Bob, As part of fixing this, we made a function in the nsssysinit.c non-static (so it could be tested). I've added a .def file to the nsssysinit.so build so that we don't pollute the ABI. This does result in ABI changes (from the ABI compatibility check): +Function symbols changes summary: 2 Removed, 0 Added function symbols not referenced by debug info +Variable symbols changes summary: 0 Removed, 0 Added variable symbol not referenced by debug info + +2 Removed function symbols not referenced by debug info: + + _fini + _init + It seems like this is very likely OK: https://sourceware.org/ml/binutils/2018-05/msg00012.html However, I just want to confirm that hiding the _init and _fini symbols isn't an issue in your experience? We hide those symbols for other libraries, and I don't see any special behavior here that warrants caution, but I thought that I'd double check.
Bug 818686 Comment 16 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
Bob, As part of fixing this, we made a function in the nsssysinit.c non-static (so it could be tested). I've added a .def file to the nsssysinit.so build so that we don't pollute the ABI. This does result in ABI changes (from the ABI compatibility check): ``` +Function symbols changes summary: 2 Removed, 0 Added function symbols not referenced by debug info +Variable symbols changes summary: 0 Removed, 0 Added variable symbol not referenced by debug info + +2 Removed function symbols not referenced by debug info: + + _fini + _init + ``` It seems like this is very likely OK: https://sourceware.org/ml/binutils/2018-05/msg00012.html However, I just want to confirm that hiding the _init and _fini symbols isn't an issue in your experience? We hide those symbols for other libraries, and I don't see any special behavior here that warrants caution, but I thought that I'd double check.