Understanding Valgrind output

I am new to Linux. How can I interpret the following output from Valgrind?

valgrind --tool=memcheck --leak-check=yes ./main

It says some blocks are lost. How can I nail down memory leaks?

==2599== HEAP SUMMARY:
==2599==     in use at exit: 17,327 bytes in 55 blocks
==2599==   total heap usage: 180,597 allocs, 180,542 frees, 15,787,989,675 bytes allocated
==2599==
==2599== 9 bytes in 1 blocks are definitely lost in loss record 5 of 19
==2599==    at 0x4025BD3: malloc (vg_replace_malloc.c:236)
==2599==    by 0x41E546F: strdup (strdup.c:43)
==2599==    by 0x804BA2A: Schema::Schema(char*, char*) (Schema.cc:86)
==2599==    by 0x804AD78: CNF::GrowFromParseTree(AndList*, Schema*, Record&) (Comparison.cc:606)
==2599==    by 0x804EE52: main (main.cc:28)
==2599==
==2599== 10 bytes in 2 blocks are definitely lost in loss record 6 of 19
==2599==    at 0x4025BD3: malloc (vg_replace_malloc.c:236)
==2599==    by 0x41E546F: strdup (strdup.c:43)
==2599==    by 0x804BB84: Schema::Schema(char*, char*) (Schema.cc:115)
==2599==    by 0x804AD78: CNF::GrowFromParseTree(AndList*, Schema*, Record&) (Comparison.cc:606)
==2599==    by 0x804EE52: main (main.cc:28)
==2599==
==2599== 13 bytes in 1 blocks are definitely lost in loss record 9 of 19
==2599==    at 0x4025BD3: malloc (vg_replace_malloc.c:236)
==2599==    by 0x41E546F: strdup (strdup.c:43)
==2599==    by 0x804BA2A: Schema::Schema(char*, char*) (Schema.cc:86)
==2599==    by 0x804EDF4: main (main.cc:23)
==2599==
==2599== 13 bytes in 1 blocks are definitely lost in loss record 10 of 19
==2599==    at 0x4025BD3: malloc (vg_replace_malloc.c:236)
==2599==    by 0x41E546F: strdup (strdup.c:43)
==2599==    by 0x804BA2A: Schema::Schema(char*, char*) (Schema.cc:86)
==2599==    by 0x804EEA4: main (main.cc:37)
==2599开发者_JAVA百科==
==2599== 188 bytes in 16 blocks are definitely lost in loss record 16 of 19
==2599==    at 0x4025BD3: malloc (vg_replace_malloc.c:236)
==2599==    by 0x41E546F: strdup (strdup.c:43)
==2599==    by 0x804BB84: Schema::Schema(char*, char*) (Schema.cc:115)
==2599==    by 0x804EDF4: main (main.cc:23)
==2599==
==2599== 188 bytes in 16 blocks are definitely lost in loss record 17 of 19
==2599==    at 0x4025BD3: malloc (vg_replace_malloc.c:236)
==2599==    by 0x41E546F: strdup (strdup.c:43)
==2599==    by 0x804BB84: Schema::Schema(char*, char*) (Schema.cc:115)
==2599==    by 0x804EEA4: main (main.cc:37)
==2599==
==2599== LEAK SUMMARY:
==2599==    definitely lost: 421 bytes in 37 blocks
==2599==    indirectly lost: 0 bytes in 0 blocks
==2599==      possibly lost: 0 bytes in 0 blocks
==2599==    still reachable: 16,906 bytes in 18 blocks
==2599==         suppressed: 0 bytes in 0 blocks
==2599== Reachable blocks (those to which a pointer was found) are not shown.
==2599== To see them, rerun with: --leak-check=full --show-reachable=yes
==2599==
==2599== For counts of detected and suppressed errors, rerun with: -v
==2599== ERROR SUMMARY: 6 errors from 6 contexts (suppressed: 19 from 8)

---

The output shows the stack for where the leaked (and lost, i.e., no pointers to which remain) memory was allocated, in lines 86 and 115 of Schema.cc (an strdup call).

---

From opengroup's strdup() description:

The strdup() function shall return a pointer to a new string, which is a duplicate of the string pointed to by s1. The returned pointer can be passed to free(). A null pointer is returned if the new string cannot be created.

In short, strdup() is calling malloc(). You will need to free() that resultant string when you are done with it. Although, as you're using C++, I'd highly recommend std::string instead if you can. Remember, if you really need the C equivalent string.c_str() will get it for you.