daemon
2014-11-04 11:14:18 +02:00
misc/launchd/local.mk \
Docs build: depend on locally built nix executable and not installed one Previously many of the documentation targets were depending on `$(bindir)/nix` which is the installed version. This meant that its install rules would be triggered (which in chain would also trigger the install of libraries, as reported in #5140). Therefore a build of the documentation without an installation would not be possible (which apart from doing unwanted operations it may also generate permission problems for example). The fix makes the rules depend on `$(nix_PATH)` instead, which is the executable in the build tree.
2021-08-17 05:26:41 +03:00
misc/upstart/local.mk
Use positive source filtering for the standalone functional tests job Additionally this skipping of the building is reimplemented to be a bit more robust and use the same idioms as the functionality for skipping the tests. In particular, it will now work even if the source files exist, so we can do this during development too.
2023-10-06 18:57:31 +03:00
endif
Rename Makefile.new -> Makefile
2013-11-25 17:25:13 +02:00
Split `--disable-tests`, fix cross builds It might seem obnoxious to have yet more configure flags, but I found controlling both the unit and functional tests with one flag was quite confusing because they are so different: - unit tests depending on building, functional tests don't (e.g. when we test already-built Nix) - unit tests can be installed, functional tests cannot - unit tests neeed extra libraries (GTest, RapidCheck), functional tests need extra executables (jq). - unit tests are run by `make check`, functional tests are run by `make installcheck` Really on a technical level, they seem wholly independent. Only on a human level ("they are both are tests") do they have anything in common. I had messed up the logic in cross builds because of this. Now I split the flag in two (and cleaned up a few other inconsistencies), and the logic fixed itself. Co-Authored-By: Robert Hensing <roberth@users.noreply.github.com>
2023-12-14 07:05:03 +02:00
ifeq ($(ENABLE_UNIT_TESTS), yes)
Add ./configure --disable-tests option Building without tests is useful for bootstrapping with a smaller footprint or running the tests in a separate derivation. Otherwise, we do compile and run them. This isn't fine grained as to allow picking `check` but not `installcheck` or vice versa, but it's good enough for now. I've tried to use Nixpkgs' `checkInputs`, but those inputs weren't discovered properly by the configure script. We can emulate its behavior very well though.
2023-02-20 21:27:50 +02:00
makefiles += \
Move tests to separate directories, and document Today, with the tests inside a `tests` intermingled with the corresponding library's source code, we have a few problems: - We have to be careful that wildcards don't end up with tests being built as part of Nix proper, or test headers being installed as part of Nix proper. - Tests in libraries but not executables is not right: - It means each executable runs the previous unit tests again, because it needs the libraries. - It doesn't work right on Windows, which doesn't want you to load a DLL just for the side global variable . It could be made to work with the dlopen equivalent, but that's gross! This reorg solves these problems. There is a remaining problem which is that sibbling headers (like `hash.hh` the test header vs `hash.hh` the main `libnixutil` header) end up shadowing each other. This PR doesn't solve that. That is left as future work for a future PR. Co-authored-by: Valentin Gagarin <valentin.gagarin@tweag.io>
2023-08-25 17:20:28 +03:00
tests/unit/libutil/local.mk \
tests/unit/libutil-support/local.mk \
tests/unit/libstore/local.mk \
tests/unit/libstore-support/local.mk \
tests/unit/libexpr/local.mk \
tests/unit/libexpr-support/local.mk
Use positive source filtering for the standalone functional tests job Additionally this skipping of the building is reimplemented to be a bit more robust and use the same idioms as the functionality for skipping the tests. In particular, it will now work even if the source files exist, so we can do this during development too.
