gh-76961: Possible fix for buildbot failures in test_pep3118 #101587
Conversation
|
🤖 New build scheduled with the buildbot fleet by @mdickinson for commit 3586c09 🤖 If you want to schedule another build, you need to add the |
|
Argh. Windows wants |
|
Added a couple of commits to silence warnings. |
|
🤖 New build scheduled with the buildbot fleet by @mdickinson for commit f273e74 🤖 If you want to schedule another build, you need to add the |
| @@ -355,8 +355,8 @@ _ctypes_alloc_format_padding(const char *prefix, Py_ssize_t padding) | |||
| return _ctypes_alloc_format_string(prefix, "x"); | |||
| } | |||
|
|
|||
| int ret = PyOS_snprintf(buf, sizeof(buf), "%zdx", padding); | |||
| assert(0 <= ret && ret < sizeof(buf)); | |||
| int ret = PyOS_snprintf(buf, sizeof(buf), "%zdx", padding); (void)ret; | |||
There was a problem hiding this comment.
FWIW, the (void)ret; trick comes from the LLVM Coding Standard: https://llvm.org/docs/CodingStandards.html#assert-liberally
|
🤖 New build scheduled with the buildbot fleet by @mdickinson for commit 59144e7 🤖 If you want to schedule another build, you need to add the |
|
One buildbot failure, for buildbot/aarch64 RHEL8 Refleaks PR, which looks unrelated ( @eric-wieser This seems to be working; do you have time to give this a second-pass review? |
This PR is an attempt to fix the buildbot failures introduced by my merge of #5561, by restricting the relevant tests to something that should work on both 32-bit and 64-bit platforms.