bpo-40516: Silence GCC 9 warnings on MacOS Catalina #19925
Conversation
I'm not sure what to do here. Is GCC 9 on Catalina supposed to be supported? I tried it because I'm having issues with debug symbols with the latest version of clang and GDB but should I just build gcc 10? |
There was a problem hiding this comment.
See the similar discussion about this issue.
#13306 (comment)
|
A Python core developer has requested some changes be made to your pull request before we can consider merging it. If you could please address their requests along with any other requests in other reviews from core developers that would be appreciated. Once you have made the requested changes, please leave a comment on this pull request containing the phrase |
|
My question is do we know if these are legitimate warnings? If so, does the current Apple-provided clang give them in an otherwise identical build environment and configure options? If not, why not :) ? Also does gcc 10 report the same errors? Do these errors show up in other BSD-based systems, like FreeBSD? BTW @remilapeyre. you mentioned problems using gdb with clang. Have you tried clang's native debugger, lldb? |
|
Hi, thanks for your feedback, I had not found #13306 sorry about that. For more context about this issue, I'm usually using clang and lldb but I'm currently trying to diagnose a segmentation fault during SSL handshakes. To understand what's going on I would like to use the gdb support (https://devguide.python.org/gdb/). Using gdb on a binary built with clang used to work well but broke recently, I opened an issue at https://sourceware.org/bugzilla/show_bug.cgi?id=25850. I would actually be quite happy to continue using lldb as using gdb on MacOS is cumbersome since a certificate needs to be generated and gdb needs to signed with it but as as I know there is no way to get the same integration with the debugger.
The one in posix is specific to a quirk of MacOS 10.13, maybe clang knows about it? I did not check if the one in nis is present on BSD.
I did not try as I did not know it existed, it does not appear at https://gcc.gnu.org/releases.html. My main goal here is to be able to use https://devguide.python.org/gdb/ on MacOS, I see multiple way to do that:
Maybe the the third one is the best as a Python / LLDB could be useful to others? |
|
Ping, @ned-deily, @corona10? |
I don't have further suggestions as it's not really my area of expertise. Perhaps @ronaldoussoren might. |
|
For what it's worth I started working on a |
| @@ -213,7 +213,9 @@ nis_cat (PyObject *self, PyObject *args, PyObject *kwdict) | |||
| dict = PyDict_New (); | |||
| if (dict == NULL) | |||
| return NULL; | |||
| #pragma GCC diagnostic ignored "-Wcast-function-type" | |||
There was a problem hiding this comment.
This needs a preprocessor guard to ensure the pragma is only used with compatible compilers.
| cb.foreach = (foreachfunc)nis_foreach; | ||
| #pragma GCC diagnostic pop |
| @@ -3392,12 +3392,15 @@ os_chown_impl(PyObject *module, path_t *path, uid_t uid, gid_t gid, | |||
| * This is for Mac OS X 10.3, which doesn't have lchown. | |||
There was a problem hiding this comment.
PR #22855 will fix the warnings in posixmodule.c (as a side effect of more interesting changes).
|
Note the nis module is deprecated in 3.11 and set for removal in 3.13. See PEP 594 – Removing dead batteries from the standard library and #91217. |
|
This PR has been "awaiting changes" for over a year, and now has a merge conflict. I am therefore closing this PR. @remilapeyre, if you're still interested in working on this issue, feel free to ping me and I'll happily reopen the PR. However, do note that the |
https://bugs.python.org/issue40516