2023-10-06 18:57:31 +03:00
endif
Split `--disable-tests`, fix cross builds It might seem obnoxious to have yet more configure flags, but I found controlling both the unit and functional tests with one flag was quite confusing because they are so different: - unit tests depending on building, functional tests don't (e.g. when we test already-built Nix) - unit tests can be installed, functional tests cannot - unit tests neeed extra libraries (GTest, RapidCheck), functional tests need extra executables (jq). - unit tests are run by `make check`, functional tests are run by `make installcheck` Really on a technical level, they seem wholly independent. Only on a human level ("they are both are tests") do they have anything in common. I had messed up the logic in cross builds because of this. Now I split the flag in two (and cleaned up a few other inconsistencies), and the logic fixed itself. Co-Authored-By: Robert Hensing <roberth@users.noreply.github.com>
2023-12-14 07:05:03 +02:00
ifeq ($(ENABLE_FUNCTIONAL_TESTS), yes)
Use positive source filtering for the standalone functional tests job Additionally this skipping of the building is reimplemented to be a bit more robust and use the same idioms as the functionality for skipping the tests. In particular, it will now work even if the source files exist, so we can do this during development too.
2023-10-06 18:57:31 +03:00
makefiles += \
Put functional tests in `tests/functional` I think it is bad for these reasons when `tests/` contains a mix of functional and integration tests - Concepts is harder to understand, the documentation makes a good unit vs functional vs integration distinction, but when the integration tests are just two subdirs within `tests/` this is not clear. - Source filtering in the `flake.nix` is more complex. We need to filter out some of the dirs from `tests/`, rather than simply pick the dirs we want and take all of them. This is a good sign the structure of what we are trying to do is not matching the structure of the files. With this change we have a clean: ```shell-session $ git show 'HEAD:tests' tree HEAD:tests functional/ installer/ nixos/ ```
2023-10-05 19:12:18 +03:00
tests/functional/local.mk \
tests/functional/ca/local.mk \
tests/functional/dyn-drv/local.mk \
Merge remote-tracking branch 'upstream/master' into overlayfs-store
2023-10-23 20:13:37 +03:00
tests/functional/local-overlay-store/local.mk \
Put functional tests in `tests/functional` I think it is bad for these reasons when `tests/` contains a mix of functional and integration tests - Concepts is harder to understand, the documentation makes a good unit vs functional vs integration distinction, but when the integration tests are just two subdirs within `tests/` this is not clear. - Source filtering in the `flake.nix` is more complex. We need to filter out some of the dirs from `tests/`, rather than simply pick the dirs we want and take all of them. This is a good sign the structure of what we are trying to do is not matching the structure of the files. With this change we have a clean: ```shell-session $ git show 'HEAD:tests' tree HEAD:tests functional/ installer/ nixos/ ```
2023-10-05 19:12:18 +03:00
tests/functional/test-libstoreconsumer/local.mk \
tests/functional/plugins/local.mk
Add ./configure --disable-tests option Building without tests is useful for bootstrapping with a smaller footprint or running the tests in a separate derivation. Otherwise, we do compile and run them. This isn't fine grained as to allow picking `check` but not `installcheck` or vice versa, but it's good enough for now. I've tried to use Nixpkgs' `checkInputs`, but those inputs weren't discovered properly by the configure script. We can emulate its behavior very well though.
2023-02-20 21:27:50 +02:00
endif
Build Windows DLLs with `-Wl,--export-all-symbols` This is not the most elegant, but will match the SOs in exporting everything for now. Later we can refine what is public/private to clean up the interface.
2024-01-09 19:41:42 +02:00
# Miscellaneous global Flags
Revert "autoconf: Allow overriding CFLAGS/CXXFLAGS from outside." This reverts commit 717e821b99797845e1bef47d862f8cb0fb69cfc9. It's much more convenient to do 'make OPTIMIZE=0'.
