How to Use LCOV for Analyzing the Test Coverage Against PostgreSQL

The LCOV (Linux Coverage Visualizer) tool is used for code coverage measurement and visualization. It works in conjunction with another tool, GCC's gcov, to provide a user-friendly way to analyze how thoroughly your tests cover your code.

Here's a breakdown of what LCOV does:

• Leverages gcov data: LCOV relies on data generated by gcov, which tracks which parts of your code are executed during testing.
• Creates HTML reports: LCOV takes this data and transforms it into interactive HTML reports.
• Visualizes coverage: These reports show your source code annotated with coverage information. You can see which lines, functions, or branches of your code were executed by the tests.
• Supports navigation: For large projects, LCOV provides features like an overview page to navigate the coverage data efficiently.

Through the coverage report, you can identify areas where tests on the codebase might be missing and improve your test suite accordingly.

Learn more about setting up tests from the world’s top PostgreSQL experts. Join EDB Postgres Training.

keyboard_arrow_up

Steps to perform test coverage for any schedule:

1. Cleanup the source under “postgressql” directory.

cd /home/prabhat/PG/PGsrc/postgresql 
make clean 
make distclean 
git clean -fdx 
git reset HEAD --hard

2. Configure with "--enable-coverage" option and install.

./configure --with-zlib --enable-debug --enable-depend --prefix=$PWD/inst/  
--enable-cassert --with-readline --with-openssl --enable-coverage CFLAGS="-g -O0"; make; make install;

3. Run the schedule (suppose, make check) From /src/test/regress directory.

cd src/test/regress 
make check

4. Run the coverage under "/home/prabhat/PG/PGsrc/postgresql" directory.

cd ../../../ 
make coverage-html

5. Open the index.html in the Firefox browser under the "coverage" directory.

cd coverage/ 
firefox index.html

Steps to perform test coverage for newly added test cases:

1. Cleanup the source under “postgressql” directory.

cd /home/prabhat/PG/PGsrc/postgresql 
make clean 
make distclean 
git clean -fdx 
git reset HEAD –hard

2. Configure with "--enable-coverage" option and install the source.

./configure --with-zlib --enable-debug --enable-depend --prefix=$PWD/inst/  
--enable-cassert --with-readline --with-openssl --enable-coverage CFLAGS="-g -O0"; make; make install;

3. Prepare the required test case files. Example:

cd /home/prabhat/PG/PGsrc/postgresql/inst/bin 
[prabhat@localhost bin]$ cat /tmp/test.sql 
-- Test to demonstrate generate_series. 
select x as "col1" from generate_series(1,5) x; 
select x as "col1" from generate_series(1,5,2) x; 
select x+5 as "col1" from generate_series(1,10,2) x;
   
[prabhat@localhost bin]$ ./psql postgres -a -f /tmp/test.sql > /tmp/test.out 
[prabhat@localhost bin]$ cat /tmp/test.out  
    
-- Test to demonstrate generate_series. 
select x as "col1" from generate_series(1,5) x; 
col1  
------ 
    
  1 
    
  2 
    
  3 
    
  4 
    
  5 
    
(5 rows) 
   
select x as "col1" from generate_series(1,5,2) x; 
    
col1  
    
------ 
  1 
    
  3 
    
  5 
    
(3 rows) 
    
select x+5 as "col1" from generate_series(1,10,2) x; 
    
col1  
------ 
    
  6 
    
  8 
    
 10 
    
 12 
    
 14 
    
(5 rows)

4. Keep test case files in specific directory under src/test/regress/ (sql, expected)

cd /home/prabhat/PG/PGsrc/postgresql/src/test/regress 
mkdir sql/generate_series 
mkdir expected/generate_series 
mkdir results/generate_series 
cp /tmp/test.sql sql/generate_series/. 
cp /tmp/test.out expected/generate_series/.

5. Run the test or schedule under "/src/test/regress" directory.

cd /src/test/regress 
    
./pg_regress --temp-instance=/tmp/data1 generate_series/test 
   
Example: 
[prabhat@localhost regress]$ ./pg_regress --temp-instance=/tmp/data1 generate_series/test 
    
============== creating temporary instance        	============== 
    
============== initializing database system       	============== 
    
============== starting postmaster                	============== 
    
running on port 58080 with PID 110199 
    
============== creating database "regression"     	============== 
    
CREATE DATABASE 
    
ALTER DATABASE 
    
============== running regression test queries    	============== 
    
test generate_series/test     	... ok       	24 ms 
    
============== shutting down postmaster           	============== 
    
============== removing temporary instance        	============== 
    
    
    
===================== 
    
 All 1 tests passed.  
    
=====================

6. Run the coverage under "/home/prabhat/PG/PGsrc/postgresql" directory.

cd ../../../ 
make coverage-html

Example:

[prabhat@localhost postgresql]$ make coverage-html 
/usr/bin/lcov --gcov-tool /usr/bin/gcov -q --no-external -c -i -d . -d . -o lcov_base.info 
/usr/bin/lcov --gcov-tool /usr/bin/gcov -q --no-external -c -d . -d . -o lcov_test.info 
rm -rf coverage 
/usr/bin/genhtml -q --legend -o coverage --title='PostgreSQL 14devel' --num-spaces=4 --prefix='/home/prabhat/PG/PGsrc/postgresql' lcov_base.info lcov_test.info 
touch coverage-html-stamp

7. Open the “index.html” in the firefox browser under the "coverage" directory.

cd coverage/ 
firefox index.html

To maximize the benefits of LCOV for PostgreSQL projects, it's important to follow best practices and be prepared to troubleshoot common issues. Here are some recommendations and tips: