• CS_ERROR_INVALID_PDG_KIND if kind is not valid,
• CS_SUCCESS otherwise.

• CS_SUCCESS on success,
• CS_PDG_NOT_FOUND if no cs_pdg with name proc_name exists.

• CS_TRUNCATED if compilation_unit is not large enough to hold the entire filename (and so the filename is truncated),
• CS_SUCCESS otherwise.

Call this function with compilation_unit NULL and capacity_bytes 0 (zero) to determine the memory needed to hold the entire path.

• CS_ELEMENT_NOT_PRESENT if no ast is associated with pdg.
• CS_SUCCESS otherwise.

• CS_NO_POSITION if pdg has kind cs_pdg_kind_undefined or its entry vertex does not have a position.
• CS_SUCCESS otherwise.

• CS_PDG_IS_UNDEFINED if pdg represents an undefined function.
• CS_SUCCESS on success.

• CS_PDG_IS_UNDEFINED if pdg represents an undefined function.
• CS_SUCCESS on success.

• CS_PDG_IS_UNDEFINED if pdg represents an undefined function.
• CS_SUCCESS on success.

• In an incremental analysis, a procedure will have the same identifier that it did in the incremental parent if and only if the containing compilation unit is not recompiled.
• In all other cases identifiers do not persist between different analyses of the same project (or the same code in a different project).

• CS_TRUNCATED if proc_name is not large enough to hold the entire procedure name (and so the procedure name is truncated),
• CS_SUCCESS otherwise.

Call this function with proc_name NULL and capacity_bytes 0 to obtain the required capacity of proc_name.

• CS_TRUNCATED if proc_name is not large enough to hold the entire procedure name (and so the procedure name is truncated),
• CS_SUCCESS otherwise.

Call this function with proc_name NULL and capacity_bytes 0 to obtain the required capacity of proc_name.

For C++ methods, proc_name will be of the form ClassName::MethodName.

• Namespace::ClassName::MethodName if the class is a member of a namespace.
• Class1Name::ClassName::MethodName if the class is nested inside another class.
• (etc)

Use this function only to retrieve user-friendly procedure names for display.

• Use cs_pdg_basename() to obtain procedure name tokens to use with cross-referencing queries (cs_xref_def_iter_first(), cs_xref_occ_iter_first(), cs_xref_tuple).
• Use the name obtained by cs_pdg_procedure_name() for all other purposes.

In some cases the "friendly name" of a cs_pdg will differ markedly from its "procedure name". For example, for the following program:

void foo() asm("bar");

void foo() {
  printf("1234");
}

int main() {
  foo();
}

For C++ methods, proc_name will be of the form ClassName::MethodName.

• Namespace::ClassName::MethodName if the class is a member of a namespace.
• Class1Name::ClassName::MethodName if the class is nested inside another class.
• (etc)

Use this function only to retrieve user-friendly procedure names for display.

• Use cs_pdg_basename() to obtain procedure name tokens to use with cross-referencing queries (cs_xref_def_iter_first(), cs_xref_occ_iter_first(), cs_xref_tuple).
• Use the name obtained by cs_pdg_procedure_name() for all other purposes.

In some cases the "friendly name" of a cs_pdg will differ markedly from its "procedure name". For example, for the following program:

void foo() asm("bar");

void foo() {
  printf("1234");
}

int main() {
  foo();
}

• CS_TRUNCATED if bname is not large enough to hold the entire procedure basename (and so the basename is truncated),
• CS_SUCCESS otherwise.

The basename of a procedure is defined with respect to the friendly name (as obtained with cs_pdg_friendly_name()) as follows.

• If the friendly name contains one or more scope resolution operators "::", the basename is the method name component (i.e., rightmost token). For example, if the friendly name is ClassName::MethodName, the basename is MethodName.
• Otherwise, the basename is the same as the friendly name.

When cross-referencing queries (cs_xref_def_iter_first(), cs_xref_occ_iter_first()) are executed, the specified tokens are compared against procedure basenames, and query results (cs_xref_tuple) from procedure name matches contain a hash of this basename. Therefore, use cs_pdg_basename() to extract any procedure name that you want to use as a cross-referencing query token or compare against query results.

• Use cs_pdg_friendly_name() to retrieve user-friendly procedure names for display.
• Use the name obtained by cs_pdg_procedure_name() for all other purposes.

Call this function with bname NULL and capacity_bytes 0 to obtain the required capacity of name.

• CS_PDG_IS_UNDEFINED if pdg represents an undefined function.
• CS_SUCCESS on success.

• CS_SUCCESS on success.
• CS_OUT_OF_ELEMENTS if pdg has no local variables.

• CS_SUCCESS on success.
• CS_OUT_OF_ELEMENTS if iter has no more local variables to retrieve.

• CS_PDG_IS_UNDEFINED if pdg represents an undefined function.
• CS_TRUNCATED if call_sites is not large enough to hold the entire list of call sites (and so the list was truncated),
• CS_SUCCESS otherwise.