2019-11-06 17:55:37 +02:00
OPTIMIZE = 1
ifeq ($(OPTIMIZE), 1)
enable LTO in optimized builds gives 2-5% performance improvement across a board of tests. LTO is broken when using clang; some libs link fine while others crash the linker with a segfault in the llvm linker plugin. 🙁
2021-12-21 20:34:40 +02:00
GLOBAL_CXXFLAGS += -O3 $(CXXLTO)
GLOBAL_LDFLAGS += $(CXXLTO)
Revert "autoconf: Allow overriding CFLAGS/CXXFLAGS from outside." This reverts commit 717e821b99797845e1bef47d862f8cb0fb69cfc9. It's much more convenient to do 'make OPTIMIZE=0'.
2019-11-06 17:55:37 +02:00
else
Disable `FORTIFY_SOURCE` when compiling without optims Otherwise the build is cluttered with ``` /nix/store/fwpn2f7a4iqszyydw7ag61zlnp6xk5d3-glibc-2.30-dev/include/features.h:382:4: warning: #warning _FORTIFY_SOURCE requires compiling with optimization (-O) [-Wcpp] 382 | # warning _FORTIFY_SOURCE requires compiling with optimization (-O) | ^~~~~~~ ``` when building with `OPTIMIZE=0`
2020-09-22 11:04:25 +03:00
GLOBAL_CXXFLAGS += -O0 -U_FORTIFY_SOURCE
Revert "autoconf: Allow overriding CFLAGS/CXXFLAGS from outside." This reverts commit 717e821b99797845e1bef47d862f8cb0fb69cfc9. It's much more convenient to do 'make OPTIMIZE=0'.
2019-11-06 17:55:37 +02:00
endif
Build Windows DLLs with `-Wl,--export-all-symbols` This is not the most elegant, but will match the SOs in exporting everything for now. Later we can refine what is public/private to clean up the interface.
2024-01-09 19:41:42 +02:00
include mk/platform.mk
ifdef HOST_WINDOWS
# Windows DLLs are stricter about symbol visibility than Unix shared
# objects --- see https://gcc.gnu.org/wiki/Visibility for details.
# This is a temporary sledgehammer to export everything like on Unix,
# and not detail with this yet.
#
# TODO do not do this, and instead do fine-grained export annotations.
GLOBAL_LDFLAGS += -Wl,--export-all-symbols
endif
GLOBAL_CXXFLAGS += -g -Wall -include $(buildprefix)config.h -std=c++2a -I src
# Include the main lib, causing rules to be defined
Rename Makefile.new -> Makefile
2013-11-25 17:25:13 +02:00
include mk/lib.mk
Precompile headers This cuts 'make install -j6' on my laptop from 170s to 134s.
2019-11-06 17:54:18 +02:00
Build Windows DLLs with `-Wl,--export-all-symbols` This is not the most elegant, but will match the SOs in exporting everything for now. Later we can refine what is public/private to clean up the interface.
2024-01-09 19:41:42 +02:00
# Fallback stub rules for better UX when things are disabled
#
# These must be defined after `mk/lib.mk`. Otherwise the first rule
# incorrectly becomes the default target.
Move down fallback targets in `Makefile` This ensures `lib.mk` still defines `default` as the first target. This fixes some builds.
2023-12-20 10:23:49 +02:00
ifneq ($(ENABLE_UNIT_TESTS), yes)
.PHONY: check
check:
@echo "Unit tests are disabled. Configure without '--disable-unit-tests', or avoid calling 'make check'."
@exit 1
endif
ifneq ($(ENABLE_FUNCTIONAL_TESTS), yes)
.PHONY: installcheck
installcheck:
@echo "Functional tests are disabled. Configure without '--disable-functional-tests', or avoid calling 'make installcheck'."
@exit 1
endif
Build Windows DLLs with `-Wl,--export-all-symbols` This is not the most elegant, but will match the SOs in exporting everything for now. Later we can refine what is public/private to clean up the interface.
2024-01-09 19:41:42 +02:00
# Documentation or else fallback stub rules.
#
# The documentation makefiles be included after `mk/lib.mk` so rules
# refer to variables defined by `mk/lib.mk`. Rules are not "lazy" like
# variables, unfortunately.
Move down fallback targets in `Makefile` This ensures `lib.mk` still defines `default` as the first target. This fixes some builds.
