|
|
|
|
# Copyright (c) Facebook, Inc. and its affiliates.
|
|
|
|
|
#
|
|
|
|
|
# This source code is licensed under the MIT license found in the
|
|
|
|
|
# LICENSE file in the root directory of this source tree.
|
|
|
|
|
|
|
|
|
|
.PHONY: default
|
|
|
|
|
default: infer
|
|
|
|
|
|
|
|
|
|
ROOT_DIR = .
|
|
|
|
|
include $(ROOT_DIR)/Makefile.config
|
|
|
|
|
|
|
|
|
|
ORIG_SHELL_BUILD_MODE = $(BUILD_MODE)
|
[build] switch to 4.05.0+flambda by default
Summary:
With flambda (`-O3`), compilation time is ~5x slower, but the backend is ~25% faster!
To mitigate the atrocious compilation times, introduce a new `opt` build mode in the jbuilder files.
- build in "opt" mode by default from the toplevel (so that install scripts and external users get the fastest infer by default), in "default" mode by default from infer/src (since the latter is only called directly by infer devs, for faster builds)
- `make byte` is as fast as before in any mode
- `make test` will build "opt" by default, which is very slow. Solution for testing (or building the models) locally: `make BUILD_MODE=default test`.
- You can even change the default locally with `export BUILD_MODE=default`.
The benchmarks are to be taken with a sizable pinch of salt because I ran them only once and other stuff could be running in the background. That said, the perf win is consistent across all projects, with 15-20% win in wallclock time and around 25% win in total CPU time, ~9% win in sys time, and ~25% fewer minor allocations, and ~5-10% fewer overall allocations. This is only for the backend; the capture is by and large unaffected (either the same or a tad faster within noise range).
Here are the results running on OpenSSL 1.0.2d on osx (12 cores, 32G RAM)
=== base
infer binary: 26193088 bytes
compile time: 40s
capture:
```lang=text
real 1m7.513s
user 3m11.437s
sys 0m55.236s
```
analysis:
```lang=text
real 5m41.580s
user 61m37.855s
sys 1m12.870s
```
Memory profile:
```lang=json
{
...
"minor_gb": 0.1534719169139862,
"promoted_gb": 0.0038930922746658325,
"major_gb": 0.4546157643198967,
"allocated_gb": 0.6041945889592171,
"minor_collections": 78,
"major_collections": 23,
"compactions": 7,
"top_heap_gb": 0.07388687133789062,
"stack_kb": 0.3984375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda with stock options (no `-Oclassic`, just the same flags as base)
Exactly the same as base.
=== flambda `-O3`
infer binary: 56870376 bytes (2.17x bigger)
compile time: 191s (4.78x slower)
capture is the same as base:
```lang=text
real 1m9.203s
user 3m12.242s
sys 0m58.905s
```
analysis is ~20% wallclock time faster, ~25% CPU time faster:
```lang=text
real 4m32.656s
user 46m43.987s
sys 1m2.424s
```
memory usage is a bit lower too:
```lang=json
{
...
"minor_gb": 0.11583046615123749, // 75% of previous
"promoted_gb": 0.00363825261592865, // 93% of previous
"major_gb": 0.45415670424699783, // about same
"allocated_gb": 0.5663489177823067, // 94% of previous
"minor_collections": 73,
"major_collections": 22,
"compactions": 7,
"top_heap_gb": 0.07165145874023438,
"stack_kb": 0.3359375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda `-O2`
Not nearly as exciting as `-O3`, but the compilation cost is still quite high:
infer: 37826856 bytes
compilation of infer: 100s
Capture and analysis timings are mostly the same as base.
Reviewed By: jberdine
Differential Revision: D4867979
fbshipit-source-id: 99230b7
7 years ago
|
|
|
# override this for faster builds (but slower infer)
|
|
|
|
|
BUILD_MODE ?= opt
|
|
|
|
|
|
[RFC][build] Use dune environments and profiles instead of contexts
Summary:
With profiles and `(env ...)` stanza it's possible to consolidate
various ocamlc/ocamlopt/etc setups in a single place.
Where previously we needed to append `dune.common` to every dune file
and specify `flags` and `ocamlopt_flags` now the flags are specified
in `env` and applied accross the board.
This allows to
1. simplify build definitions,
2. avoid the need to generate dune files,
3. use plain sexps instead of OCaml and JBuilder plugin in build
files.
(I'll try to address 2 and 3 in the followup patches).
Existing `make` targets should continue working as before. Also, we
can use dune CLI like so:
```
infer/src$ dune printenv --profile opt # <- very useful for introspection
infer/src$ dune build check
infer/src$ dune build check --profile test
infer/src$ dune build infer.exe --profile dev
infer/src$ dune build infer.exe --profile opt
```
Also, with just 1 context something like `dune runtest` will run unit
tests only once instead of N times, where N is the number of contexts.
Now, there's one difference compared to the previous setup with
contexts:
- Previously, each context had its own build folder, and building infer
in opt context didn't invalidate any of the build artifacts in default
context. Therefore, alternating between `make` and `make opt` had low
overhead at the expense of having N copies of all build artifacts (1
for every context).
- Now, there's just 1 build folder and switching between profiles does
invalidate some artifacts (but not all) and rebuild takes a bit more
time.
So, if you're alternating like crazy between profiles your experience
may get worse (but not necessarily, more on that below). If you want
to trigger an opt build occasionally, you shouldn't notice much
difference.
For those who are concerned about slower build times when alternating
between different build profiles, there's a solution: [dune
cache](https://dune.readthedocs.io/en/stable/caching.html).
You can enable it by creating a file `~/.config/dune/config` with the
following contents:
```
(lang dune 2.0)
(cache enabled)
```
With cache enabled switching between different build profiles (but
also branches and commits) has ~0 overhead.
Dune cache works fine on Linux, but unfortunately there are [certain
problems with
MacOS](https://github.com/ocaml/dune/issues/3233) (hopefully, those
will be fixed soon and then there will be no downsides to using
profiles compared to contexts for our case).
Reviewed By: jvillard
Differential Revision: D20247864
fbshipit-source-id: 5f8afa0db
5 years ago
|
|
|
MAKE_SOURCE = $(MAKE) -C $(SRC_DIR)
|
[build] switch to 4.05.0+flambda by default
Summary:
With flambda (`-O3`), compilation time is ~5x slower, but the backend is ~25% faster!
To mitigate the atrocious compilation times, introduce a new `opt` build mode in the jbuilder files.
- build in "opt" mode by default from the toplevel (so that install scripts and external users get the fastest infer by default), in "default" mode by default from infer/src (since the latter is only called directly by infer devs, for faster builds)
- `make byte` is as fast as before in any mode
- `make test` will build "opt" by default, which is very slow. Solution for testing (or building the models) locally: `make BUILD_MODE=default test`.
- You can even change the default locally with `export BUILD_MODE=default`.
The benchmarks are to be taken with a sizable pinch of salt because I ran them only once and other stuff could be running in the background. That said, the perf win is consistent across all projects, with 15-20% win in wallclock time and around 25% win in total CPU time, ~9% win in sys time, and ~25% fewer minor allocations, and ~5-10% fewer overall allocations. This is only for the backend; the capture is by and large unaffected (either the same or a tad faster within noise range).
Here are the results running on OpenSSL 1.0.2d on osx (12 cores, 32G RAM)
=== base
infer binary: 26193088 bytes
compile time: 40s
capture:
```lang=text
real 1m7.513s
user 3m11.437s
sys 0m55.236s
```
analysis:
```lang=text
real 5m41.580s
user 61m37.855s
sys 1m12.870s
```
Memory profile:
```lang=json
{
...
"minor_gb": 0.1534719169139862,
"promoted_gb": 0.0038930922746658325,
"major_gb": 0.4546157643198967,
"allocated_gb": 0.6041945889592171,
"minor_collections": 78,
"major_collections": 23,
"compactions": 7,
"top_heap_gb": 0.07388687133789062,
"stack_kb": 0.3984375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda with stock options (no `-Oclassic`, just the same flags as base)
Exactly the same as base.
=== flambda `-O3`
infer binary: 56870376 bytes (2.17x bigger)
compile time: 191s (4.78x slower)
capture is the same as base:
```lang=text
real 1m9.203s
user 3m12.242s
sys 0m58.905s
```
analysis is ~20% wallclock time faster, ~25% CPU time faster:
```lang=text
real 4m32.656s
user 46m43.987s
sys 1m2.424s
```
memory usage is a bit lower too:
```lang=json
{
...
"minor_gb": 0.11583046615123749, // 75% of previous
"promoted_gb": 0.00363825261592865, // 93% of previous
"major_gb": 0.45415670424699783, // about same
"allocated_gb": 0.5663489177823067, // 94% of previous
"minor_collections": 73,
"major_collections": 22,
"compactions": 7,
"top_heap_gb": 0.07165145874023438,
"stack_kb": 0.3359375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda `-O2`
Not nearly as exciting as `-O3`, but the compilation cost is still quite high:
infer: 37826856 bytes
compilation of infer: 100s
Capture and analysis timings are mostly the same as base.
Reviewed By: jberdine
Differential Revision: D4867979
fbshipit-source-id: 99230b7
7 years ago
|
|
|
|
[make] s/ocamlbuild/jbuilder/g
Summary:
Use jbuilder to build infer instead of ocamlbuild. This is mainly to get faster builds:
```
times in 10ms, ±differences measured in speedups, 4 cores
| | ocb total | jb | ±total | ocb user | jb | ±user | ocb cpu | jb | ±cpu | ocb sys | jb | ±sys |
|-----------------------------------+-----------+------+--------+----------+------+-------+---------+-----+------+---------+------+------|
| byte from scratch | 6428 | 2456 | 2.62 | 7743 | 6662 | 1.16 | 138 | 331 | 2.40 | 1184 | 1477 | 0.80 |
| native from scratch | 9841 | 4289 | 2.29 | 9530 | 8834 | 1.08 | 110 | 245 | 2.23 | 1373 | 1712 | 0.80 |
| byte after native | 29578 | 1602 | 18.46 | 4514 | 4640 | 0.97 | 170 | 325 | 1.91 | 543 | 576 | 0.94 |
| change infer.ml byte | 344 | 282 | 1.22 | 292 | 215 | 1.36 | 96 | 99 | 1.03 | 040 | 066 | 0.61 |
| change infer.ml native | 837 | 223 | 3.75 | 789 | 174 | 4.53 | 98 | 99 | 1.01 | 036 | 47 | 0.77 |
| change Config.ml byte | 451 | 339 | 1.33 | 382 | 336 | 1.14 | 97 | 122 | 1.26 | 056 | 80 | 0.70 |
| change Config.ml native | 4024 | 1760 | 2.29 | 4585 | 4225 | 1.09 | 127 | 276 | 2.17 | 559 | 644 | 0.87 |
| change cFrontend_config.ml byte | 348 | 643 | 0.54 | 297 | 330 | 0.90 | 96 | 67 | 0.70 | 038 | 102 | 0.37 |
| change cFrontend_config.ml native | 1480 | 584 | 2.53 | 1435 | 906 | 1.58 | 106 | 185 | 1.75 | 136 | 178 | 0.76 |
#+TBLFM: $4=$2/$3;f2::$7=$5/$6;f2::$10=$9/$8;f2::$13=$11/$12;f2
50 cores
| | ocb total | jb | ±total | ocb user | jb | ±user | ocb cpu | jb | ±cpu | ocb sys | jb | ±sys |
|---------------------+-----------+------+--------+----------+------+-------+---------+----+------+---------+------+------|
| byte from scratch | 9114 | 2061 | 4.42 | 9334 | 5133 | 1.82 | | | 0/0 | 2566 | 1726 | 1.49 |
| native from scratch | 13481 | 3967 | 3.40 | 12291 | 7608 | 1.62 | | | 0/0 | 3003 | 2100 | 1.43 |
| byte after native | 3467 | 1476 | 2.35 | 5067 | 3912 | 1.30 | | | 0/0 | 971 | 801 | 1.21 |
#+TBLFM: $4=$2/$3;f2::$7=$5/$6;f2::$10=$9/$8;f2::$13=$11/$12;f2
```
Menu:
1. Write a jbuild file, autogenerated from jbuild.in because we need to fill in
some information at build-time (really, at configure time, but TODO), such as
whether or not clang is enabled.
2. Nuke lots of stuff from infer/src/Makefile that is now in the jbuild file
3. The jbuild file lives in infer/src/ so it can see all the sources. If we put it somewhere else, eg, infer/, then `jbuilder` scans too many files (all irrelevant) and takes 2.5s to start instead of .8s. Adding irrelevant directories to jbuild-ignore does not help.
4. jbuilder does not support subdirectories, so resort to listing all the
source files in the generated jbuild (only source directories need to be
manually listed in jbuild.in though). Still, the generated .merlin is wrong
and makes merlin find source files in _build, so manually tune it to get
good merlin support. We also lose some of merlin for unit tests as it
cannot see their build artefacts anymore.
5. checkCopyright gets its own jbuild because it's standalone. Also, remove
some deprecation warnings in checkCopyright due to the new version of Core from
a while ago.
6. Drop less-used Makefile features (they had regressed anyway) such as
building individual modules. Also, building mod_dep.pdf now takes all the
source files available so they better build (before, it would only take the
source files from the config, eg with or without clang) (that's pretty minor).
7. The toplevel is now built as a custom toplevel because that was easier. It
should soon be even easier: https://github.com/janestreet/jbuilder/issues/210
8. Move BUILTINS.mli to BUILTINS.ml because jbuilder is not happy about
interface files without implementations.
In particular, I did not try to migrate too much of the Makefile logic to jbuilder,
more can be done in the future.
Reviewed By: jberdine
Differential Revision: D5573661
fbshipit-source-id: 4ca6d8f
7 years ago
|
|
|
ifneq ($(UTOP),no)
|
|
|
|
|
BUILD_SYSTEMS_TESTS += infertop
|
[make] s/ocamlbuild/jbuilder/g
Summary:
Use jbuilder to build infer instead of ocamlbuild. This is mainly to get faster builds:
```
times in 10ms, ±differences measured in speedups, 4 cores
| | ocb total | jb | ±total | ocb user | jb | ±user | ocb cpu | jb | ±cpu | ocb sys | jb | ±sys |
|-----------------------------------+-----------+------+--------+----------+------+-------+---------+-----+------+---------+------+------|
| byte from scratch | 6428 | 2456 | 2.62 | 7743 | 6662 | 1.16 | 138 | 331 | 2.40 | 1184 | 1477 | 0.80 |
| native from scratch | 9841 | 4289 | 2.29 | 9530 | 8834 | 1.08 | 110 | 245 | 2.23 | 1373 | 1712 | 0.80 |
| byte after native | 29578 | 1602 | 18.46 | 4514 | 4640 | 0.97 | 170 | 325 | 1.91 | 543 | 576 | 0.94 |
| change infer.ml byte | 344 | 282 | 1.22 | 292 | 215 | 1.36 | 96 | 99 | 1.03 | 040 | 066 | 0.61 |
| change infer.ml native | 837 | 223 | 3.75 | 789 | 174 | 4.53 | 98 | 99 | 1.01 | 036 | 47 | 0.77 |
| change Config.ml byte | 451 | 339 | 1.33 | 382 | 336 | 1.14 | 97 | 122 | 1.26 | 056 | 80 | 0.70 |
| change Config.ml native | 4024 | 1760 | 2.29 | 4585 | 4225 | 1.09 | 127 | 276 | 2.17 | 559 | 644 | 0.87 |
| change cFrontend_config.ml byte | 348 | 643 | 0.54 | 297 | 330 | 0.90 | 96 | 67 | 0.70 | 038 | 102 | 0.37 |
| change cFrontend_config.ml native | 1480 | 584 | 2.53 | 1435 | 906 | 1.58 | 106 | 185 | 1.75 | 136 | 178 | 0.76 |
#+TBLFM: $4=$2/$3;f2::$7=$5/$6;f2::$10=$9/$8;f2::$13=$11/$12;f2
50 cores
| | ocb total | jb | ±total | ocb user | jb | ±user | ocb cpu | jb | ±cpu | ocb sys | jb | ±sys |
|---------------------+-----------+------+--------+----------+------+-------+---------+----+------+---------+------+------|
| byte from scratch | 9114 | 2061 | 4.42 | 9334 | 5133 | 1.82 | | | 0/0 | 2566 | 1726 | 1.49 |
| native from scratch | 13481 | 3967 | 3.40 | 12291 | 7608 | 1.62 | | | 0/0 | 3003 | 2100 | 1.43 |
| byte after native | 3467 | 1476 | 2.35 | 5067 | 3912 | 1.30 | | | 0/0 | 971 | 801 | 1.21 |
#+TBLFM: $4=$2/$3;f2::$7=$5/$6;f2::$10=$9/$8;f2::$13=$11/$12;f2
```
Menu:
1. Write a jbuild file, autogenerated from jbuild.in because we need to fill in
some information at build-time (really, at configure time, but TODO), such as
whether or not clang is enabled.