• CS_SUCCESS on success.
• CS_OUT_OF_ELEMENTS if pdg calls nothing.

• CS_SUCCESS on success.
• CS_OUT_OF_ELEMENTS if itr has no more call sites to retrieve.

• CS_SUCCESS on success.
• CS_OUT_OF_ELEMENTS if pdg is not called.

• CS_SUCCESS on success.
• CS_OUT_OF_ELEMENTS if itr has no more call sites to retrieve.

• CS_PDG_IS_UNDEFINED if pdg represents an undefined function.
• CS_SUCCESS on success.

Functions cs_pdg_formals_in() and cs_pdg_formals_in_as_list() retrieve different sets of vertices:

• cs_pdg_formals_in() retrieves the formal-in and the global-formal-in vertices of a cs_pdg.
• cs_pdg_formals_in_as_list() retrieves only the formal-in vertices of a cs_pdg.

• CS_PDG_IS_UNDEFINED if pdg represents an undefined function.
• CS_TRUNCATED if formal_ins is not large enough to hold the entire list of formal-in vertices (and so the list was truncated).
• CS_SUCCESS otherwise.

Functions cs_pdg_formals_in() and cs_pdg_formals_in_as_list() retrieve different sets of vertices:

• cs_pdg_formals_in() retrieves the formal-in and the global-formal-in vertices of a cs_pdg.
• cs_pdg_formals_in_as_list() retrieves only the formal-in vertices of a cs_pdg.

Call this function with formal_ins NULL and capacity_bytes 0 (zero) to determine the memory needed to hold the entire list.

• CS_SUCCESS on success.
• CS_ERROR_INVALID_ARGUMENT if fi_rank is negative.
• CS_ELEMENT_NOT_PRESENT if there is no formal-in parameter with rank fi_rank.
• CS_PDG_IS_UNDEFINED if pdg is undefined.

• CS_PDG_IS_UNDEFINED if pdg represents an undefined function.
• CS_SUCCESS on success.

• CS_PDG_IS_UNDEFINED if pdg represents an undefined function.
• CS_TRUNCATED if formal_outs is not large enough to hold the entire list of formal-out vertices (and so the list was truncated),
• CS_SUCCESS otherwise.

Call this function with formal_outs NULL and capacity_bytes 0 (zero) to determine the memory needed to hold the entire list.

• CS_PDG_IS_UNDEFINED if pdg represents an undefined function.
• CS_NO_SUCH_PDG_VERTEX if a cs_pdg_vertex with the given ID does not exist in pdg.
• CS_SUCCESS otherwise.

This function entails less user overhead than cs_pdg_procedure_name(), and so can be more convenient to use; the tradeoff is that the returned value is not under user control and has a constrained lifetime.

Do not modify or free the returned buffer.

This function entails less user overhead than cs_pdg_friendly_name(), and so can be more convenient to use; the tradeoff is that the returned value is not under user control and has a constrained lifetime.

Do not modify or free the returned buffer.

This function entails less user overhead than cs_pdg_basename(), and so can be more convenient to use; the tradeoff is that the returned value is not under user control and has a constrained lifetime.

If you do not need hash values to be stable across analyses, use cs_pdg_hash(): it has better performance.

• N<0 if pdg_a considered less than pdg_b
• N==0 if pdg_a considered equal to pdg_b
• N>0 if pdg_a considered greater than pdg_b

The comparison is stable in the following sense. Suppose there are two analyses A1 and A2 generated with exactly the same inputs (including identical analyzed code, underlying build commands and ordering, command line and configuration settings, increment order and contents). Let a1 and b1 be two cs_pdg values in A1, and a2 and b2 be the cs_pdg values in A2 that correspond to a1 and b1 respectively. Then cs_pdg_stable_compare(a1,b1)==cs_pdg_stable_compare(a2,b2).

If you do not need comparison results to be stable across different analyses, use cs_pdg_compare(): it has better performance.

• CS_SUCCESS on success,
• CS_TRUNCATED if the provided buffer is not large enough to hold the handle.

• CS_SUCCESS on success,
• CS_ERROR_INVALID_ARGUMENT if pdg_handle is not valid,
• CS_ERROR_IO_READ_PAST_EOF if the handle parameter is too short to be a valid handle.
• CS_ERROR_SDG_NOT_PRESENT if there is no current System Dependence Graph.
• CS_ELEMENT_NOT_PRESENT if there is no pdg matching pdg_handle.

• N<0 if pdg_a is considered less than pdg_b
• N==0 if pdg_a and pdg_b are the same cs_pdg
• N>0 if pdg_a is considered greater than pdg_b

For fixed cs_pdg values a and b in a single analysis, cs_pdg_compare(a,b) will always return the same value. However, cs_pdg_compare() will not necessarily return this same value when called on the corresponding cs_pdg values a' and b' in a different analysis. Function cs_pdg_stable_compare() provides some degree of stability across analyses, although restrictions still apply. If you do not need comparison results to be stable across different analyses, use cs_pdg_compare(): it has better performance.