1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
|
---
layout: documentation
---
C++ Basics
==========
Use fully qualified include paths
---------------------------------
Includes are relative to the root of your workspace. For example, suppose
you have the following directory structure:
```
[workspace]/
WORKSPACE
a/
BUILD
a.h
a.cc
b/
BUILD
b.h
b.cc
main.cc
```
If _b/main.cc_ needs to include b.h then we'd create the following _b/BUILD_
file:
```python
cc_library(
name = "b",
srcs = ["b.cc"],
hdrs = ["b.h"],
)
cc_binary(
name = "main",
srcs = ["main.cc"],
deps = [":b"],
)
```
_b/main.cc_ would have the following include statement:
```cpp
#include "b/b.h"
```
Note that the full path from the package root is used. If we want _b/main.cc_ to
also depend on _a/a.h_, we'd add the rule to _a/BUILD_:
```python
cc_library(
name = "a",
srcs = ["a.cc"],
hdrs = ["a.h"],
visibility = ["//b:__pkg__"],
)
```
Then we'd add a dependency to _b/BUILD_:
```python
cc_binary(
name = "main",
srcs = ["main.cc"],
deps = [
":b",
"//a",
],
)
```
And the following include to _b/main.cc_:
```cpp
#include "a/a.h"
```
_b/main.cc_ will then be able to access symbols from _a/a.h_ or _b/b.h_.
Transitive includes
-------------------
If a file includes a header then the file's rule should depend on that header's
library. Conversely, only direct dependencies need to be specified as
dependencies. For example, suppose _sandwich.h_ includes _bread.h_ and
_bread.h_ includes _flour.h_. _sandwich.h_ doesn't include _flour.h_ (who wants
flour in their sandwich?), so the BUILD file would look like:
```python
cc_library(
name = "sandwich",
srcs = ["sandwich.cc"],
hdrs = ["sandwich.h"],
deps = [":bread"],
)
cc_library(
name = "bread",
srcs = ["bread.cc"],
hdrs = ["bread.h"],
deps = [":flour"],
)
cc_library(
name = "flour",
srcs = ["flour.cc"],
hdrs = ["flour.h"],
)
```
This expresses that the "sandwich" library depends on the "bread" library,
which depends on the "flour" library.
Adding include paths
--------------------
Sometimes you cannot (or do not want to) base include paths at the workspace
root. Existing libaries might already have a include directory that doesn't
match its path in your workspace. For example, suppose you have the following
directory structure:
```
[workspace]/
WORKSPACE
third_party/
some_lib/
include/
some_lib.h
BUILD
some_lib.cc
```
Bazel will expect _some_lib.h_ to be included as
`third_party/some_lib/include/some_lib.h`, but suppose _some_lib.cc_ includes
`"include/some_lib.h"`. To make that include path valid,
_third\_party/some_lib/BUILD_ will need to specify that the _some_lib/_
directory is an include directory:
```python
cc_library(
name = "some_lib",
srcs = ["some_lib.cc"],
hdrs = ["some_lib.h"],
includes = ["."],
)
```
This is especially useful for external dependencies, as their header files
must otherwise be included with an "external/[repository-name]/" prefix.
Including external libraries: an example
----------------------------------------
Suppose you are using [Google Test](https://code.google.com/p/googletest/). You
can use one of the `new_` repository functions in the _WORKSPACE_ file to
download Google Test and make it available in your repository:
```python
new_http_archive(
name = "gtest-repo",
url = "https://googletest.googlecode.com/files/gtest-1.7.0.zip",
sha256 = "247ca18dd83f53deb1328be17e4b1be31514cedfc1e3424f672bf11fd7e0d60d",
build_file = "gtest.BUILD",
)
bind(
name = "gtest/main",
actual = "@gtest-repo//:main",
)
```
Then create _gtest.BUILD_, a BUILD file to use to compile Google Test.
Google Test has several "special" requirements that make its `cc_library` rule
more complicated:
* _gtest-1.7.0/src/gtest-all.cc_ `#include`s all of the other files in
_gtest-1.7.0/src/_, so we need to exclude it from the compile or we'll get
link errors for duplicate symbols.
* It uses header files that relative to the _gtest-1.7.0/include/_ directory
(`"gtest/gtest.h"`), so we must add that directory the includes.
* It uses "private" header files in src/, so we add "." to the includes so it
can `#include "src/gtest-internal-inl.h"`.
* It needs to link in pthread, so we add that as a `linkopt`.
The final rule looks like this:
```python
cc_library(
name = "main",
srcs = glob(
["gtest-1.7.0/src/*.cc"],
exclude = ["gtest-1.7.0/src/gtest-all.cc"]
),
hdrs = glob(["gtest-1.7.0/include/**/*.h"]),
includes = [
"gtest-1.7.0",
"gtest-1.7.0/include"
],
linkopts = ["-pthread"],
visibility = ["//visibility:public"],
)
```
Now `cc_` rules can depend on `//external:gtest/main`.
For example, we could create a test such as:
```cpp
#include "gtest/gtest.h"
TEST(FactorialTest, Negative) {
EXPECT_EQ(1, 1);
}
```
Then create a BUILD file for your tests:
```python
cc_test(
name = "my_test",
srcs = ["my_test.cc"],
deps = ["//external:gtest/main"],
)
```
You can then use `bazel test` to run the test.
|
