mk-lib1521.pl: fix the long return code check - #15372
Conversation
|
Hm, this isn't easy to fix in a nice and clean way... |
|
The reason the existing macro is bad to use in lib1521 is that it effectively prevents us from setting a break-point in lib1521.c with a debugger inside that macro, and it also breaks single-stepping. This makes debugging super annoying and downright complicated. |
|
GGGGGGGGGG |
|
I'll drop the deprecation part from this PR and come back to it later. |
It worked mostly by accident since it checked the variable from the previous setopt invoke.
|
Analysis of PR #15372 at 92e6b3fd: Test 1193 failed, which has NOT been flaky recently, so there could be a real issue in this PR. Test 1196 failed, which has NOT been flaky recently, so there could be a real issue in this PR. Test 2201 failed, which has NOT been flaky recently, so there could be a real issue in this PR. Test 2205 failed, which has NOT been flaky recently, so there could be a real issue in this PR. Test 1452 failed, which has NOT been flaky recently, so there could be a real issue in this PR. Generated by Testclutch |
It worked mostly by accident since it checked the variable from the previous setopt invoke. Closes curl#15372
2 participants
It worked mostly by accident since it checked the variable from the previous setopt invoke.
Bonus: this disables deprecation differently. Using the CURL_IGNORE_DEPRECATION() style works, but it makes the code infuriatingly annoying to work with when single-stepping in a debugger.