So _ZN5boost9unit_testL16INV_TEST_UNIT_IDE is the only label that's used through a call opcode, and so the original parser decides to hide the other labels, but somehow keep showing the internal data.
I think the showing of the data while not "belonging" to the first label is probably wrong. So the original asm-parser has a bug here, I think.
(I reintroduced fake usage through the .type line in the c++ parser to show the rest of the labels - to try and get some kind of compatibility going, but it's not easy to be compatible while trying to stay correct)
So _ZN5boost9unit_testL16INV_TEST_UNIT_IDE is the only label that's used through a call opcode, and so the original parser decides to hide the other labels, but somehow keep showing the internal data.
I think the showing of the data while not "belonging" to the first label is probably wrong. So the original asm-parser has a bug here, I think.
(I reintroduced fake usage through the .type line in the c++ parser to show the rest of the labels - to try and get some kind of compatibility going, but it's not easy to be compatible while trying to stay correct)