This gives some flexibility to embedders.

Currently, embedders have to build a binary for each test file.

After this CL, embedders can still choose to have a binary for each test
file, by linking each test file with a googletest target that includes
main() (usually "gtest_main"). Embedders can also choose to build a
single binary for almost all test files, and link with a googletest
target that includes main(). The latter is more convenient for projects
that have very few test binaries, like Chromium.

PiperOrigin-RevId: 419470798

PiperOrigin-RevId: 282373286

PiperOrigin-RevId: 281815695

Use clang-format to correct formatting to be in agreement with the [Google C++ Style Guide](https://google.github.io/styleguide/cppguide.html). Doing this simplifies the process of accepting changes. Also fixed a few warnings flagged by clang-tidy.

PiperOrigin-RevId: 246350737

This will result in smaller code generation when Status instances are
passed around.

Benchmarks don't indicate a significant change either way.
CPU:        48 * Intel(R) Xeon(R) CPU E5-2690 v3 @ 2.60GHz
CPUCache:   30720 KB
Keys:       16 bytes each
Values:     100 bytes each (50 bytes after compression)
Entries:    1000000
RawSize:    110.6 MB (estimated)
FileSize:   62.9 MB (estimated)

Baseline:
fillseq      :       3.589 micros/op;   30.8 MB/s
fillsync     :    4165.299 micros/op;    0.0 MB/s (1000 ops)
fillrandom   :       5.864 micros/op;   18.9 MB/s
overwrite    :       7.830 micros/op;   14.1 MB/s
readrandom   :       5.534 micros/op; (1000000 of 1000000 found)
readrandom   :       4.292 micros/op; (1000000 of 1000000 found)
readseq      :       0.312 micros/op;  354.1 MB/s
readreverse  :       0.501 micros/op;  220.8 MB/s
compact      :  886211.000 micros/op;
readrandom   :       3.518 micros/op; (1000000 of 1000000 found)
readseq      :       0.251 micros/op;  441.2 MB/s
readreverse  :       0.456 micros/op;  242.4 MB/s
fill100K     :    1329.723 micros/op;   71.7 MB/s (1000 ops)
crc32c       :       1.976 micros/op; 1976.7 MB/s (4K per op)
snappycomp   :       4.705 micros/op;  830.2 MB/s (output: 55.1%)
snappyuncomp :       0.958 micros/op; 4079.1 MB/s
acquireload  :       0.727 micros/op; (each op is 1000 loads)

New:
fillseq      :       3.129 micros/op;   35.4 MB/s
fillsync     :    2748.099 micros/op;    0.0 MB/s (1000 ops)
fillrandom   :       5.394 micros/op;   20.5 MB/s
overwrite    :       7.253 micros/op;   15.3 MB/s
readrandom   :       5.655 micros/op; (1000000 of 1000000 found)
readrandom   :       4.425 micros/op; (1000000 of 1000000 found)
readseq      :       0.298 micros/op;  371.3 MB/s
readreverse  :       0.508 micros/op;  217.9 MB/s
compact      :  885842.000 micros/op;
readrandom   :       3.545 micros/op; (1000000 of 1000000 found)
readseq      :       0.252 micros/op;  438.2 MB/s
readreverse  :       0.425 micros/op;  260.2 MB/s
fill100K     :    1418.347 micros/op;   67.2 MB/s (1000 ops)
crc32c       :       1.987 micros/op; 1966.0 MB/s (4K per op)
snappycomp   :       4.767 micros/op;  819.4 MB/s (output: 55.1%)
snappyuncomp :       0.916 micros/op; 4264.9 MB/s
acquireload  :       0.665 micros/op; (each op is 1000 loads)

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=194002392