2. Nuke lots of stuff from infer/src/Makefile that is now in the jbuild file
3. The jbuild file lives in infer/src/ so it can see all the sources. If we put it somewhere else, eg, infer/, then `jbuilder` scans too many files (all irrelevant) and takes 2.5s to start instead of .8s. Adding irrelevant directories to jbuild-ignore does not help.
4. jbuilder does not support subdirectories, so resort to listing all the
source files in the generated jbuild (only source directories need to be
manually listed in jbuild.in though). Still, the generated .merlin is wrong
and makes merlin find source files in _build, so manually tune it to get
good merlin support. We also lose some of merlin for unit tests as it
cannot see their build artefacts anymore.
5. checkCopyright gets its own jbuild because it's standalone. Also, remove
some deprecation warnings in checkCopyright due to the new version of Core from
a while ago.
6. Drop less-used Makefile features (they had regressed anyway) such as
building individual modules. Also, building mod_dep.pdf now takes all the
source files available so they better build (before, it would only take the
source files from the config, eg with or without clang) (that's pretty minor).
7. The toplevel is now built as a custom toplevel because that was easier. It
should soon be even easier: https://github.com/janestreet/jbuilder/issues/210
8. Move BUILTINS.mli to BUILTINS.ml because jbuilder is not happy about
interface files without implementations.
In particular, I did not try to migrate too much of the Makefile logic to jbuilder,
more can be done in the future.
Reviewed By: jberdine
Differential Revision: D5573661
fbshipit-source-id: 4ca6d8f
7 years ago
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
ifeq ($(BUILD_C_ANALYZERS),yes)
|
|
|
|
|
BUILD_SYSTEMS_TESTS += \
|
|
|
|
|
annotation-reachability-sources-override \
|
|
|
|
|
assembly \
|
|
|
|
|
backtrack_level \
|
|
|
|
|
ck_analytics ck_imports \
|
|
|
|
|
clang_compilation_db_escaped clang_compilation_db_relpath \
|
|
|
|
|
clang_multiple_files \
|
|
|
|
|
clang_translation \
|
|
|
|
|
clang_unknown_ext \
|
|
|
|
|
clang_with_blacklisted_flags \
|
|
|
|
|
clang_with_E_flag \
|
|
|
|
|
clang_with_M_flag \
|
|
|
|
|
clang_with_MD_flag \
|
|
|
|
|
deduplicate_template_warnings \
|
|
|
|
|
delete_results_dir \
|
|
|
|
|
duplicate_symbols \
|
|
|
|
|
fail_on_issue \
|
|
|
|
|
j1 \
|
|
|
|
|
linters \
|
|
|
|
|
project_root_rel \
|
|
|
|
|
reactive \
|
|
|
|
|
run_hidden_linters \
|
|
|
|
|
tracebugs \
|
|
|
|
|
utf8_in_procname \
|
|
|
|
|
export_changed_functions \
|
|
|
|
|
incremental_analysis_remove_file \
|
|
|
|
|
incremental_analysis_change_procedure \
|
|
|
|
|
incremental_analysis_add_procedure \
|
|
|
|
|
|
|
|
|
|
DIRECT_TESTS += \
|
|
|
|
|
c_biabduction \
|
|
|
|
|
c_bufferoverrun \
|
|
|
|
|
c_frontend \
|
|
|
|
|
c_performance \
|
|
|
|
|
c_pulse \
|
|
|
|
|
c_purity \
|
|
|
|
|
c_uninit \
|
|
|
|
|
cpp_annotation-reachability \
|
|
|
|
|
cpp_biabduction \
|
|
|
|
|
cpp_bufferoverrun \
|
|
|
|
|
cpp_conflicts \
|
|
|
|
|
cpp_frontend \
|
|
|
|
|
cpp_impurity \
|
|
|
|
|
cpp_linters \
|
|
|
|
|
cpp_linters-for-test-only \
|
|
|
|
|
cpp_liveness \
|
|
|
|
|
cpp_pulse \
|
|
|
|
|
cpp_quandary \
|
|
|
|
|
cpp_racerd \
|
|
|
|
|
cpp_siof \
|
|
|
|
|
cpp_starvation \
|
|
|
|
|
cpp_uninit \
|
|
|
|
|
|
|
|
|
|
ifeq ($(IS_FACEBOOK_TREE),yes)
|
|
|
|
|
DIRECT_TESTS += cpp_fb-taint
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
ifneq ($(BUCK),no)
|
|
|
|
|
BUILD_SYSTEMS_TESTS += \
|
|
|
|
|
buck_blacklist \
|
|
|
|
|
buck-clang-db \
|
|
|
|
|
buck_clang_test_determinator \
|
|
|
|
|
buck_flavors \
|
|
|
|
|
buck_flavors_diff \
|
|
|
|
|
buck_flavors_run \
|
|
|
|
|
buck_flavors_deterministic \
|
|
|
|
|
buck_export_changed_functions \
|
|
|
|
|
|
|
|
|
|
endif
|
|
|
|
|
ifneq ($(CMAKE),no)
|
|
|
|
|
BUILD_SYSTEMS_TESTS += clang_compilation_db cmake inferconfig inferconfig_not_strict
|
|
|
|
|
endif
|
|
|
|
|
ifneq ($(NDKBUILD),no)
|
|
|
|
|
BUILD_SYSTEMS_TESTS += ndk_build
|
|
|
|
|
endif
|
|
|
|
|
ifeq ($(HAS_OBJC),yes)
|
|
|
|
|
BUILD_SYSTEMS_TESTS += \
|
|
|
|
|
clang_test_determinator \
|
|
|
|
|
objc_getters_setters \
|
|
|
|
|
objc_missing_fld \
|
|
|
|
|
objc_retain_cycles \
|
|
|
|
|
objc_retain_cycles_weak
|
|
|
|
|
|
|
|
|
|
DIRECT_TESTS += \
|
|
|
|
|
objc_biabduction \
|
|
|
|
|
objc_frontend \
|
|
|
|
|
objc_ioslints \
|
|
|
|
|
objc_linters \
|
|
|
|
|
objc_linters-def-folder \
|
|
|
|
|
objc_linters-for-test-only \
|
|
|
|
|
objc_liveness \
|
|
|
|
|
objc_performance \
|
|
|
|
|
objc_pulse \
|
|
|
|
|
objc_quandary \
|
|
|
|
|
objc_self-in-block \
|
|
|
|
|
objc_uninit \
|
|
|
|
|
objcpp_biabduction \
|
|
|
|
|
objcpp_frontend \
|
|
|
|
|
objcpp_linters \
|
|
|
|
|
objcpp_linters-for-test-only \
|
|
|
|
|
objcpp_liveness \
|
|
|
|
|
objcpp_pulse \
|
|
|
|
|
objcpp_racerd \
|
|
|
|
|
objcpp_retain-cycles \
|
|
|
|
|
|
|
|
|
|
ifneq ($(XCODE_SELECT),no)
|
|
|
|
|
BUILD_SYSTEMS_TESTS += xcodebuild_no_xcpretty
|
|
|
|
|
endif
|
|
|
|
|
ifneq ($(XCPRETTY),no)
|
|
|
|
|
BUILD_SYSTEMS_TESTS += xcodebuild
|
|
|
|
|
endif
|
|
|
|
|
endif # HAS_OBJC
|
|
|
|
|
endif # BUILD_C_ANALYZERS
|
|
|
|
|
|
|
|
|
|
ifeq ($(BUILD_JAVA_ANALYZERS),yes)
|
|
|
|
|
BUILD_SYSTEMS_TESTS += \
|
|
|
|
|
differential_interesting_paths_filter \
|
|
|
|
|
differential_of_costs_report \
|
|
|
|
|
incremental_analysis_cost_change \
|
|
|
|
|
differential_skip_anonymous_class_renamings \
|
|
|
|
|
differential_skip_duplicated_types_on_filenames \
|
|
|
|
|
differential_skip_duplicated_types_on_filenames_with_renamings \
|
|
|
|
|
gradle \
|
|
|
|
|
java_test_determinator \
|
|
|
|
|
javac \
|
|
|
|
|
resource_leak_exception_lines \
|
|
|
|
|
racerd_dedup
|
|
|
|
|
|
|
|
|
|
#TODO T41549034: Jdk11 translates string append differently, causing
|
|
|
|
|
#test failures in NullPointerExceptions:stringVarEqualsFalseNPE
|
|
|
|
|
|
|
|
|
|
DIRECT_TESTS += \
|
|
|
|
|
java_annotreach \
|
|
|
|
|
java_biabduction \
|
|
|
|
|
java_bufferoverrun \
|
|
|
|
|
java_checkers \
|
|
|
|
|
java_classloads \
|
|
|
|
|
java_nullsafe \
|
|
|
|
|
java_hoisting \
|
|
|
|
|
java_hoistingExpensive \
|
|
|
|
|
java_impurity \
|
|
|
|
|
java_inefficientKeysetIterator \
|
|
|
|
|
java_litho-required-props \
|
|
|
|
|
java_performance \
|
|
|
|
|
java_performance-exclusive \
|
|
|
|
|
java_purity \
|
|
|
|
|
java_quandary \
|
|
|
|
|
java_racerd \
|
|
|
|
|
java_starvation \
|
|
|
|
|
java_starvation-dedup \
|
|
|
|
|
java_starvation-whole-program \
|
|
|
|
|
java_topl \
|
|
|
|
|
|
|
|
|
|
ifeq ($(IS_FACEBOOK_TREE),yes)
|
|
|
|
|
DIRECT_TESTS += java_fb-performance
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
ifneq ($(ANT),no)
|
|
|
|
|
BUILD_SYSTEMS_TESTS += ant
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ifneq ($(BUCK),no)
|
|
|
|
|
BUILD_SYSTEMS_TESTS += genrulecapture
|
|
|
|
|
endif
|
|
|
|
|
ifneq ($(MVN),no)
|
|
|
|
|
BUILD_SYSTEMS_TESTS += mvn
|
|
|
|
|
endif
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
ifeq ($(BUILD_C_ANALYZERS)+$(BUILD_JAVA_ANALYZERS),yes+yes)
|
|
|
|
|
BUILD_SYSTEMS_TESTS += make utf8_in_pwd waf
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
ifeq ($(IS_INFER_RELEASE),no)
|
|
|
|
|
configure: configure.ac $(wildcard m4/*.m4)
|
|
|
|
|
# rerun ./autogen.sh in case of failure as the failure may be due to needing to rerun
|
|
|
|
|
# ./configure
|
|
|
|
|
$(QUIET)($(call silent_on_success,Generate ./configure,./autogen.sh)) || \
|
|
|
|
|
./autogen.sh
|
|
|
|
|
|
|
|
|
|
Makefile.autoconf: configure Makefile.autoconf.in
|
|
|
|
|
# rerun ./configure with the flags that were used last time it was run (if available)
|
|
|
|
|
# retry in case of failure as the failure may be due to needing to rerun ./configure
|
|
|
|
|
$(QUIET)($(call silent_on_success,Running\
|
|
|
|
|
./configure $(shell ./config.status --config || true),\
|
|
|
|
|
./configure $(shell ./config.status --config || true))) || \
|
|
|
|
|
./configure $(shell ./config.status --config || true)
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
.PHONY: fb-setup
|
|
|
|
|
fb-setup:
|
|
|
|
|
$(QUIET)$(call silent_on_success,Facebook setup,\
|
|
|
|
|
$(MAKE) -C facebook setup)
|
|
|
|
|
|
|
|
|
|
.PHONY: fmt
|
|
|
|
|
fmt:
|
|
|
|
|
parallel $(OCAMLFORMAT_EXE) $(OCAMLFORMAT_ARGS) -i ::: $$(git diff --name-only --diff-filter=ACMRU $$(git merge-base origin/master HEAD) | grep "\.mli\?$$")
|
|
|
|
|
|
|
|
|
|
DUNE_ML:=$(shell find * -name 'dune*.in' | grep -v workspace | grep -v infer-source)
|
|
|
|
|
|
|
|
|
|
.PHONY: fmt_dune
|
|
|
|
|
fmt_dune:
|
|
|
|
|
parallel $(OCAMLFORMAT_EXE) $(OCAMLFORMAT_ARGS) -i ::: $(DUNE_ML)
|
|
|
|
|
|
|
|
|
|
SRC_ML:=$(shell find * \( -name _build -or -name facebook-clang-plugins -or -path facebook/dependencies -or -path sledge/llvm -or -path sledge/.llvm_build \) -not -prune -or -type f -and -name '*'.ml -or -name '*'.mli 2>/dev/null)
|
|
|
|
|
|
|
|
|
|
.PHONY: fmt_all
|
|
|
|
|
fmt_all:
|
|
|
|
|
parallel $(OCAMLFORMAT_EXE) $(OCAMLFORMAT_ARGS) -i ::: $(SRC_ML) $(DUNE_ML)
|
|
|
|
|
|
|
|
|
|
# pre-building these avoids race conditions when building, eg src_build and test_build in parallel
|
|
|
|
|
.PHONY: src_build_common
|
|
|
|
|
src_build_common:
|
|
|
|
|
$(QUIET)$(call silent_on_success,Generating source dependencies,\
|
[build] switch to 4.05.0+flambda by default
Summary:
With flambda (`-O3`), compilation time is ~5x slower, but the backend is ~25% faster!
To mitigate the atrocious compilation times, introduce a new `opt` build mode in the jbuilder files.
- build in "opt" mode by default from the toplevel (so that install scripts and external users get the fastest infer by default), in "default" mode by default from infer/src (since the latter is only called directly by infer devs, for faster builds)
- `make byte` is as fast as before in any mode
- `make test` will build "opt" by default, which is very slow. Solution for testing (or building the models) locally: `make BUILD_MODE=default test`.
- You can even change the default locally with `export BUILD_MODE=default`.
The benchmarks are to be taken with a sizable pinch of salt because I ran them only once and other stuff could be running in the background. That said, the perf win is consistent across all projects, with 15-20% win in wallclock time and around 25% win in total CPU time, ~9% win in sys time, and ~25% fewer minor allocations, and ~5-10% fewer overall allocations. This is only for the backend; the capture is by and large unaffected (either the same or a tad faster within noise range).
Here are the results running on OpenSSL 1.0.2d on osx (12 cores, 32G RAM)
=== base
infer binary: 26193088 bytes
compile time: 40s
capture:
```lang=text
real 1m7.513s
user 3m11.437s
sys 0m55.236s
```
analysis:
```lang=text
real 5m41.580s
user 61m37.855s
sys 1m12.870s
```
Memory profile:
```lang=json
{
...
"minor_gb": 0.1534719169139862,
"promoted_gb": 0.0038930922746658325,
"major_gb": 0.4546157643198967,
"allocated_gb": 0.6041945889592171,
"minor_collections": 78,
"major_collections": 23,
"compactions": 7,
"top_heap_gb": 0.07388687133789062,
"stack_kb": 0.3984375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda with stock options (no `-Oclassic`, just the same flags as base)
Exactly the same as base.
=== flambda `-O3`
infer binary: 56870376 bytes (2.17x bigger)
compile time: 191s (4.78x slower)
capture is the same as base:
```lang=text
real 1m9.203s
user 3m12.242s
sys 0m58.905s
```
analysis is ~20% wallclock time faster, ~25% CPU time faster:
```lang=text
real 4m32.656s
user 46m43.987s
sys 1m2.424s
```
memory usage is a bit lower too:
```lang=json
{
...
"minor_gb": 0.11583046615123749, // 75% of previous
"promoted_gb": 0.00363825261592865, // 93% of previous
"major_gb": 0.45415670424699783, // about same
"allocated_gb": 0.5663489177823067, // 94% of previous
"minor_collections": 73,
"major_collections": 22,
"compactions": 7,
"top_heap_gb": 0.07165145874023438,
"stack_kb": 0.3359375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda `-O2`
Not nearly as exciting as `-O3`, but the compilation cost is still quite high:
infer: 37826856 bytes
compilation of infer: 100s
Capture and analysis timings are mostly the same as base.
Reviewed By: jberdine
Differential Revision: D4867979
fbshipit-source-id: 99230b7
7 years ago
|
|
|
$(MAKE_SOURCE) src_build_common)
|
|
|
|
|
|
|
|
|
|
.PHONY: src_build
|
|
|
|
|
src_build: src_build_common
|
|
|
|
|
$(QUIET)$(call silent_on_success,Building native($(BUILD_MODE)) Infer,\
|
[build] switch to 4.05.0+flambda by default
Summary:
With flambda (`-O3`), compilation time is ~5x slower, but the backend is ~25% faster!
To mitigate the atrocious compilation times, introduce a new `opt` build mode in the jbuilder files.
- build in "opt" mode by default from the toplevel (so that install scripts and external users get the fastest infer by default), in "default" mode by default from infer/src (since the latter is only called directly by infer devs, for faster builds)
- `make byte` is as fast as before in any mode
- `make test` will build "opt" by default, which is very slow. Solution for testing (or building the models) locally: `make BUILD_MODE=default test`.
- You can even change the default locally with `export BUILD_MODE=default`.
The benchmarks are to be taken with a sizable pinch of salt because I ran them only once and other stuff could be running in the background. That said, the perf win is consistent across all projects, with 15-20% win in wallclock time and around 25% win in total CPU time, ~9% win in sys time, and ~25% fewer minor allocations, and ~5-10% fewer overall allocations. This is only for the backend; the capture is by and large unaffected (either the same or a tad faster within noise range).
Here are the results running on OpenSSL 1.0.2d on osx (12 cores, 32G RAM)
=== base
infer binary: 26193088 bytes
compile time: 40s
capture:
```lang=text
real 1m7.513s
user 3m11.437s
sys 0m55.236s
```
analysis:
```lang=text
real 5m41.580s
user 61m37.855s
sys 1m12.870s
```
Memory profile:
```lang=json
{
...
"minor_gb": 0.1534719169139862,
"promoted_gb": 0.0038930922746658325,
"major_gb": 0.4546157643198967,
"allocated_gb": 0.6041945889592171,
"minor_collections": 78,
"major_collections": 23,
"compactions": 7,
"top_heap_gb": 0.07388687133789062,
"stack_kb": 0.3984375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda with stock options (no `-Oclassic`, just the same flags as base)
Exactly the same as base.
=== flambda `-O3`
infer binary: 56870376 bytes (2.17x bigger)
compile time: 191s (4.78x slower)
capture is the same as base:
```lang=text
real 1m9.203s
user 3m12.242s
sys 0m58.905s
```
analysis is ~20% wallclock time faster, ~25% CPU time faster:
```lang=text
real 4m32.656s
user 46m43.987s
sys 1m2.424s
```
memory usage is a bit lower too:
```lang=json
{
...
"minor_gb": 0.11583046615123749, // 75% of previous
"promoted_gb": 0.00363825261592865, // 93% of previous
"major_gb": 0.45415670424699783, // about same
"allocated_gb": 0.5663489177823067, // 94% of previous
"minor_collections": 73,
"major_collections": 22,
"compactions": 7,
"top_heap_gb": 0.07165145874023438,
"stack_kb": 0.3359375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda `-O2`
Not nearly as exciting as `-O3`, but the compilation cost is still quite high:
infer: 37826856 bytes
compilation of infer: 100s
Capture and analysis timings are mostly the same as base.
Reviewed By: jberdine
Differential Revision: D4867979
fbshipit-source-id: 99230b7
7 years ago
|
|
|
$(MAKE_SOURCE) infer)
|
|
|
|
|
|
|
|
|
|
.PHONY: byte
|
|
|
|
|
byte: src_build_common
|
|
|
|
|
$(QUIET)$(call silent_on_success,Building byte Infer,\
|
[build] switch to 4.05.0+flambda by default
Summary:
With flambda (`-O3`), compilation time is ~5x slower, but the backend is ~25% faster!
To mitigate the atrocious compilation times, introduce a new `opt` build mode in the jbuilder files.
- build in "opt" mode by default from the toplevel (so that install scripts and external users get the fastest infer by default), in "default" mode by default from infer/src (since the latter is only called directly by infer devs, for faster builds)
- `make byte` is as fast as before in any mode
- `make test` will build "opt" by default, which is very slow. Solution for testing (or building the models) locally: `make BUILD_MODE=default test`.
- You can even change the default locally with `export BUILD_MODE=default`.
The benchmarks are to be taken with a sizable pinch of salt because I ran them only once and other stuff could be running in the background. That said, the perf win is consistent across all projects, with 15-20% win in wallclock time and around 25% win in total CPU time, ~9% win in sys time, and ~25% fewer minor allocations, and ~5-10% fewer overall allocations. This is only for the backend; the capture is by and large unaffected (either the same or a tad faster within noise range).
Here are the results running on OpenSSL 1.0.2d on osx (12 cores, 32G RAM)
=== base
infer binary: 26193088 bytes
compile time: 40s
capture:
```lang=text
real 1m7.513s
user 3m11.437s
sys 0m55.236s
```
analysis:
```lang=text
real 5m41.580s
user 61m37.855s
sys 1m12.870s
```
Memory profile:
```lang=json
{
...
"minor_gb": 0.1534719169139862,
"promoted_gb": 0.0038930922746658325,
"major_gb": 0.4546157643198967,
"allocated_gb": 0.6041945889592171,
"minor_collections": 78,
"major_collections": 23,
"compactions": 7,
"top_heap_gb": 0.07388687133789062,
"stack_kb": 0.3984375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda with stock options (no `-Oclassic`, just the same flags as base)
Exactly the same as base.
=== flambda `-O3`
infer binary: 56870376 bytes (2.17x bigger)
compile time: 191s (4.78x slower)
capture is the same as base:
```lang=text
real 1m9.203s
user 3m12.242s
sys 0m58.905s
```
analysis is ~20% wallclock time faster, ~25% CPU time faster:
```lang=text
real 4m32.656s
user 46m43.987s
sys 1m2.424s
```
memory usage is a bit lower too:
```lang=json
{
...
"minor_gb": 0.11583046615123749, // 75% of previous
"promoted_gb": 0.00363825261592865, // 93% of previous
"major_gb": 0.45415670424699783, // about same
"allocated_gb": 0.5663489177823067, // 94% of previous
"minor_collections": 73,
"major_collections": 22,
"compactions": 7,
"top_heap_gb": 0.07165145874023438,
"stack_kb": 0.3359375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda `-O2`
Not nearly as exciting as `-O3`, but the compilation cost is still quite high:
infer: 37826856 bytes
compilation of infer: 100s
Capture and analysis timings are mostly the same as base.
Reviewed By: jberdine
Differential Revision: D4867979
fbshipit-source-id: 99230b7
7 years ago
|
|
|
$(MAKE_SOURCE) byte)
|
|
|
|
|
|
|
|
|
|
.PHONY: check
|
|
|
|
|
check: src_build_common
|
|
|
|
|
$(QUIET)$(call silent_on_success,Building artifacts for tooling support,\
|
|
|
|
|
$(MAKE_SOURCE) check)
|
|
|
|
|
|
|
|
|
|
.PHONY: test_build
|
|
|
|
|
test_build: src_build_common
|
|
|
|
|
$(QUIET)$(call silent_on_success,Testing Infer builds without warnings,\
|
[build] switch to 4.05.0+flambda by default
Summary:
With flambda (`-O3`), compilation time is ~5x slower, but the backend is ~25% faster!
To mitigate the atrocious compilation times, introduce a new `opt` build mode in the jbuilder files.
- build in "opt" mode by default from the toplevel (so that install scripts and external users get the fastest infer by default), in "default" mode by default from infer/src (since the latter is only called directly by infer devs, for faster builds)
- `make byte` is as fast as before in any mode
- `make test` will build "opt" by default, which is very slow. Solution for testing (or building the models) locally: `make BUILD_MODE=default test`.
- You can even change the default locally with `export BUILD_MODE=default`.
The benchmarks are to be taken with a sizable pinch of salt because I ran them only once and other stuff could be running in the background. That said, the perf win is consistent across all projects, with 15-20% win in wallclock time and around 25% win in total CPU time, ~9% win in sys time, and ~25% fewer minor allocations, and ~5-10% fewer overall allocations. This is only for the backend; the capture is by and large unaffected (either the same or a tad faster within noise range).
Here are the results running on OpenSSL 1.0.2d on osx (12 cores, 32G RAM)
=== base
infer binary: 26193088 bytes
compile time: 40s
capture:
```lang=text
real 1m7.513s
user 3m11.437s
sys 0m55.236s
```
analysis:
```lang=text
real 5m41.580s
user 61m37.855s
sys 1m12.870s
```
Memory profile:
```lang=json
{
...
"minor_gb": 0.1534719169139862,
"promoted_gb": 0.0038930922746658325,
"major_gb": 0.4546157643198967,
"allocated_gb": 0.6041945889592171,
"minor_collections": 78,
"major_collections": 23,
"compactions": 7,
"top_heap_gb": 0.07388687133789062,
"stack_kb": 0.3984375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda with stock options (no `-Oclassic`, just the same flags as base)
Exactly the same as base.
=== flambda `-O3`
infer binary: 56870376 bytes (2.17x bigger)
compile time: 191s (4.78x slower)
capture is the same as base:
```lang=text
real 1m9.203s
user 3m12.242s
sys 0m58.905s
```
analysis is ~20% wallclock time faster, ~25% CPU time faster:
```lang=text
real 4m32.656s
user 46m43.987s
sys 1m2.424s
```
memory usage is a bit lower too:
```lang=json
{
...
"minor_gb": 0.11583046615123749, // 75% of previous
"promoted_gb": 0.00363825261592865, // 93% of previous
"major_gb": 0.45415670424699783, // about same
"allocated_gb": 0.5663489177823067, // 94% of previous
"minor_collections": 73,
"major_collections": 22,
"compactions": 7,
"top_heap_gb": 0.07165145874023438,
"stack_kb": 0.3359375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda `-O2`
Not nearly as exciting as `-O3`, but the compilation cost is still quite high:
infer: 37826856 bytes
compilation of infer: 100s
Capture and analysis timings are mostly the same as base.
Reviewed By: jberdine
Differential Revision: D4867979
fbshipit-source-id: 99230b7
7 years ago
|
|
|
$(MAKE_SOURCE) test)
|
|
|
|
|
|
|
|
|
|
# deadcode analysis: only do the deadcode detection on Facebook builds and if GNU sed is available
|
|
|
|
|
.PHONY: real_deadcode
|
|
|
|
|
real_deadcode: src_build_common
|
|
|
|
|
$(QUIET)$(call silent_on_success,Testing there is no dead OCaml code,\
|
|
|
|
|
$(MAKE) -C $(SRC_DIR)/deadcode)
|
|
|
|
|
|
|
|
|
|
.PHONY: deadcode
|
|
|
|
|
deadcode:
|
|
|
|
|
ifeq ($(IS_FACEBOOK_TREE),no)
|
|
|
|
|
$(QUIET)echo "Deadcode detection only works in Facebook builds, skipping"
|
|
|
|
|
endif
|
|
|
|
|
ifeq ($(GNU_SED),no)
|
|
|
|
|
$(QUIET)echo "Deadcode detection only works with GNU sed installed, skipping"
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
ifeq ($(IS_FACEBOOK_TREE),yes)
|
|
|
|
|
ifneq ($(GNU_SED),no)
|
|
|
|
|
deadcode: real_deadcode
|
|
|
|
|
endif
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
.PHONY: toplevel toplevel_test
|
|
|
|
|
toplevel toplevel_test: src_build_common
|
|
|
|
|
|
|
|
|
|
toplevel:
|
|
|
|
|
$(QUIET)$(call silent_on_success,Building Infer REPL,\
|
|
|
|
|
$(MAKE_SOURCE) toplevel)
|
|
|
|
|
$(QUIET)echo
|
|
|
|
|
$(QUIET)echo "You can now use the infer REPL:"
|
|
|
|
|
$(QUIET)echo " \"$(ABSOLUTE_ROOT_DIR)/scripts/infer_repl\""
|
|
|
|
|
|
[RFC][build] Use dune environments and profiles instead of contexts
Summary:
With profiles and `(env ...)` stanza it's possible to consolidate
various ocamlc/ocamlopt/etc setups in a single place.
Where previously we needed to append `dune.common` to every dune file
and specify `flags` and `ocamlopt_flags` now the flags are specified
in `env` and applied accross the board.
This allows to
1. simplify build definitions,
2. avoid the need to generate dune files,
3. use plain sexps instead of OCaml and JBuilder plugin in build
files.
(I'll try to address 2 and 3 in the followup patches).
Existing `make` targets should continue working as before. Also, we
can use dune CLI like so:
```
infer/src$ dune printenv --profile opt # <- very useful for introspection
infer/src$ dune build check
infer/src$ dune build check --profile test
infer/src$ dune build infer.exe --profile dev
infer/src$ dune build infer.exe --profile opt
```
Also, with just 1 context something like `dune runtest` will run unit
tests only once instead of N times, where N is the number of contexts.
Now, there's one difference compared to the previous setup with
contexts:
- Previously, each context had its own build folder, and building infer
in opt context didn't invalidate any of the build artifacts in default
context. Therefore, alternating between `make` and `make opt` had low
overhead at the expense of having N copies of all build artifacts (1
for every context).
- Now, there's just 1 build folder and switching between profiles does
invalidate some artifacts (but not all) and rebuild takes a bit more
time.
So, if you're alternating like crazy between profiles your experience
may get worse (but not necessarily, more on that below). If you want
to trigger an opt build occasionally, you shouldn't notice much
difference.
For those who are concerned about slower build times when alternating
between different build profiles, there's a solution: [dune
cache](https://dune.readthedocs.io/en/stable/caching.html).
You can enable it by creating a file `~/.config/dune/config` with the
following contents:
```
(lang dune 2.0)
(cache enabled)
```
With cache enabled switching between different build profiles (but
also branches and commits) has ~0 overhead.
Dune cache works fine on Linux, but unfortunately there are [certain
problems with
MacOS](https://github.com/ocaml/dune/issues/3233) (hopefully, those
will be fixed soon and then there will be no downsides to using
profiles compared to contexts for our case).
Reviewed By: jvillard
Differential Revision: D20247864
fbshipit-source-id: 5f8afa0db
5 years ago
|
|
|
toplevel_test:
|
|
|
|
|
$(QUIET)$(call silent_on_success,Building Infer REPL,\
|
|
|
|
|
$(MAKE_SOURCE) toplevel)
|
|
|
|
|
|
|
|
|
|
ifeq ($(IS_FACEBOOK_TREE),yes)
|
[make] s/ocamlbuild/jbuilder/g
Summary:
Use jbuilder to build infer instead of ocamlbuild. This is mainly to get faster builds:
```
times in 10ms, ±differences measured in speedups, 4 cores
| | ocb total | jb | ±total | ocb user | jb | ±user | ocb cpu | jb | ±cpu | ocb sys | jb | ±sys |
|-----------------------------------+-----------+------+--------+----------+------+-------+---------+-----+------+---------+------+------|
| byte from scratch | 6428 | 2456 | 2.62 | 7743 | 6662 | 1.16 | 138 | 331 | 2.40 | 1184 | 1477 | 0.80 |
| native from scratch | 9841 | 4289 | 2.29 | 9530 | 8834 | 1.08 | 110 | 245 | 2.23 | 1373 | 1712 | 0.80 |
| byte after native | 29578 | 1602 | 18.46 | 4514 | 4640 | 0.97 | 170 | 325 | 1.91 | 543 | 576 | 0.94 |
| change infer.ml byte | 344 | 282 | 1.22 | 292 | 215 | 1.36 | 96 | 99 | 1.03 | 040 | 066 | 0.61 |
| change infer.ml native | 837 | 223 | 3.75 | 789 | 174 | 4.53 | 98 | 99 | 1.01 | 036 | 47 | 0.77 |
| change Config.ml byte | 451 | 339 | 1.33 | 382 | 336 | 1.14 | 97 | 122 | 1.26 | 056 | 80 | 0.70 |
| change Config.ml native | 4024 | 1760 | 2.29 | 4585 | 4225 | 1.09 | 127 | 276 | 2.17 | 559 | 644 | 0.87 |
| change cFrontend_config.ml byte | 348 | 643 | 0.54 | 297 | 330 | 0.90 | 96 | 67 | 0.70 | 038 | 102 | 0.37 |
| change cFrontend_config.ml native | 1480 | 584 | 2.53 | 1435 | 906 | 1.58 | 106 | 185 | 1.75 | 136 | 178 | 0.76 |
#+TBLFM: $4=$2/$3;f2::$7=$5/$6;f2::$10=$9/$8;f2::$13=$11/$12;f2
50 cores
| | ocb total | jb | ±total | ocb user | jb | ±user | ocb cpu | jb | ±cpu | ocb sys | jb | ±sys |
|---------------------+-----------+------+--------+----------+------+-------+---------+----+------+---------+------+------|
| byte from scratch | 9114 | 2061 | 4.42 | 9334 | 5133 | 1.82 | | | 0/0 | 2566 | 1726 | 1.49 |
| native from scratch | 13481 | 3967 | 3.40 | 12291 | 7608 | 1.62 | | | 0/0 | 3003 | 2100 | 1.43 |
| byte after native | 3467 | 1476 | 2.35 | 5067 | 3912 | 1.30 | | | 0/0 | 971 | 801 | 1.21 |
#+TBLFM: $4=$2/$3;f2::$7=$5/$6;f2::$10=$9/$8;f2::$13=$11/$12;f2
```
Menu:
1. Write a jbuild file, autogenerated from jbuild.in because we need to fill in
some information at build-time (really, at configure time, but TODO), such as
whether or not clang is enabled.
2. Nuke lots of stuff from infer/src/Makefile that is now in the jbuild file
3. The jbuild file lives in infer/src/ so it can see all the sources. If we put it somewhere else, eg, infer/, then `jbuilder` scans too many files (all irrelevant) and takes 2.5s to start instead of .8s. Adding irrelevant directories to jbuild-ignore does not help.
4. jbuilder does not support subdirectories, so resort to listing all the
source files in the generated jbuild (only source directories need to be
manually listed in jbuild.in though). Still, the generated .merlin is wrong
and makes merlin find source files in _build, so manually tune it to get
good merlin support. We also lose some of merlin for unit tests as it
cannot see their build artefacts anymore.
5. checkCopyright gets its own jbuild because it's standalone. Also, remove
some deprecation warnings in checkCopyright due to the new version of Core from
a while ago.
6. Drop less-used Makefile features (they had regressed anyway) such as
building individual modules. Also, building mod_dep.pdf now takes all the
source files available so they better build (before, it would only take the
source files from the config, eg with or without clang) (that's pretty minor).
7. The toplevel is now built as a custom toplevel because that was easier. It
should soon be even easier: https://github.com/janestreet/jbuilder/issues/210
8. Move BUILTINS.mli to BUILTINS.ml because jbuilder is not happy about
interface files without implementations.
In particular, I did not try to migrate too much of the Makefile logic to jbuilder,
more can be done in the future.
Reviewed By: jberdine
Differential Revision: D5573661
fbshipit-source-id: 4ca6d8f
7 years ago
|
|
|
byte src_build_common src_build test_build: fb-setup
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
ifeq ($(BUILD_C_ANALYZERS),yes)
|
|
|
|
|
byte src_build src_build_common test_build: clang_plugin
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
$(INFER_COMMAND_MANUALS): src_build $(MAKEFILE_LIST)
|
|
|
|
|
$(QUIET)$(MKDIR_P) $(@D)
|
|
|
|
|
$(QUIET)$(INFER_BIN) $(patsubst infer-%.1,%,$(@F)) --help-scrubbed --help-format=groff > $@
|
|
|
|
|
|
|
|
|
|
$(INFER_COMMAND_TEXT_MANUALS): src_build $(MAKEFILE_LIST)
|
|
|
|
|
$(QUIET)$(MKDIR_P) $(@D)
|
|
|
|
|
$(QUIET)$(INFER_BIN) $(patsubst infer-%.txt,%,$(@F)) --help-scrubbed --help-format=plain > $@
|
|
|
|
|
|
|
|
|
|
$(INFER_MANUAL): src_build $(MAKEFILE_LIST)
|
|
|
|
|
$(QUIET)$(MKDIR_P) $(@D)
|
|
|
|
|
$(QUIET)$(INFER_BIN) --help-scrubbed --help-format=groff > $@
|
|
|
|
|
|
|
|
|
|
$(INFER_TEXT_MANUAL): src_build $(MAKEFILE_LIST)
|
|
|
|
|
$(QUIET)$(MKDIR_P) $(@D)
|
|
|
|
|
$(QUIET)$(INFER_BIN) --help-scrubbed --help-format=plain > $@
|
|
|
|
|
|
|
|
|
|
$(INFER_FULL_TEXT_MANUAL): src_build $(MAKEFILE_LIST)
|
|
|
|
|
$(QUIET)$(MKDIR_P) $(@D)
|
|
|
|
|
$(QUIET)$(INFER_BIN) --help-scrubbed-full --help-format=plain > $@
|
|
|
|
|
|
|
|
|
|
$(INFER_GROFF_MANUALS_GZIPPED): %.gz: %
|
|
|
|
|
$(QUIET)$(REMOVE) $@
|
|
|
|
|
gzip $<
|
|
|
|
|
|
|
|
|
|
manuals:
|
|
|
|
|
$(QUIET)$(call silent_on_success,Building Infer manuals,\
|
|
|
|
|
$(MAKE) $(INFER_MANUALS))
|
|
|
|
|
|
|
|
|
|
infer_models: src_build
|
|
|
|
|
ifeq ($(BUILD_JAVA_ANALYZERS),yes)
|
[RFC][build] Use dune environments and profiles instead of contexts
Summary:
With profiles and `(env ...)` stanza it's possible to consolidate
various ocamlc/ocamlopt/etc setups in a single place.
Where previously we needed to append `dune.common` to every dune file
and specify `flags` and `ocamlopt_flags` now the flags are specified
in `env` and applied accross the board.
This allows to
1. simplify build definitions,
2. avoid the need to generate dune files,
3. use plain sexps instead of OCaml and JBuilder plugin in build
files.
(I'll try to address 2 and 3 in the followup patches).
Existing `make` targets should continue working as before. Also, we
can use dune CLI like so:
```
infer/src$ dune printenv --profile opt # <- very useful for introspection
infer/src$ dune build check
infer/src$ dune build check --profile test
infer/src$ dune build infer.exe --profile dev
infer/src$ dune build infer.exe --profile opt
```
Also, with just 1 context something like `dune runtest` will run unit
tests only once instead of N times, where N is the number of contexts.
Now, there's one difference compared to the previous setup with
contexts:
- Previously, each context had its own build folder, and building infer
in opt context didn't invalidate any of the build artifacts in default
context. Therefore, alternating between `make` and `make opt` had low
overhead at the expense of having N copies of all build artifacts (1
for every context).
- Now, there's just 1 build folder and switching between profiles does
invalidate some artifacts (but not all) and rebuild takes a bit more
time.
So, if you're alternating like crazy between profiles your experience
may get worse (but not necessarily, more on that below). If you want
to trigger an opt build occasionally, you shouldn't notice much
difference.
For those who are concerned about slower build times when alternating
between different build profiles, there's a solution: [dune
cache](https://dune.readthedocs.io/en/stable/caching.html).
You can enable it by creating a file `~/.config/dune/config` with the
following contents:
```
(lang dune 2.0)
(cache enabled)
```
With cache enabled switching between different build profiles (but
also branches and commits) has ~0 overhead.
Dune cache works fine on Linux, but unfortunately there are [certain
problems with
MacOS](https://github.com/ocaml/dune/issues/3233) (hopefully, those
will be fixed soon and then there will be no downsides to using
profiles compared to contexts for our case).
Reviewed By: jvillard
Differential Revision: D20247864
fbshipit-source-id: 5f8afa0db
5 years ago
|
|
|
$(QUIET)$(call silent_on_success,Building infer annotations,\
|
|
|
|
|
$(MAKE) -C $(ANNOTATIONS_DIR))
|
|
|
|
|
endif
|
[RFC][build] Use dune environments and profiles instead of contexts
Summary:
With profiles and `(env ...)` stanza it's possible to consolidate
various ocamlc/ocamlopt/etc setups in a single place.
Where previously we needed to append `dune.common` to every dune file
and specify `flags` and `ocamlopt_flags` now the flags are specified
in `env` and applied accross the board.
This allows to
1. simplify build definitions,
2. avoid the need to generate dune files,
3. use plain sexps instead of OCaml and JBuilder plugin in build
files.
(I'll try to address 2 and 3 in the followup patches).
Existing `make` targets should continue working as before. Also, we
can use dune CLI like so:
```
infer/src$ dune printenv --profile opt # <- very useful for introspection
infer/src$ dune build check
infer/src$ dune build check --profile test
infer/src$ dune build infer.exe --profile dev
infer/src$ dune build infer.exe --profile opt
```
Also, with just 1 context something like `dune runtest` will run unit
tests only once instead of N times, where N is the number of contexts.
Now, there's one difference compared to the previous setup with
contexts:
- Previously, each context had its own build folder, and building infer
in opt context didn't invalidate any of the build artifacts in default
context. Therefore, alternating between `make` and `make opt` had low
overhead at the expense of having N copies of all build artifacts (1
for every context).
- Now, there's just 1 build folder and switching between profiles does
invalidate some artifacts (but not all) and rebuild takes a bit more
time.
So, if you're alternating like crazy between profiles your experience
may get worse (but not necessarily, more on that below). If you want
to trigger an opt build occasionally, you shouldn't notice much
difference.
For those who are concerned about slower build times when alternating
between different build profiles, there's a solution: [dune
cache](https://dune.readthedocs.io/en/stable/caching.html).
You can enable it by creating a file `~/.config/dune/config` with the
following contents:
```
(lang dune 2.0)
(cache enabled)
```
With cache enabled switching between different build profiles (but
also branches and commits) has ~0 overhead.
Dune cache works fine on Linux, but unfortunately there are [certain
problems with
MacOS](https://github.com/ocaml/dune/issues/3233) (hopefully, those
will be fixed soon and then there will be no downsides to using
profiles compared to contexts for our case).
Reviewed By: jvillard
Differential Revision: D20247864
fbshipit-source-id: 5f8afa0db
5 years ago
|
|
|
$(QUIET)$(call silent_on_success,Building infer models,\
|
|
|
|
|
$(MAKE) -C $(MODELS_DIR) all)
|
|
|
|
|
|
|
|
|
|
.PHONY: infer byte_infer
|
[RFC][build] Use dune environments and profiles instead of contexts
Summary:
With profiles and `(env ...)` stanza it's possible to consolidate
various ocamlc/ocamlopt/etc setups in a single place.
Where previously we needed to append `dune.common` to every dune file
and specify `flags` and `ocamlopt_flags` now the flags are specified
in `env` and applied accross the board.
This allows to
1. simplify build definitions,
2. avoid the need to generate dune files,
3. use plain sexps instead of OCaml and JBuilder plugin in build
files.
(I'll try to address 2 and 3 in the followup patches).
Existing `make` targets should continue working as before. Also, we
can use dune CLI like so:
```
infer/src$ dune printenv --profile opt # <- very useful for introspection
infer/src$ dune build check
infer/src$ dune build check --profile test
infer/src$ dune build infer.exe --profile dev
infer/src$ dune build infer.exe --profile opt
```
Also, with just 1 context something like `dune runtest` will run unit
tests only once instead of N times, where N is the number of contexts.
Now, there's one difference compared to the previous setup with
contexts:
- Previously, each context had its own build folder, and building infer
in opt context didn't invalidate any of the build artifacts in default
context. Therefore, alternating between `make` and `make opt` had low
overhead at the expense of having N copies of all build artifacts (1
for every context).
- Now, there's just 1 build folder and switching between profiles does
invalidate some artifacts (but not all) and rebuild takes a bit more
time.
So, if you're alternating like crazy between profiles your experience
may get worse (but not necessarily, more on that below). If you want
to trigger an opt build occasionally, you shouldn't notice much
difference.
For those who are concerned about slower build times when alternating
between different build profiles, there's a solution: [dune
cache](https://dune.readthedocs.io/en/stable/caching.html).
You can enable it by creating a file `~/.config/dune/config` with the
following contents:
```
(lang dune 2.0)
(cache enabled)
```
With cache enabled switching between different build profiles (but
also branches and commits) has ~0 overhead.
Dune cache works fine on Linux, but unfortunately there are [certain
problems with
MacOS](https://github.com/ocaml/dune/issues/3233) (hopefully, those
will be fixed soon and then there will be no downsides to using
profiles compared to contexts for our case).
Reviewed By: jvillard
Differential Revision: D20247864
fbshipit-source-id: 5f8afa0db
5 years ago
|
|
|
infer byte_infer: infer_models
|
|
|
|
|
infer: src_build
|
|
|
|
|
byte_infer: byte
|
|
|
|
|
|
|
|
|
|
.PHONY: opt
|
|
|
|
|
opt:
|
|
|
|
|
$(QUIET)$(MAKE) BUILD_MODE=opt infer
|
|
|
|
|
|
|
|
|
|
.PHONY: clang_setup
|
|
|
|
|
clang_setup:
|
|
|
|
|
# if clang is already built then let the user know they might not need to rebuild clang
|
|
|
|
|
$(QUIET)export CC="$(CC)" CFLAGS="$(CFLAGS)"; \
|
|
|
|
|
export CXX="$(CXX)" CXXFLAGS="$(CXXFLAGS)"; \
|
|
|
|
|
export CPP="$(CPP)" LDFLAGS="$(LDFLAGS)" LIBS="$(LIBS)"; \
|
|
|
|
|
$(FCP_DIR)/clang/setup.sh --only-check-install || { \
|
|
|
|
|
if [ -x '$(FCP_DIR)'/clang/install/bin/clang ]; then \
|
|
|
|
|
echo '$(TERM_INFO)*** Now building clang, this will take a while...$(TERM_RESET)' >&2; \
|
|
|
|
|
echo '$(TERM_INFO)*** If you believe that facebook-clang-plugins/clang/install is up-to-date you can$(TERM_RESET)' >&2; \
|
|
|
|
|
echo '$(TERM_INFO)*** interrupt the compilation (Control-C) and run this to prevent clang from being rebuilt:$(TERM_RESET)' >&2; \
|
|
|
|
|
echo >&2 ; \
|
|
|
|
|
echo '$(TERM_INFO) $(FCP_DIR)/clang/setup.sh --only-record-install$(TERM_RESET)' >&2; \
|
|
|
|
|
echo >&2 ; \
|
|
|
|
|
echo '$(TERM_INFO)(TIP: you can also force a clang rebuild by removing $(FCP_DIR)/clang/installed.version)$(TERM_RESET)' >&2; \
|
|
|
|
|
echo >&2 ; \
|
|
|
|
|
fi; \
|
|
|
|
|
$(FCP_DIR)/clang/setup.sh; \
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
.PHONY: clang_plugin
|
|
|
|
|
clang_plugin: clang_setup
|
|
|
|
|
$(QUIET)$(call silent_on_success,Building clang plugin,\
|
|
|
|
|
$(MAKE) -C $(FCP_DIR)/libtooling all \
|
|
|
|
|
CC="$(CC)" CXX="$(CXX)" \
|
|
|
|
|
CFLAGS="$(CFLAGS)" CXXFLAGS="$(CXXFLAGS)" \
|
|
|
|
|
CPP="$(CPP)" LDFLAGS="$(LDFLAGS)" LIBS="$(LIBS)" \
|
|
|
|
|
LOCAL_CLANG=$(CLANG_PREFIX)/bin/clang \
|
|
|
|
|
CLANG_PREFIX=$(CLANG_PREFIX) \
|
|
|
|
|
CLANG_INCLUDES=$(CLANG_INCLUDES) \
|
|
|
|
|
SDKPATH=$(XCODE_ISYSROOT) \
|
|
|
|
|
)
|
|
|
|
|
$(QUIET)$(call silent_on_success,Building clang plugin OCaml interface,\
|
|
|
|
|
$(MAKE) -C $(FCP_DIR)/clang-ocaml all \
|
|
|
|
|
build/clang_ast_proj.ml build/clang_ast_proj.mli \
|
|
|
|
|
CC=$(CC) CXX=$(CXX) \
|
|
|
|
|
CFLAGS="$(CFLAGS)" CXXFLAGS="$(CXXFLAGS)" \
|
|
|
|
|
CPP="$(CPP)" LDFLAGS="$(LDFLAGS)" LIBS="$(LIBS)" \
|
|
|
|
|
LOCAL_CLANG=$(CLANG_PREFIX)/bin/clang \
|
|
|
|
|
CLANG_PREFIX=$(CLANG_PREFIX) \
|
|
|
|
|
CLANG_INCLUDES=$(CLANG_INCLUDES) \
|
|
|
|
|
SDKPATH=$(XCODE_ISYSROOT) \
|
|
|
|
|
)
|
|
|
|
|
.PHONY: clang_plugin_test
|
|
|
|
|
clang_plugin_test: clang_setup
|
|
|
|
|
$(QUIET)$(call silent_on_success,Running facebook-clang-plugins/libtooling/ tests,\
|
|
|
|
|
$(MAKE) -C $(FCP_DIR)/libtooling test \
|
|
|
|
|
CC=$(CC) CXX=$(CXX) \
|
|
|
|
|
CFLAGS="$(CFLAGS)" CXXFLAGS="$(CXXFLAGS)" \
|
|
|
|
|
CPP="$(CPP)" LDFLAGS="$(LDFLAGS)" LIBS="$(LIBS)" \
|
|
|
|
|
LOCAL_CLANG=$(CLANG_PREFIX)/bin/clang \
|
|
|
|
|
CLANG_PREFIX=$(CLANG_PREFIX) \
|
|
|
|
|
CLANG_INCLUDES=$(CLANG_INCLUDES) \
|
|
|
|
|
SDKPATH=$(XCODE_ISYSROOT) \
|
|
|
|
|
)
|
|
|
|
|
$(QUIET)$(call silent_on_success,Running facebook-clang-plugins/clang-ocaml/ tests,\
|
|
|
|
|
$(MAKE) -C $(FCP_DIR)/clang-ocaml test \
|
|
|
|
|
CC=$(CC) CXX=$(CXX) \
|
|
|
|
|
CFLAGS="$(CFLAGS)" CXXFLAGS="$(CXXFLAGS)" \
|
|
|
|
|
CPP="$(CPP)" LDFLAGS="$(LDFLAGS)" LIBS="$(LIBS)" \
|
|
|
|
|
LOCAL_CLANG=$(CLANG_PREFIX)/bin/clang \
|
|
|
|
|
CLANG_PREFIX=$(CLANG_PREFIX) \
|
|
|
|
|
CLANG_INCLUDES=$(CLANG_INCLUDES) \
|
|
|
|
|
SDKPATH=$(XCODE_ISYSROOT) \
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
.PHONY: clang_plugin_test
|
|
|
|
|
clang_plugin_test_replace: clang_setup
|
|
|
|
|
$(QUIET)$(call silent_on_success,Running facebook-clang-plugins/libtooling/ record tests,\
|
|
|
|
|
$(MAKE) -C $(FCP_DIR)/libtooling record-test-outputs \
|
|
|
|
|
CC=$(CC) CXX=$(CXX) \
|
|
|
|
|
CFLAGS="$(CFLAGS)" CXXFLAGS="$(CXXFLAGS)" \
|
|
|
|
|
CPP="$(CPP)" LDFLAGS="$(LDFLAGS)" LIBS="$(LIBS)" \
|
|
|
|
|
LOCAL_CLANG=$(CLANG_PREFIX)/bin/clang \
|
|
|
|
|
CLANG_PREFIX=$(CLANG_PREFIX) \
|
|
|
|
|
CLANG_INCLUDES=$(CLANG_INCLUDES) \
|
|
|
|
|
SDKPATH=$(XCODE_ISYSROOT) \
|
|
|
|
|
)
|
|
|
|
|
$(QUIET)$(call silent_on_success,Running facebook-clang-plugins/clang-ocaml/ record tests,\
|
|
|
|
|
$(MAKE) -C $(FCP_DIR)/clang-ocaml record-test-outputs \
|
|
|
|
|
CC=$(CC) CXX=$(CXX) \
|
|
|
|
|
CFLAGS="$(CFLAGS)" CXXFLAGS="$(CXXFLAGS)" \
|
|
|
|
|
CPP="$(CPP)" LDFLAGS="$(LDFLAGS)" LIBS="$(LIBS)" \
|
|
|
|
|
LOCAL_CLANG=$(CLANG_PREFIX)/bin/clang \
|
|
|
|
|
CLANG_PREFIX=$(CLANG_PREFIX) \
|
|
|
|
|
CLANG_INCLUDES=$(CLANG_INCLUDES) \
|
|
|
|
|
SDKPATH=$(XCODE_ISYSROOT) \
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
.PHONY: ocaml_unit_test
|
|
|
|
|
ocaml_unit_test: test_build
|
|
|
|
|
$(QUIET)$(REMOVE_DIR) infer-out-unit-tests
|
|
|
|
|
$(QUIET)$(call silent_on_success,Running OCaml unit tests,\
|
|
|
|
|
INFER_ARGS=--results-dir^infer-out-unit-tests $(INFERUNIT_BIN))
|
|
|
|
|
|
|
|
|
|
define silence_make
|
|
|
|
|
$(1) 2> >(grep -v 'warning: \(ignoring old\|overriding\) \(commands\|recipe\) for target')
|
|
|
|
|
endef
|
|
|
|
|
|
|
|
|
|
.PHONY: $(DIRECT_TESTS:%=direct_%_test)
|
|
|
|
|
$(DIRECT_TESTS:%=direct_%_test): infer
|
|
|
|
|
$(QUIET)$(call silent_on_success,Running test: $(subst _, ,$@),\
|
|
|
|
|
$(call silence_make,\
|
|
|
|
|
$(MAKE) -C \
|
|
|
|
|
$(INFER_DIR)/tests/codetoanalyze/$(shell printf $@ | cut -f 2 -d _)/$(shell printf $@ | cut -f 3 -d _) \
|
|
|
|
|
test))
|
|
|
|
|
|
|
|
|
|
.PHONY: $(DIRECT_TESTS:%=direct_%_print)
|
|
|
|
|
$(DIRECT_TESTS:%=direct_%_print): infer
|
|
|
|
|
$(QUIET)$(call silent_on_success,Running: $(subst _, ,$@),\
|
|
|
|
|
$(call silence_make,\
|
|
|
|
|
$(MAKE) -C \
|
|
|
|
|
$(INFER_DIR)/tests/codetoanalyze/$(shell printf $@ | cut -f 2 -d _)/$(shell printf $@ | cut -f 3 -d _) \
|
|
|
|
|
print))
|
|
|
|
|
|
|
|
|
|
.PHONY: $(DIRECT_TESTS:%=direct_%_clean)
|
|
|
|
|
$(DIRECT_TESTS:%=direct_%_clean):
|
|
|
|
|
$(QUIET)$(call silent_on_success,Cleaning: $(subst _, ,$@),\
|
|
|
|
|
$(call silence_make,\
|
|
|
|
|
$(MAKE) -C \
|
|
|
|
|
$(INFER_DIR)/tests/codetoanalyze/$(shell printf $@ | cut -f 2 -d _)/$(shell printf $@ | cut -f 3 -d _) \
|
|
|
|
|
clean))
|
|
|
|
|
|
|
|
|
|
.PHONY: $(DIRECT_TESTS:%=direct_%_replace)
|
|
|
|
|
$(DIRECT_TESTS:%=direct_%_replace): infer
|
|
|
|
|
$(QUIET)$(call silent_on_success,Recording: $(subst _, ,$@),\
|
|
|
|
|
$(call silence_make,\
|
|
|
|
|
$(MAKE) -C \
|
|
|
|
|
$(INFER_DIR)/tests/codetoanalyze/$(shell printf $@ | cut -f 2 -d _)/$(shell printf $@ | cut -f 3 -d _) \
|
|
|
|
|
replace))
|
|
|
|
|
|
|
|
|
|
.PHONY: direct_tests
|
|
|
|
|
direct_tests: $(DIRECT_TESTS:%=direct_%_test)
|
|
|
|
|
|
|
|
|
|
COST_TESTS += \
|
|
|
|
|
c_performance \
|
|
|
|
|
java_hoistingExpensive \
|
|
|
|
|
java_performance \
|
|
|
|
|
objc_performance \
|
|
|
|
|
|
|
|
|
|
.PHONY: cost_tests
|
|
|
|
|
cost_tests: $(COST_TESTS:%=direct_%_test)
|
|
|
|
|
|
|
|
|
|
.PHONY: cost_tests_clean
|
|
|
|
|
cost_tests_clean: $(COST_TESTS:%=direct_%_clean)
|
|
|
|
|
|
|
|
|
|
.PHONY: cost_tests_replace
|
|
|
|
|
cost_tests_replace: $(COST_TESTS:%=direct_%_replace)
|
|
|
|
|
|
|
|
|
|
.PHONY: cost_tests_print
|
|
|
|
|
cost_tests_print: $(COST_TESTS:%=direct_%_print)
|
|
|
|
|
|
|
|
|
|
.PHONY: $(BUILD_SYSTEMS_TESTS:%=build_%_test)
|
|
|
|
|
$(BUILD_SYSTEMS_TESTS:%=build_%_test): infer
|
|
|
|
|
$(QUIET)$(call silent_on_success,Running test: $(subst _, ,$@),\
|
|
|
|
|
$(call silence_make,\
|
|
|
|
|
$(MAKE) -C $(INFER_DIR)/tests/build_systems/$(patsubst build_%_test,%,$@) test))
|
|
|
|
|
|
|
|
|
|
.PHONY: $(BUILD_SYSTEMS_TESTS:%=build_%_print)
|
|
|
|
|
$(BUILD_SYSTEMS_TESTS:%=build_%_print): infer
|
|
|
|
|
$(QUIET)$(call silent_on_success,Running: $(subst _, ,$@),\
|
|
|
|
|
$(call silence_make,\
|
|
|
|
|
$(MAKE) -C $(INFER_DIR)/tests/build_systems/$(patsubst build_%_print,%,$@) print))
|
|
|
|
|
|
|
|
|
|
.PHONY: $(BUILD_SYSTEMS_TESTS:%=build_%_clean)
|
|
|
|
|
$(BUILD_SYSTEMS_TESTS:%=build_%_clean):
|
|
|
|
|
$(QUIET)$(call silent_on_success,Cleaning: $(subst _, ,$@),\
|
|
|
|
|
$(call silence_make,\
|
|
|
|
|
$(MAKE) -C $(INFER_DIR)/tests/build_systems/$(patsubst build_%_clean,%,$@) clean))
|
|
|
|
|
|
|
|
|
|
.PHONY: $(BUILD_SYSTEMS_TESTS:%=build_%_replace)
|
|
|
|
|
$(BUILD_SYSTEMS_TESTS:%=build_%_replace): infer
|
|
|
|
|
$(QUIET)$(call silent_on_success,Recording: $(subst _, ,$@),\
|
|
|
|
|
$(call silence_make,\
|
|
|
|
|
$(MAKE) -C $(INFER_DIR)/tests/build_systems/$(patsubst build_%_replace,%,$@) replace))
|
|
|
|
|
|
|
|
|
|
build_infertop_print build_infertop_test build_infertop_replace: toplevel_test
|
|
|
|
|
|
|
|
|
|
.PHONY: build_systems_tests
|
|
|
|
|
build_systems_tests: $(BUILD_SYSTEMS_TESTS:%=build_%_test)
|
|
|
|
|
|
|
|
|
|
.PHONY: endtoend_test
|
|
|
|
|
endtoend_test: $(BUILD_SYSTEMS_TESTS:%=build_%_test) $(DIRECT_TESTS:%=direct_%_test)
|
|
|
|
|
|
|
|
|
|
.PHONY: check_missing_mli
|
|
|
|
|
check_missing_mli:
|
|
|
|
|
$(QUIET)for x in $$(find $(INFER_DIR)/src -name "*.ml"); do \
|
|
|
|
|
test -f "$$x"i || echo Missing "$$x"i; done
|
|
|
|
|
|
|
|
|
|
.PHONY: checkCopyright
|
|
|
|
|
checkCopyright: src_build_common
|
|
|
|
|
$(QUIET)$(call silent_on_success,Building checkCopyright,\
|
|
|
|
|
$(MAKE) -C $(SRC_DIR) checkCopyright)
|
|
|
|
|
|
|
|
|
|
.PHONY: validate-skel
|
|
|
|
|
validate-skel:
|
|
|
|
|
ifeq ($(IS_FACEBOOK_TREE),yes)
|
|
|
|
|
$(QUIET)$(call silent_on_success,Validating facebook/,\
|
|
|
|
|
$(MAKE) -C facebook validate)
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
.PHONY: crash_if_not_all_analyzers_enabled
|
|
|
|
|
crash_if_not_all_analyzers_enabled:
|
|
|
|
|
ifneq ($(BUILD_C_ANALYZERS)+$(BUILD_JAVA_ANALYZERS),yes+yes)
|
|
|
|
|
ifneq ($(BUILD_C_ANALYZERS),yes)
|
|
|
|
|
@echo '*** ERROR: Cannot run the full tests: the Clang analyzers are disabled.'
|
|
|
|
|
@echo '*** ERROR: You can run clang-only tests with:'
|
|
|
|
|
@echo '*** ERROR:'
|
|
|
|
|
@echo '*** ERROR: make config_tests'
|
|
|
|
|
@echo '*** ERROR:'
|
|
|
|
|
endif
|
|
|
|
|
ifneq ($(BUILD_JAVA_ANALYZERS),yes)
|
|
|
|
|
@echo '*** ERROR: Cannot run the full tests: the Java analyzers are disabled.'
|
|
|
|
|
@echo '*** ERROR: You can run Java-only tests with:'
|
|
|
|
|
@echo '*** ERROR:'
|
|
|
|
|
@echo '*** ERROR: make config_tests'
|
|
|
|
|
@echo '*** ERROR:'
|
|
|
|
|
endif
|
|
|
|
|
@echo '*** ERROR: To run the full set of tests, please enable all the analyzers.'
|
|
|
|
|
@exit 1
|
|
|
|
|
else
|
|
|
|
|
@:
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
.PHONY: mod_dep
|
|
|
|
|
mod_dep: src_build_common
|
|
|
|
|
$(QUIET)$(call silent_on_success,Building Infer source dependency graph,\
|
|
|
|
|
$(MAKE) -C $(SRC_DIR) mod_dep.dot)
|
|
|
|
|
|
|
|
|
|
# `test_build` and `src_build` (which is a dependency of `endtoend_test`) should not be run in
|
|
|
|
|
# parallel since they build infer with different profiles (and therefore conflict). Therefore,
|
|
|
|
|
# `test_build` is in the dependency, and `endtoend_test` in the recipe.
|
|
|
|
|
.PHONY: config_tests
|
[RFC][build] Use dune environments and profiles instead of contexts
Summary:
With profiles and `(env ...)` stanza it's possible to consolidate
various ocamlc/ocamlopt/etc setups in a single place.
Where previously we needed to append `dune.common` to every dune file
and specify `flags` and `ocamlopt_flags` now the flags are specified
in `env` and applied accross the board.
This allows to
1. simplify build definitions,
2. avoid the need to generate dune files,
3. use plain sexps instead of OCaml and JBuilder plugin in build
files.
(I'll try to address 2 and 3 in the followup patches).
Existing `make` targets should continue working as before. Also, we
can use dune CLI like so:
```
infer/src$ dune printenv --profile opt # <- very useful for introspection
infer/src$ dune build check
infer/src$ dune build check --profile test
infer/src$ dune build infer.exe --profile dev
infer/src$ dune build infer.exe --profile opt
```
Also, with just 1 context something like `dune runtest` will run unit
tests only once instead of N times, where N is the number of contexts.
Now, there's one difference compared to the previous setup with
contexts:
- Previously, each context had its own build folder, and building infer
in opt context didn't invalidate any of the build artifacts in default
context. Therefore, alternating between `make` and `make opt` had low
overhead at the expense of having N copies of all build artifacts (1
for every context).
- Now, there's just 1 build folder and switching between profiles does
invalidate some artifacts (but not all) and rebuild takes a bit more
time.
So, if you're alternating like crazy between profiles your experience
may get worse (but not necessarily, more on that below). If you want
to trigger an opt build occasionally, you shouldn't notice much
difference.
For those who are concerned about slower build times when alternating
between different build profiles, there's a solution: [dune
cache](https://dune.readthedocs.io/en/stable/caching.html).
You can enable it by creating a file `~/.config/dune/config` with the
following contents:
```
(lang dune 2.0)
(cache enabled)
```
With cache enabled switching between different build profiles (but
also branches and commits) has ~0 overhead.
Dune cache works fine on Linux, but unfortunately there are [certain
problems with
MacOS](https://github.com/ocaml/dune/issues/3233) (hopefully, those
will be fixed soon and then there will be no downsides to using
profiles compared to contexts for our case).
Reviewed By: jvillard
Differential Revision: D20247864
fbshipit-source-id: 5f8afa0db
5 years ago
|
|
|
config_tests: test_build ocaml_unit_test validate-skel mod_dep
|
|
|
|
|
$(MAKE) endtoend_test checkCopyright
|
|
|
|
|
$(MAKE) manuals
|
|
|
|
|
|
[RFC][build] Use dune environments and profiles instead of contexts
Summary:
With profiles and `(env ...)` stanza it's possible to consolidate
various ocamlc/ocamlopt/etc setups in a single place.
Where previously we needed to append `dune.common` to every dune file
and specify `flags` and `ocamlopt_flags` now the flags are specified
in `env` and applied accross the board.
This allows to
1. simplify build definitions,
2. avoid the need to generate dune files,
3. use plain sexps instead of OCaml and JBuilder plugin in build
files.
(I'll try to address 2 and 3 in the followup patches).
Existing `make` targets should continue working as before. Also, we
can use dune CLI like so:
```
infer/src$ dune printenv --profile opt # <- very useful for introspection
infer/src$ dune build check
infer/src$ dune build check --profile test
infer/src$ dune build infer.exe --profile dev
infer/src$ dune build infer.exe --profile opt
```
Also, with just 1 context something like `dune runtest` will run unit
tests only once instead of N times, where N is the number of contexts.
Now, there's one difference compared to the previous setup with
contexts:
- Previously, each context had its own build folder, and building infer
in opt context didn't invalidate any of the build artifacts in default
context. Therefore, alternating between `make` and `make opt` had low
overhead at the expense of having N copies of all build artifacts (1
for every context).
- Now, there's just 1 build folder and switching between profiles does
invalidate some artifacts (but not all) and rebuild takes a bit more
time.
So, if you're alternating like crazy between profiles your experience
may get worse (but not necessarily, more on that below). If you want
to trigger an opt build occasionally, you shouldn't notice much
difference.
For those who are concerned about slower build times when alternating
between different build profiles, there's a solution: [dune
cache](https://dune.readthedocs.io/en/stable/caching.html).
You can enable it by creating a file `~/.config/dune/config` with the
following contents:
```
(lang dune 2.0)
(cache enabled)
```
With cache enabled switching between different build profiles (but
also branches and commits) has ~0 overhead.
Dune cache works fine on Linux, but unfortunately there are [certain
problems with
MacOS](https://github.com/ocaml/dune/issues/3233) (hopefully, those
will be fixed soon and then there will be no downsides to using
profiles compared to contexts for our case).
Reviewed By: jvillard
Differential Revision: D20247864
fbshipit-source-id: 5f8afa0db
5 years ago
|
|
|
ifneq ($(filter endtoend_test,${MAKECMDGOALS}),)
|
|
|
|
|
checkCopyright: src_build
|
|
|
|
|
toplevel_test: checkCopyright
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
.PHONY: test
|
|
|
|
|
test: crash_if_not_all_analyzers_enabled config_tests
|
|
|
|
|
ifeq (,$(findstring s,$(MAKEFLAGS)))
|
|
|
|
|
$(QUIET)echo "$(TERM_INFO)ALL TESTS PASSED$(TERM_RESET)"
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
.PHONY: test-replace
|
|
|
|
|
test-replace: $(BUILD_SYSTEMS_TESTS:%=build_%_replace) $(DIRECT_TESTS:%=direct_%_replace) \
|
|
|
|
|
clang_plugin_test_replace
|
|
|
|
|
|
|
|
|
|
.PHONY: uninstall
|
|
|
|
|
uninstall:
|
|
|
|
|
$(REMOVE_DIR) $(DESTDIR)$(libdir)/infer/
|
|
|
|
|
$(REMOVE) $(DESTDIR)$(bindir)/infer
|
|
|
|
|
$(REMOVE) $(INFER_COMMANDS:%=$(DESTDIR)$(bindir)/%)
|
|
|
|
|
$(REMOVE) $(foreach manual,$(INFER_GROFF_MANUALS_GZIPPED),\
|
|
|
|
|
$(DESTDIR)$(mandir)/man1/$(notdir $(manual)))
|
|
|
|
|
ifeq ($(IS_FACEBOOK_TREE),yes)
|
|
|
|
|
$(MAKE) -C facebook uninstall
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
.PHONY: test_clean
|
|
|
|
|
test_clean: $(DIRECT_TESTS:%=direct_%_clean) $(BUILD_SYSTEMS_TESTS:%=build_%_clean)
|
|
|
|
|
|
|
|
|
|
.PHONY: install
|
|
|
|
|
install: infer $(INFER_GROFF_MANUALS_GZIPPED)
|
|
|
|
|
# create directory structure
|
|
|
|
|
test -d '$(DESTDIR)$(bindir)' || \
|
|
|
|
|
$(MKDIR_P) '$(DESTDIR)$(bindir)'
|
|
|
|
|
test -d '$(DESTDIR)$(mandir)/man1' || \
|
|
|
|
|
$(MKDIR_P) '$(DESTDIR)$(mandir)/man1'
|
|
|
|
|
test -d '$(DESTDIR)$(libdir)/infer/' || \
|
|
|
|
|
$(MKDIR_P) '$(DESTDIR)$(libdir)/infer/'
|
|
|
|
|
ifeq ($(BUILD_C_ANALYZERS),yes)
|
|
|
|
|
test -d '$(DESTDIR)$(libdir)/infer/facebook-clang-plugins/libtooling/build/' || \
|
|
|
|
|
$(MKDIR_P) '$(DESTDIR)$(libdir)/infer/facebook-clang-plugins/libtooling/build/'
|
|
|
|
|
find facebook-clang-plugins/clang/install/. -type d -print0 | xargs -0 -n 1 \
|
|
|
|
|
$(SHELL) -x -c "test -d '$(DESTDIR)$(libdir)'/infer/\$$1 || \
|
|
|
|
|
$(MKDIR_P) '$(DESTDIR)$(libdir)'/infer/\$$1" --
|
|
|
|
|
test -d '$(DESTDIR)$(libdir)/infer/infer/lib/clang_wrappers/' || \
|
|
|
|
|
$(MKDIR_P) '$(DESTDIR)$(libdir)/infer/infer/lib/clang_wrappers/'
|
|
|
|
|
test -d '$(DESTDIR)$(libdir)/infer/infer/lib/linter_rules/' || \
|
|
|
|
|
$(MKDIR_P) '$(DESTDIR)$(libdir)/infer/infer/lib/linter_rules/'
|
|
|
|
|
test -d '$(DESTDIR)$(libdir)/infer/infer/etc/' || \
|
|
|
|
|
$(MKDIR_P) '$(DESTDIR)$(libdir)/infer/infer/etc'
|
|
|
|
|
endif
|
|
|
|
|
ifeq ($(BUILD_JAVA_ANALYZERS),yes)
|
|
|
|
|
test -d '$(DESTDIR)$(libdir)/infer/infer/lib/java/' || \
|
|
|
|
|
$(MKDIR_P) '$(DESTDIR)$(libdir)/infer/infer/lib/java/'
|
|
|
|
|
endif
|
|
|
|
|
test -d '$(DESTDIR)$(libdir)/infer/infer/annotations/' || \
|
|
|
|
|
$(MKDIR_P) '$(DESTDIR)$(libdir)/infer/infer/annotations/'
|
|
|
|
|
test -d '$(DESTDIR)$(libdir)/infer/infer/lib/wrappers/' || \
|
|
|
|
|
$(MKDIR_P) '$(DESTDIR)$(libdir)/infer/infer/lib/wrappers/'
|
|
|
|
|
test -d '$(DESTDIR)$(libdir)/infer/infer/lib/specs/' || \
|
|
|
|
|
$(MKDIR_P) '$(DESTDIR)$(libdir)/infer/infer/lib/specs/'
|
|
|
|
|
test -d '$(DESTDIR)$(libdir)/infer/infer/bin/' || \
|
|
|
|
|
$(MKDIR_P) '$(DESTDIR)$(libdir)/infer/infer/bin/'
|
|
|
|
|
# copy files
|
|
|
|
|
ifeq ($(BUILD_C_ANALYZERS),yes)
|
|
|
|
|
$(INSTALL_DATA) -C 'facebook-clang-plugins/libtooling/build/FacebookClangPlugin.dylib' \
|
|
|
|
|
'$(DESTDIR)$(libdir)/infer/facebook-clang-plugins/libtooling/build/FacebookClangPlugin.dylib'
|
|
|
|
|
# do not use "install" for symbolic links as this will copy the destination file instead
|
|
|
|
|
find facebook-clang-plugins/clang/install/. -not -type d -not -type l -not -name '*.a' -print0 \
|
|
|
|
|
| xargs -0 -I \{\} $(INSTALL_PROGRAM) -C \{\} '$(DESTDIR)$(libdir)'/infer/\{\}
|
|
|
|
|
# all the symlinks in clang are relative and safe to brutally copy over
|
|
|
|
|
find facebook-clang-plugins/clang/install/. -type l -not -name '*.a' -print0 \
|
|
|
|
|
| xargs -0 -I \{\} $(COPY) -a \{\} '$(DESTDIR)$(libdir)'/infer/\{\}
|
|
|
|
|
find infer/lib/clang_wrappers/* -print0 | xargs -0 -I \{\} \
|
|
|
|
|
$(INSTALL_PROGRAM) -C \{\} '$(DESTDIR)$(libdir)'/infer/\{\}
|
|
|
|
|
# only for files that point to infer
|
|
|
|
|
(cd '$(DESTDIR)$(libdir)/infer/infer/lib/wrappers/' && \
|
|
|
|
|
$(foreach cc,$(shell find '$(LIB_DIR)/wrappers' -type l), \
|
|
|
|
|
[ $(cc) -ef '$(INFER_BIN)' ] && \
|
|
|
|
|
$(REMOVE) '$(notdir $(cc))' && \
|
|
|
|
|
$(LN_S) ../../bin/infer '$(notdir $(cc))';))
|
|
|
|
|
find infer/lib/specs/* -print0 | xargs -0 -I \{\} \
|
|
|
|
|
$(INSTALL_DATA) -C \{\} '$(DESTDIR)$(libdir)'/infer/\{\}
|
|
|
|
|
$(INSTALL_DATA) -C 'infer/lib/linter_rules/linters.al' \
|
|
|
|
|
'$(DESTDIR)$(libdir)/infer/infer/lib/linter_rules/linters.al'
|
|
|
|
|
$(INSTALL_DATA) -C 'infer/etc/clang_ast.dict' \
|
|
|
|
|
'$(DESTDIR)$(libdir)/infer/infer/etc/clang_ast.dict'
|
|
|
|
|
endif
|
|
|
|
|
ifeq ($(BUILD_JAVA_ANALYZERS),yes)
|
|
|
|
|
$(INSTALL_DATA) -C 'infer/annotations/annotations.jar' \
|
|
|
|
|
'$(DESTDIR)$(libdir)/infer/infer/annotations/annotations.jar'
|
|
|
|
|
find infer/lib/java/*.jar -print0 | xargs -0 -I \{\} \
|
|
|
|
|
$(INSTALL_DATA) -C \{\} '$(DESTDIR)$(libdir)'/infer/\{\}
|
|
|
|
|
$(INSTALL_PROGRAM) -C '$(LIB_DIR)'/wrappers/javac \
|
|
|
|
|
'$(DESTDIR)$(libdir)'/infer/infer/lib/wrappers/
|
|
|
|
|
endif
|
|
|
|
|
$(INSTALL_PROGRAM) -C '$(INFER_BIN)' '$(DESTDIR)$(libdir)'/infer/infer/bin/
|
|
|
|
|
(cd '$(DESTDIR)$(bindir)/' && \
|
|
|
|
|
$(REMOVE) infer && \
|
|
|
|
|
$(LN_S) '$(libdir_relative_to_bindir)'/infer/infer/bin/infer infer)
|
[CLI] switch to infer-<command> (symlinks) executables
Summary:
Introduce `infer-<command>` for each command, except for internal commands
(only `infer-clang` for now) which are not exported. Install these executables
(which are just symlinks to `infer`) on `make install`. The main executable
looks at the name it was invoked with to figure out if it should behave as a
particular command.
Get rid of `InferClang`, `InferAnalyze`, and `InferPrint`. As a bonus, we now
only need to build one executable: `infer`, which should be a few seconds
faster (less link time).
`InferAnalyze` is now `infer-analyze` and `InferPrint` is `infer-print`. To run
`InferClang`, use a symlink named `clang`, `clang++`, etc. to `infer`. There
are such symlinks available in "infer/lib/wrappers/" already.
I also noticed that the scripts in xcodebuild_wrappers/ don't seem useful
anymore, so use wrappers/ instead, as for `make`.
Reviewed By: mbouaziz
Differential Revision: D5036495
fbshipit-source-id: 4a90030
8 years ago
|
|
|
for alias in $(INFER_COMMANDS); do \
|
|
|
|
|
(cd '$(DESTDIR)$(bindir)'/ && \
|
|
|
|
|
$(REMOVE) "$$alias" && \
|
|
|
|
|
$(LN_S) infer "$$alias"); done
|
[CLI] switch to infer-<command> (symlinks) executables
Summary:
Introduce `infer-<command>` for each command, except for internal commands
(only `infer-clang` for now) which are not exported. Install these executables
(which are just symlinks to `infer`) on `make install`. The main executable
looks at the name it was invoked with to figure out if it should behave as a
particular command.
Get rid of `InferClang`, `InferAnalyze`, and `InferPrint`. As a bonus, we now
only need to build one executable: `infer`, which should be a few seconds
faster (less link time).
`InferAnalyze` is now `infer-analyze` and `InferPrint` is `infer-print`. To run
`InferClang`, use a symlink named `clang`, `clang++`, etc. to `infer`. There
are such symlinks available in "infer/lib/wrappers/" already.
I also noticed that the scripts in xcodebuild_wrappers/ don't seem useful
anymore, so use wrappers/ instead, as for `make`.
Reviewed By: mbouaziz
Differential Revision: D5036495
fbshipit-source-id: 4a90030
8 years ago
|
|
|
for alias in $(INFER_COMMANDS); do \
|
|
|
|
|
(cd '$(DESTDIR)$(libdir)'/infer/infer/bin && \
|
|
|
|
|
$(REMOVE) "$$alias" && \
|
|
|
|
|
$(LN_S) infer "$$alias"); done
|
|
|
|
|
$(foreach man,$(INFER_GROFF_MANUALS_GZIPPED), \
|
|
|
|
|
$(INSTALL_DATA) -C $(man) '$(DESTDIR)$(mandir)/man1/$(notdir $(man))';)
|
|
|
|
|
ifeq ($(IS_FACEBOOK_TREE),yes)
|
|
|
|
|
ifdef DESTDIR
|
|
|
|
|
ifeq (,$(findstring :/,:$(DESTDIR)))
|
|
|
|
|
# DESTDIR is set and relative
|
|
|
|
|
$(MAKE) -C facebook install 'DESTDIR=../$(DESTDIR)'
