Add Catecholamine (CA), a tool for breaking code objects out of fat binaries created via LPL. As with LPL, --help provides usage details. Fixed LPL's implementation of path_to_self which was a nasty infinite loop iff for whatever reason readlink would always fail, orthogonally to the string's size. Link against pthread for LPL, which uses call_once and once_flag (this could, under certain circumstances, yield an arcane "Unknown error=-1" exception.

LPL_CA/ca.cpp

@@ -0,0 +1,49 @@
+#include "ca.hpp"
+
+#include <cstdlib>
+#include <exception>
+#include <iostream>
+#include <stdexcept>
+#include <string>
+#include <vector>
+
+using namespace clara;
+using namespace hip_impl;
+using namespace std;
+
+int main(int argc, char** argv)
+{
+    try {
+        bool help = false;
+        vector<string> inputs;
+        string targets;
+
+        auto cmd = cmdline_parser(help, inputs, targets);
+
+        const auto r = cmd.parse(Args{argc, argv});
+
+        if (!r) throw runtime_error{r.errorMessage()};
+
+        if (help) cout << cmd << endl;
+        else {
+            if (inputs.empty()) throw runtime_error{"No inputs specified."};
+
+            validate_inputs(inputs);
+
+            auto tmp = tokenize_targets(targets);
+            if (tmp.empty()) {
+                tmp.assign(amdgpu_targets().cbegin(), amdgpu_targets().cend());
+            }
+            else validate_targets(tmp);
+
+            extract_code_objects(inputs, tmp);
+        }
+    }
+    catch (const exception& ex) {
+        cerr << ex.what() << endl;
+
+        return EXIT_FAILURE;
+    }
+
+    return EXIT_SUCCESS;
+}