2023-12-20 10:23:49 +02:00
ifeq ($(ENABLE_DOC_GEN), yes)
Docs build: depend on locally built nix executable and not installed one Previously many of the documentation targets were depending on `$(bindir)/nix` which is the installed version. This meant that its install rules would be triggered (which in chain would also trigger the install of libraries, as reported in #5140). Therefore a build of the documentation without an installation would not be possible (which apart from doing unwanted operations it may also generate permission problems for example). The fix makes the rules depend on `$(nix_PATH)` instead, which is the executable in the build tree.
2021-08-17 05:26:41 +03:00
$(eval $(call include-sub-makefile, doc/manual/local.mk))
Split `--disable-tests`, fix cross builds It might seem obnoxious to have yet more configure flags, but I found controlling both the unit and functional tests with one flag was quite confusing because they are so different: - unit tests depending on building, functional tests don't (e.g. when we test already-built Nix) - unit tests can be installed, functional tests cannot - unit tests neeed extra libraries (GTest, RapidCheck), functional tests need extra executables (jq). - unit tests are run by `make check`, functional tests are run by `make installcheck` Really on a technical level, they seem wholly independent. Only on a human level ("they are both are tests") do they have anything in common. I had messed up the logic in cross builds because of this. Now I split the flag in two (and cleaned up a few other inconsistencies), and the logic fixed itself. Co-Authored-By: Robert Hensing <roberth@users.noreply.github.com>
2023-12-14 07:05:03 +02:00
else
.PHONY: manual-html manpages
manual-html manpages:
@echo "Generated docs are disabled. Configure without '--disable-doc-gen', or avoid calling 'make manpages' and 'make manual-html'."
@exit 1
Docs build: depend on locally built nix executable and not installed one Previously many of the documentation targets were depending on `$(bindir)/nix` which is the installed version. This meant that its install rules would be triggered (which in chain would also trigger the install of libraries, as reported in #5140). Therefore a build of the documentation without an installation would not be possible (which apart from doing unwanted operations it may also generate permission problems for example). The fix makes the rules depend on `$(nix_PATH)` instead, which is the executable in the build tree.
2021-08-17 05:26:41 +03:00
endif
Move down fallback targets in `Makefile` This ensures `lib.mk` still defines `default` as the first target. This fixes some builds.
2023-12-20 10:23:49 +02:00
ifeq ($(ENABLE_INTERNAL_API_DOCS), yes)
Fix underlying build system so `--disable-build` works better - Internal API docs once again work - configure skips checks for a bunch of things it doesn't need
2023-12-04 01:12:22 +02:00
$(eval $(call include-sub-makefile, doc/internal-api/local.mk))
Split `--disable-tests`, fix cross builds It might seem obnoxious to have yet more configure flags, but I found controlling both the unit and functional tests with one flag was quite confusing because they are so different: - unit tests depending on building, functional tests don't (e.g. when we test already-built Nix) - unit tests can be installed, functional tests cannot - unit tests neeed extra libraries (GTest, RapidCheck), functional tests need extra executables (jq). - unit tests are run by `make check`, functional tests are run by `make installcheck` Really on a technical level, they seem wholly independent. Only on a human level ("they are both are tests") do they have anything in common. I had messed up the logic in cross builds because of this. Now I split the flag in two (and cleaned up a few other inconsistencies), and the logic fixed itself. Co-Authored-By: Robert Hensing <roberth@users.noreply.github.com>
2023-12-14 07:05:03 +02:00
else
.PHONY: internal-api-html
internal-api-html:
@echo "Internal API docs are disabled. Configure with '--enable-internal-api-docs', or avoid calling 'make internal-api-html'."
@exit 1
endif
Reference in a new issue Copy permalink