|
|
|
|
|
else
|
|
|
|
|
# DESTDIR is set and absolute
|
|
|
|
|
$(MAKE) -C facebook install
|
|
|
|
|
endif
|
|
|
|
|
else
|
|
|
|
|
# DESTDIR not set
|
|
|
|
|
$(MAKE) -C facebook install
|
|
|
|
|
endif
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
# Nuke objects built from OCaml. Useful when changing the OCaml compiler, for instance.
|
|
|
|
|
.PHONY: ocaml_clean
|
|
|
|
|
ocaml_clean:
|
|
|
|
|
ifeq ($(BUILD_C_ANALYZERS),yes)
|
|
|
|
|
$(QUIET)$(call silent_on_success,Cleaning facebook-clang-plugins OCaml build,\
|
|
|
|
|
$(MAKE) -C $(FCP_DIR)/clang-ocaml clean)
|
|
|
|
|
endif
|
|
|
|
|
$(QUIET)$(call silent_on_success,Cleaning infer OCaml build,\
|
|
|
|
|
$(MAKE) -C $(SRC_DIR) clean)
|
|
|
|
|
$(QUIET)$(call silent_on_success,Cleaning ocamldot,\
|
|
|
|
|
$(MAKE) -C $(DEPENDENCIES_DIR)/ocamldot clean)
|
|
|
|
|
|
|
|
|
|
.PHONY: clean
|
|
|
|
|
clean: ocaml_clean test_clean
|
|
|
|
|
ifeq ($(BUILD_C_ANALYZERS),yes)
|
|
|
|
|
$(QUIET)$(call silent_on_success,Cleaning facebook-clang-plugins C++ build,\
|
|
|
|
|
$(MAKE) -C $(FCP_DIR) clean)
|
|
|
|
|
endif
|
|
|
|
|
$(QUIET)$(call silent_on_success,Cleaning Java annotations,\
|
|
|
|
|
$(MAKE) -C $(ANNOTATIONS_DIR) clean)
|
|
|
|
|
$(QUIET)$(call silent_on_success,Cleaning infer models,\
|
|
|
|
|
$(MAKE) -C $(MODELS_DIR) clean)
|
|
|
|
|
ifeq ($(IS_FACEBOOK_TREE),yes)
|
|
|
|
|
$(QUIET)$(call silent_on_success,Cleaning facebook/,\
|
|
|
|
|
$(MAKE) -C facebook clean)
|
|
|
|
|
endif
|
|
|
|
|
$(QUIET)$(call silent_on_success,Removing *.o and *.o.sh,\
|
|
|
|
|
find $(INFER_DIR)/tests \( -name '*.o' -o -name '*.o.sh' \) -delete)
|
|
|
|
|
$(QUIET)$(call silent_on_success,Removing build logs,\
|
|
|
|
|
$(REMOVE_DIR) _build_logs)
|
|
|
|
|
|
|
|
|
|
.PHONY: conf-clean
|
|
|
|
|
conf-clean: clean
|
|
|
|
|
$(REMOVE) .buckversion
|
|
|
|
|
$(REMOVE) Makefile.autoconf
|
|
|
|
|
$(REMOVE) acinclude.m4
|
|
|
|
|
$(REMOVE) aclocal.m4
|
|
|
|
|
$(REMOVE_DIR) autom4te.cache/
|
|
|
|
|
$(REMOVE) config.log
|
|
|
|
|
$(REMOVE) config.status
|
|
|
|
|
$(REMOVE) configure
|
|
|
|
|
$(REMOVE_DIR) $(MODELS_DIR)/c/out/
|
|
|
|
|
$(REMOVE_DIR) $(MODELS_DIR)/cpp/out/
|
|
|
|
|
$(REMOVE_DIR) $(MODELS_DIR)/java/infer-out/
|
|
|
|
|
$(REMOVE_DIR) $(MODELS_DIR)/objc/out/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# phony because it depends on opam's internal state
|
|
|
|
|
.PHONY: opam.locked
|
|
|
|
|
opam.locked: opam
|
|
|
|
|
# allow users to not force a run of opam update since it's very slow
|
|
|
|
|
ifeq ($(NO_OPAM_UPDATE),)
|
|
|
|
|
$(QUIET)$(call silent_on_success,opam update,$(OPAM) update)
|
|
|
|
|
endif
|
|
|
|
|
$(QUIET)$(call silent_on_success,generating opam.locked,\
|
|
|
|
|
$(OPAM) lock .)
|
|
|
|
|
|
|
|
|
|
OPAM_DEV_DEPS = ocp-indent merlin utop webbrowser
|
|
|
|
|
|
|
|
|
|
ifneq ($(EMACS),no)
|
|
|
|
|
OPAM_DEV_DEPS += tuareg
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
.PHONY: devsetup
|
|
|
|
|
devsetup: Makefile.autoconf
|
|
|
|
|
$(QUIET)[ $(OPAM) != "no" ] || (echo 'No `opam` found, aborting setup.' >&2; exit 1)
|
|
|
|
|
$(QUIET)$(call silent_on_success,installing $(OPAM_DEV_DEPS),\
|
|
|
|
|
OPAMSWITCH=$(OPAMSWITCH); $(OPAM) install --yes --no-checksum user-setup $(OPAM_DEV_DEPS))
|
|
|
|
|
$(QUIET)echo '$(TERM_INFO)*** Running `opam user-setup`$(TERM_RESET)' >&2
|
|
|
|
|
$(QUIET)OPAMSWITCH=$(OPAMSWITCH); OPAMYES=1; $(OPAM) user-setup install
|
|
|
|
|
$(QUIET)if [ "$(PLATFORM)" = "Darwin" ] && [ x"$(GNU_SED)" = x"no" ]; then \
|
|
|
|
|
echo '$(TERM_INFO)*** Installing GNU sed$(TERM_RESET)' >&2; \
|
|
|
|
|
brew install gnu-sed; \
|
|
|
|
|
fi
|
|
|
|
|
$(QUIET)if [ "$(PLATFORM)" = "Darwin" ] && ! $$(parallel -h | grep -q GNU); then \
|
|
|
|
|
echo '$(TERM_INFO)*** Installing GNU parallel$(TERM_RESET)' >&2; \
|
|
|
|
|
brew install parallel; \
|
|
|
|
|
fi
|
|
|
|
|
$(QUIET)if [ ! -d "$$HOME"/.parallel ]; then mkdir "$$HOME"/.parallel; fi
|
|
|
|
|
$(QUIET)touch "$$HOME"/.parallel/will-cite
|
|
|
|
|
# expand all occurrences of "~" in PATH and MANPATH
|
|
|
|
|
$(QUIET)infer_repo_is_in_path=$$(echo $${PATH//\~/$$HOME} | grep -q "$(ABSOLUTE_ROOT_DIR)"/infer/bin; echo $$?); \
|
|
|
|
|
infer_repo_is_in_manpath=$$(echo $${MANPATH//\~/$$HOME} | grep -q "$(ABSOLUTE_ROOT_DIR)"/infer/man; echo $$?); \
|
|
|
|
|
shell_config_file="<could not auto-detect, please fill in yourself>"; \
|
|
|
|
|
if [ $$(basename "$(ORIG_SHELL)") = "bash" ]; then \
|
|
|
|
|
if [ "$(PLATFORM)" = "Linux" ]; then \
|
|
|
|
|
shell_config_file="$$HOME"/.bashrc; \
|
|
|
|
|
else \
|
|
|
|
|
shell_config_file="$$HOME"/.bash_profile; \
|
|
|
|
|
fi; \
|
|
|
|
|
elif [ $$(basename "$(ORIG_SHELL)") = "zsh" ]; then \
|
|
|
|
|
shell_config_file="$$HOME"/.zshrc; \
|
|
|
|
|
fi; \
|
|
|
|
|
if [ "$$infer_repo_is_in_path" != "0" ] || [ "$$infer_repo_is_in_manpath" != "0" ]; then \
|
|
|
|
|
echo >&2; \
|
|
|
|
|
echo '$(TERM_INFO)*** NOTE: `infer` is not in your PATH or MANPATH. If you are hacking on infer, you may$(TERM_RESET)' >&2; \
|
|
|
|
|
echo '$(TERM_INFO)*** NOTE: want to make infer executables and manuals available in your terminal. Type$(TERM_RESET)' >&2; \
|
|
|
|
|
echo '$(TERM_INFO)*** NOTE: the following commands to configure the current terminal and record the$(TERM_RESET)' >&2; \
|
|
|
|
|
printf '$(TERM_INFO)*** NOTE: changes in your shell configuration file (%s):$(TERM_RESET)\n' "$$shell_config_file">&2; \
|
|
|
|
|
echo >&2; \
|
|
|
|
|
if [ "$$infer_repo_is_in_path" != "0" ]; then \
|
|
|
|
|
printf '$(TERM_INFO) export PATH="%s/infer/bin":$$PATH$(TERM_RESET)\n' "$(ABSOLUTE_ROOT_DIR)" >&2; \
|
|
|
|
|
fi; \
|
|
|
|
|
if [ "$$infer_repo_is_in_manpath" != "0" ]; then \
|
|
|
|
|
printf '$(TERM_INFO) export MANPATH="%s/infer/man":$$MANPATH$(TERM_RESET)\n' "$(ABSOLUTE_ROOT_DIR)" >&2; \
|
|
|
|
|
fi; \
|
|
|
|
|
if [ "$$infer_repo_is_in_path" != "0" ]; then \
|
|
|
|
|
printf "$(TERM_INFO) echo 'export PATH=\"%s/infer/bin\":\$$PATH' >> \"$$shell_config_file\"$(TERM_RESET)\n" "$(ABSOLUTE_ROOT_DIR)" >&2; \
|
|
|
|
|
fi; \
|
|
|
|
|
if [ "$$infer_repo_is_in_manpath" != "0" ]; then \
|
|
|
|
|
printf "$(TERM_INFO) echo 'export MANPATH=\"%s/infer/man\":\$$MANPATH' >> \"$$shell_config_file\"$(TERM_RESET)\n" "$(ABSOLUTE_ROOT_DIR)" >&2; \
|
|
|
|
|
fi; \
|
[build] switch to 4.05.0+flambda by default
Summary:
With flambda (`-O3`), compilation time is ~5x slower, but the backend is ~25% faster!
To mitigate the atrocious compilation times, introduce a new `opt` build mode in the jbuilder files.
- build in "opt" mode by default from the toplevel (so that install scripts and external users get the fastest infer by default), in "default" mode by default from infer/src (since the latter is only called directly by infer devs, for faster builds)
- `make byte` is as fast as before in any mode
- `make test` will build "opt" by default, which is very slow. Solution for testing (or building the models) locally: `make BUILD_MODE=default test`.
- You can even change the default locally with `export BUILD_MODE=default`.
The benchmarks are to be taken with a sizable pinch of salt because I ran them only once and other stuff could be running in the background. That said, the perf win is consistent across all projects, with 15-20% win in wallclock time and around 25% win in total CPU time, ~9% win in sys time, and ~25% fewer minor allocations, and ~5-10% fewer overall allocations. This is only for the backend; the capture is by and large unaffected (either the same or a tad faster within noise range).
Here are the results running on OpenSSL 1.0.2d on osx (12 cores, 32G RAM)
=== base
infer binary: 26193088 bytes
compile time: 40s
capture:
```lang=text
real 1m7.513s
user 3m11.437s
sys 0m55.236s
```
analysis:
```lang=text
real 5m41.580s
user 61m37.855s
sys 1m12.870s
```
Memory profile:
```lang=json
{
...
"minor_gb": 0.1534719169139862,
"promoted_gb": 0.0038930922746658325,
"major_gb": 0.4546157643198967,
"allocated_gb": 0.6041945889592171,
"minor_collections": 78,
"major_collections": 23,
"compactions": 7,
"top_heap_gb": 0.07388687133789062,
"stack_kb": 0.3984375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda with stock options (no `-Oclassic`, just the same flags as base)
Exactly the same as base.
=== flambda `-O3`
infer binary: 56870376 bytes (2.17x bigger)
compile time: 191s (4.78x slower)
capture is the same as base:
```lang=text
real 1m9.203s
user 3m12.242s
sys 0m58.905s
```
analysis is ~20% wallclock time faster, ~25% CPU time faster:
```lang=text
real 4m32.656s
user 46m43.987s
sys 1m2.424s
```
memory usage is a bit lower too:
```lang=json
{
...
"minor_gb": 0.11583046615123749, // 75% of previous
"promoted_gb": 0.00363825261592865, // 93% of previous
"major_gb": 0.45415670424699783, // about same
"allocated_gb": 0.5663489177823067, // 94% of previous
"minor_collections": 73,
"major_collections": 22,
"compactions": 7,
"top_heap_gb": 0.07165145874023438,
"stack_kb": 0.3359375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda `-O2`
Not nearly as exciting as `-O3`, but the compilation cost is still quite high:
infer: 37826856 bytes
compilation of infer: 100s
Capture and analysis timings are mostly the same as base.
Reviewed By: jberdine
Differential Revision: D4867979
fbshipit-source-id: 99230b7
7 years ago
|
|
|
fi; \
|
|
|
|
|
if [ -z "$(ORIG_SHELL_BUILD_MODE)" ]; then \
|
[build] switch to 4.05.0+flambda by default
Summary:
With flambda (`-O3`), compilation time is ~5x slower, but the backend is ~25% faster!
To mitigate the atrocious compilation times, introduce a new `opt` build mode in the jbuilder files.
- build in "opt" mode by default from the toplevel (so that install scripts and external users get the fastest infer by default), in "default" mode by default from infer/src (since the latter is only called directly by infer devs, for faster builds)
- `make byte` is as fast as before in any mode
- `make test` will build "opt" by default, which is very slow. Solution for testing (or building the models) locally: `make BUILD_MODE=default test`.
- You can even change the default locally with `export BUILD_MODE=default`.
The benchmarks are to be taken with a sizable pinch of salt because I ran them only once and other stuff could be running in the background. That said, the perf win is consistent across all projects, with 15-20% win in wallclock time and around 25% win in total CPU time, ~9% win in sys time, and ~25% fewer minor allocations, and ~5-10% fewer overall allocations. This is only for the backend; the capture is by and large unaffected (either the same or a tad faster within noise range).
Here are the results running on OpenSSL 1.0.2d on osx (12 cores, 32G RAM)
=== base
infer binary: 26193088 bytes
compile time: 40s
capture:
```lang=text
real 1m7.513s
user 3m11.437s
sys 0m55.236s
```
analysis:
```lang=text
real 5m41.580s
user 61m37.855s
sys 1m12.870s
```
Memory profile:
```lang=json
{
...
"minor_gb": 0.1534719169139862,
"promoted_gb": 0.0038930922746658325,
"major_gb": 0.4546157643198967,
"allocated_gb": 0.6041945889592171,
"minor_collections": 78,
"major_collections": 23,
"compactions": 7,
"top_heap_gb": 0.07388687133789062,
"stack_kb": 0.3984375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda with stock options (no `-Oclassic`, just the same flags as base)
Exactly the same as base.
=== flambda `-O3`
infer binary: 56870376 bytes (2.17x bigger)
compile time: 191s (4.78x slower)
capture is the same as base:
```lang=text
real 1m9.203s
user 3m12.242s
sys 0m58.905s
```
analysis is ~20% wallclock time faster, ~25% CPU time faster:
```lang=text
real 4m32.656s
user 46m43.987s
sys 1m2.424s
```
memory usage is a bit lower too:
```lang=json
{
...
"minor_gb": 0.11583046615123749, // 75% of previous
"promoted_gb": 0.00363825261592865, // 93% of previous
"major_gb": 0.45415670424699783, // about same
"allocated_gb": 0.5663489177823067, // 94% of previous
"minor_collections": 73,
"major_collections": 22,
"compactions": 7,
"top_heap_gb": 0.07165145874023438,
"stack_kb": 0.3359375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda `-O2`
Not nearly as exciting as `-O3`, but the compilation cost is still quite high:
infer: 37826856 bytes
compilation of infer: 100s
Capture and analysis timings are mostly the same as base.
Reviewed By: jberdine
Differential Revision: D4867979
fbshipit-source-id: 99230b7
7 years ago
|
|
|
echo >&2; \
|
[RFC][build] Use dune environments and profiles instead of contexts
Summary:
With profiles and `(env ...)` stanza it's possible to consolidate
various ocamlc/ocamlopt/etc setups in a single place.
Where previously we needed to append `dune.common` to every dune file
and specify `flags` and `ocamlopt_flags` now the flags are specified
in `env` and applied accross the board.
This allows to
1. simplify build definitions,
2. avoid the need to generate dune files,
3. use plain sexps instead of OCaml and JBuilder plugin in build
files.
(I'll try to address 2 and 3 in the followup patches).
Existing `make` targets should continue working as before. Also, we
can use dune CLI like so:
```
infer/src$ dune printenv --profile opt # <- very useful for introspection
infer/src$ dune build check
infer/src$ dune build check --profile test
infer/src$ dune build infer.exe --profile dev
infer/src$ dune build infer.exe --profile opt
```
Also, with just 1 context something like `dune runtest` will run unit
tests only once instead of N times, where N is the number of contexts.
Now, there's one difference compared to the previous setup with
contexts:
- Previously, each context had its own build folder, and building infer
in opt context didn't invalidate any of the build artifacts in default
context. Therefore, alternating between `make` and `make opt` had low
overhead at the expense of having N copies of all build artifacts (1
for every context).
- Now, there's just 1 build folder and switching between profiles does
invalidate some artifacts (but not all) and rebuild takes a bit more
time.
So, if you're alternating like crazy between profiles your experience
may get worse (but not necessarily, more on that below). If you want
to trigger an opt build occasionally, you shouldn't notice much
difference.
For those who are concerned about slower build times when alternating
between different build profiles, there's a solution: [dune
cache](https://dune.readthedocs.io/en/stable/caching.html).
You can enable it by creating a file `~/.config/dune/config` with the
following contents:
```
(lang dune 2.0)
(cache enabled)
```
With cache enabled switching between different build profiles (but
also branches and commits) has ~0 overhead.
Dune cache works fine on Linux, but unfortunately there are [certain
problems with
MacOS](https://github.com/ocaml/dune/issues/3233) (hopefully, those
will be fixed soon and then there will be no downsides to using
profiles compared to contexts for our case).
Reviewed By: jvillard
Differential Revision: D20247864
fbshipit-source-id: 5f8afa0db
5 years ago
|
|
|
echo '$(TERM_INFO)*** NOTE: Set `BUILD_MODE=dev` in your shell to disable flambda by default.$(TERM_RESET)' >&2; \
|
[build] switch to 4.05.0+flambda by default
Summary:
With flambda (`-O3`), compilation time is ~5x slower, but the backend is ~25% faster!
To mitigate the atrocious compilation times, introduce a new `opt` build mode in the jbuilder files.
- build in "opt" mode by default from the toplevel (so that install scripts and external users get the fastest infer by default), in "default" mode by default from infer/src (since the latter is only called directly by infer devs, for faster builds)
- `make byte` is as fast as before in any mode
- `make test` will build "opt" by default, which is very slow. Solution for testing (or building the models) locally: `make BUILD_MODE=default test`.
- You can even change the default locally with `export BUILD_MODE=default`.
The benchmarks are to be taken with a sizable pinch of salt because I ran them only once and other stuff could be running in the background. That said, the perf win is consistent across all projects, with 15-20% win in wallclock time and around 25% win in total CPU time, ~9% win in sys time, and ~25% fewer minor allocations, and ~5-10% fewer overall allocations. This is only for the backend; the capture is by and large unaffected (either the same or a tad faster within noise range).
Here are the results running on OpenSSL 1.0.2d on osx (12 cores, 32G RAM)
=== base
infer binary: 26193088 bytes
compile time: 40s
capture:
```lang=text
real 1m7.513s
user 3m11.437s
sys 0m55.236s
```
analysis:
```lang=text
real 5m41.580s
user 61m37.855s
sys 1m12.870s
```
Memory profile:
```lang=json
{
...
"minor_gb": 0.1534719169139862,
"promoted_gb": 0.0038930922746658325,
"major_gb": 0.4546157643198967,
"allocated_gb": 0.6041945889592171,
"minor_collections": 78,
"major_collections": 23,
"compactions": 7,
"top_heap_gb": 0.07388687133789062,
"stack_kb": 0.3984375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda with stock options (no `-Oclassic`, just the same flags as base)
Exactly the same as base.
=== flambda `-O3`
infer binary: 56870376 bytes (2.17x bigger)
compile time: 191s (4.78x slower)
capture is the same as base:
```lang=text
real 1m9.203s
user 3m12.242s
sys 0m58.905s
```
analysis is ~20% wallclock time faster, ~25% CPU time faster:
```lang=text
real 4m32.656s
user 46m43.987s
sys 1m2.424s
```
memory usage is a bit lower too:
```lang=json
{
...
"minor_gb": 0.11583046615123749, // 75% of previous
"promoted_gb": 0.00363825261592865, // 93% of previous
"major_gb": 0.45415670424699783, // about same
"allocated_gb": 0.5663489177823067, // 94% of previous
"minor_collections": 73,
"major_collections": 22,
"compactions": 7,
"top_heap_gb": 0.07165145874023438,
"stack_kb": 0.3359375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda `-O2`
Not nearly as exciting as `-O3`, but the compilation cost is still quite high:
infer: 37826856 bytes
compilation of infer: 100s
Capture and analysis timings are mostly the same as base.
Reviewed By: jberdine
Differential Revision: D4867979
fbshipit-source-id: 99230b7
7 years ago
|
|
|
echo '$(TERM_INFO)*** NOTE: Compiling with flambda is ~5 times slower than without, so unless you are$(TERM_RESET)' >&2; \
|
|
|
|
|
echo '$(TERM_INFO)*** NOTE: testing infer on a very large project it will not be worth it. Use the$(TERM_RESET)' >&2; \
|
|
|
|
|
echo '$(TERM_INFO)*** NOTE: commands below to set the default build mode. You can then use `make opt`$(TERM_RESET)' >&2; \
|
[build] switch to 4.05.0+flambda by default
Summary:
With flambda (`-O3`), compilation time is ~5x slower, but the backend is ~25% faster!
To mitigate the atrocious compilation times, introduce a new `opt` build mode in the jbuilder files.
- build in "opt" mode by default from the toplevel (so that install scripts and external users get the fastest infer by default), in "default" mode by default from infer/src (since the latter is only called directly by infer devs, for faster builds)
- `make byte` is as fast as before in any mode
- `make test` will build "opt" by default, which is very slow. Solution for testing (or building the models) locally: `make BUILD_MODE=default test`.
- You can even change the default locally with `export BUILD_MODE=default`.
The benchmarks are to be taken with a sizable pinch of salt because I ran them only once and other stuff could be running in the background. That said, the perf win is consistent across all projects, with 15-20% win in wallclock time and around 25% win in total CPU time, ~9% win in sys time, and ~25% fewer minor allocations, and ~5-10% fewer overall allocations. This is only for the backend; the capture is by and large unaffected (either the same or a tad faster within noise range).
Here are the results running on OpenSSL 1.0.2d on osx (12 cores, 32G RAM)
=== base
infer binary: 26193088 bytes
compile time: 40s
capture:
```lang=text
real 1m7.513s
user 3m11.437s
sys 0m55.236s
```
analysis:
```lang=text
real 5m41.580s
user 61m37.855s
sys 1m12.870s
```
Memory profile:
```lang=json
{
...
"minor_gb": 0.1534719169139862,
"promoted_gb": 0.0038930922746658325,
"major_gb": 0.4546157643198967,
"allocated_gb": 0.6041945889592171,
"minor_collections": 78,
"major_collections": 23,
"compactions": 7,
"top_heap_gb": 0.07388687133789062,
"stack_kb": 0.3984375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda with stock options (no `-Oclassic`, just the same flags as base)
Exactly the same as base.
=== flambda `-O3`
infer binary: 56870376 bytes (2.17x bigger)
compile time: 191s (4.78x slower)
capture is the same as base:
```lang=text
real 1m9.203s
user 3m12.242s
sys 0m58.905s
```
analysis is ~20% wallclock time faster, ~25% CPU time faster:
```lang=text
real 4m32.656s
user 46m43.987s
sys 1m2.424s
```
memory usage is a bit lower too:
```lang=json
{
...
"minor_gb": 0.11583046615123749, // 75% of previous
"promoted_gb": 0.00363825261592865, // 93% of previous
"major_gb": 0.45415670424699783, // about same
"allocated_gb": 0.5663489177823067, // 94% of previous
"minor_collections": 73,
"major_collections": 22,
"compactions": 7,
"top_heap_gb": 0.07165145874023438,
"stack_kb": 0.3359375,
"minor_heap_kb": 8192.0,
...
}
```
=== flambda `-O2`
Not nearly as exciting as `-O3`, but the compilation cost is still quite high:
infer: 37826856 bytes
compilation of infer: 100s
Capture and analysis timings are mostly the same as base.
Reviewed By: jberdine
Differential Revision: D4867979
fbshipit-source-id: 99230b7
7 years ago
|
|
|
echo '$(TERM_INFO)*** NOTE: when you really do want to enable flambda.$(TERM_RESET)' >&2; \
|
|
|
|
|
echo >&2; \
|
[RFC][build] Use dune environments and profiles instead of contexts
Summary:
With profiles and `(env ...)` stanza it's possible to consolidate
various ocamlc/ocamlopt/etc setups in a single place.
Where previously we needed to append `dune.common` to every dune file
and specify `flags` and `ocamlopt_flags` now the flags are specified
in `env` and applied accross the board.
This allows to
1. simplify build definitions,
2. avoid the need to generate dune files,
3. use plain sexps instead of OCaml and JBuilder plugin in build
files.
(I'll try to address 2 and 3 in the followup patches).
Existing `make` targets should continue working as before. Also, we
can use dune CLI like so:
```
infer/src$ dune printenv --profile opt # <- very useful for introspection
infer/src$ dune build check
infer/src$ dune build check --profile test
infer/src$ dune build infer.exe --profile dev
infer/src$ dune build infer.exe --profile opt
```
Also, with just 1 context something like `dune runtest` will run unit
tests only once instead of N times, where N is the number of contexts.
Now, there's one difference compared to the previous setup with
contexts:
- Previously, each context had its own build folder, and building infer
in opt context didn't invalidate any of the build artifacts in default
context. Therefore, alternating between `make` and `make opt` had low
overhead at the expense of having N copies of all build artifacts (1
for every context).
- Now, there's just 1 build folder and switching between profiles does
invalidate some artifacts (but not all) and rebuild takes a bit more
time.
So, if you're alternating like crazy between profiles your experience
may get worse (but not necessarily, more on that below). If you want
to trigger an opt build occasionally, you shouldn't notice much
difference.
For those who are concerned about slower build times when alternating
between different build profiles, there's a solution: [dune
cache](https://dune.readthedocs.io/en/stable/caching.html).
You can enable it by creating a file `~/.config/dune/config` with the
following contents:
```
(lang dune 2.0)
(cache enabled)
```
With cache enabled switching between different build profiles (but
also branches and commits) has ~0 overhead.
Dune cache works fine on Linux, but unfortunately there are [certain
problems with
MacOS](https://github.com/ocaml/dune/issues/3233) (hopefully, those
will be fixed soon and then there will be no downsides to using
profiles compared to contexts for our case).
Reviewed By: jvillard
Differential Revision: D20247864
fbshipit-source-id: 5f8afa0db
5 years ago
|
|
|
printf "$(TERM_INFO) export BUILD_MODE=dev$(TERM_RESET)\n" >&2; \
|
|
|
|
|
printf "$(TERM_INFO) echo 'export BUILD_MODE=dev' >> \"$$shell_config_file\"$(TERM_RESET)\n" >&2; \
|
|
|
|
|
fi
|
|
|
|
|
$(QUIET)PATH=$(ORIG_SHELL_PATH); if [ "$$(ocamlc -where 2>/dev/null)" != "$$($(OCAMLC) -where)" ]; then \
|
|
|
|
|
echo >&2; \
|
|
|
|
|
echo '$(TERM_INFO)*** NOTE: The current shell is not set up for the right opam switch.$(TERM_RESET)' >&2; \
|
|
|
|
|
echo '$(TERM_INFO)*** NOTE: Please run:$(TERM_RESET)' >&2; \
|
|
|
|
|
echo >&2; \
|
|
|
|
|
echo "$(TERM_INFO) eval \$$($(OPAM) env)$(TERM_RESET)" >&2; \
|
|
|
|
|
fi
|
|
|
|
|
|
|
|
|
|
GHPAGES ?= no
|
|
|
|
|
|
|
|
|
|
.PHONY: doc
|
|
|
|
|
doc: src_build_common
|
|
|
|
|
$(QUIET)$(call silent_on_success,Generating infer documentation,\
|
|
|
|
|
$(MAKE_SOURCE) doc)
|
|
|
|
|
# do not call the browser if we are publishing the docs
|
|
|
|
|
ifeq ($(filter doc-publish,${MAKECMDGOALS}),)
|
|
|
|
|
$(QUIET)$(call silent_on_success,Opening in browser,\
|
|
|
|
|
browse $(INFER_DIR)/_build/default/_doc/_html/index.html)
|
|
|
|
|
$(QUIET)echo "Tip: you can generate the doc for all the opam dependencies of infer like this:"
|
|
|
|
|
$(QUIET)echo
|
|
|
|
|
$(QUIET)echo " odig odoc # takes a while, run it only when the dependencies change"
|
|
|
|
|
$(QUIET)echo " odig doc"
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
.PHONY: doc-publish
|
|
|
|
|
doc-publish: doc $(INFER_GROFF_MANUALS)
|
|
|
|
|
ifeq ($(GHPAGES),no)
|
|
|
|
|
$(QUIET)echo "$(TERM_ERROR)Please set GHPAGES to a checkout of the gh-pages branch of the GitHub repo of infer$(TERM_RESET)" >&2
|
|
|
|
|
$(QUIET)exit 1
|
|
|
|
|
endif
|
|
|
|
|
# sanity check to avoid cryptic error messages and potentially annoying side-effects
|
|
|
|
|
$(QUIET)if ! [ -d "$(GHPAGES)"/static/man ]; then \
|
|
|
|
|
echo "$(TERM_ERROR)ERROR: GHPAGES doesn't seem to point to a checkout of the gh-pages branch of the GitHub repo of infer:$(TERM_RESET)" >&2; \
|
|
|
|
|
echo "$(TERM_ERROR)ERROR: '$(GHPAGES)/static/man' not found or not a directory.$(TERM_RESET)" >&2; \
|
|
|
|
|
echo "$(TERM_ERROR)ERROR: Please fix this and try again.$(TERM_RESET)" >&2; \
|
|
|
|
|
exit 1; \
|
|
|
|
|
fi
|
|
|
|
|
$(QUIET)$(call silent_on_success,Copying man pages,\
|
|
|
|
|
$(REMOVE_DIR) "$(GHPAGES)"/static/man/*; \
|
|
|
|
|
for man in $(INFER_GROFF_MANUALS); do \
|
|
|
|
|
groff -Thtml "$$man" > "$(GHPAGES)"/static/man/$$(basename "$$man").html; \
|
|
|
|
|
done)
|
|
|
|
|
ifeq ($(IS_FACEBOOK_TREE),no)
|
|
|
|
|
$(QUIET)$(call silent_on_success,Copying OCaml modules documentation,\
|
|
|
|
|
version=$$($(INFER_BIN) --version | head -1 | cut -d ' ' -f 3 | cut -c 2-); \
|
[RFC][build] Use dune environments and profiles instead of contexts
Summary:
With profiles and `(env ...)` stanza it's possible to consolidate
various ocamlc/ocamlopt/etc setups in a single place.
Where previously we needed to append `dune.common` to every dune file
and specify `flags` and `ocamlopt_flags` now the flags are specified
in `env` and applied accross the board.
This allows to
1. simplify build definitions,
2. avoid the need to generate dune files,
3. use plain sexps instead of OCaml and JBuilder plugin in build
files.
(I'll try to address 2 and 3 in the followup patches).
Existing `make` targets should continue working as before. Also, we
can use dune CLI like so:
```
infer/src$ dune printenv --profile opt # <- very useful for introspection
infer/src$ dune build check
infer/src$ dune build check --profile test
infer/src$ dune build infer.exe --profile dev
infer/src$ dune build infer.exe --profile opt
```
Also, with just 1 context something like `dune runtest` will run unit
tests only once instead of N times, where N is the number of contexts.
Now, there's one difference compared to the previous setup with
contexts:
- Previously, each context had its own build folder, and building infer
in opt context didn't invalidate any of the build artifacts in default
context. Therefore, alternating between `make` and `make opt` had low
overhead at the expense of having N copies of all build artifacts (1
for every context).
- Now, there's just 1 build folder and switching between profiles does
invalidate some artifacts (but not all) and rebuild takes a bit more
time.
So, if you're alternating like crazy between profiles your experience
may get worse (but not necessarily, more on that below). If you want
to trigger an opt build occasionally, you shouldn't notice much
difference.
For those who are concerned about slower build times when alternating
between different build profiles, there's a solution: [dune
cache](https://dune.readthedocs.io/en/stable/caching.html).
You can enable it by creating a file `~/.config/dune/config` with the
following contents:
```
(lang dune 2.0)
(cache enabled)
```
With cache enabled switching between different build profiles (but
also branches and commits) has ~0 overhead.
Dune cache works fine on Linux, but unfortunately there are [certain
problems with
MacOS](https://github.com/ocaml/dune/issues/3233) (hopefully, those
will be fixed soon and then there will be no downsides to using
profiles compared to contexts for our case).
Reviewed By: jvillard
Differential Revision: D20247864
fbshipit-source-id: 5f8afa0db
5 years ago
|
|
|
rsync -a --delete $(SRC_DIR)/_build/default/_doc/_html/ "$(GHPAGES)"/static/odoc/"$$version"; \
|
|
|
|
|
$(REMOVE) "$(GHPAGES)"/static/odoc/latest; \
|
|
|
|
|
$(LN_S) "$$version" "$(GHPAGES)"/static/odoc/latest)
|
|
|
|
|
else
|
|
|
|
|
$(QUIET)echo "Not an open-source tree, skipping the API docs generation"
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
# print list of targets
|
|
|
|
|
.PHONY: show-targets
|
|
|
|
|
show-targets:
|
|
|
|
|
$(QUIET)$(MAKE) -pqrR . | grep --only-matching -e '^[a-zA-Z0-9][^ ]*:' | cut -d ':' -f 1 | sort
|
