Handle using nil as argument for andReturn in ObjC++ code. #404
Conversation
dmaclach
force-pushed
the
realNil
branch
2 times, most recently
from
de60c8b
to
23c83e7
Compare
dmaclach
force-pushed
the
realNil
branch
2 times, most recently
from
ed508f1
to
18214dd
Compare
|
Anyone have any comments on this one? I think it's a no-brainer. |
|
Oops.. looks like Erik is looking at it as I type :) Apologies. |
Fix for erikdoe#403. This improves the special case handling of passing nil in return values to handle C++ cases that previously weren't working. Adds tests for C++11 and C++98 variants. To test fully tests need to be executed in both 32bit and 64 bit modes.
|
Erik, do you have any plans you could share with regards to timing for OCMock 3.7, 3.8 etc? My pile of patches on my end is becoming a little scary :) |
|
So, the plan was/is to get out 3.7 very soon. I got sidetracked by the PR's regarding Swift Package Manager, but I have now concluded that they won't be in 3.7. So, there's only two open PR's which seem reasonably straight-forward. I was also debating with myself whether #398 should be in, but I haven't read all the comments over there in detail. |
|
We don't want #398 in its current state IMHO. I would think about it after
#414 was available and see if it is still wanted/needed.
…On Mon, Jul 13, 2020 at 2:04 PM Erik Doernenburg ***@***.***> wrote:
So, the plan was/is to get out 3.7 very soon. I got sidetracked by the
PR's regarding Swift Package Manager, but I have now concluded that they
won't be in 3.7. So, there's only two open PR's which seem reasonably
straight-forward. I was also debating with myself whether #398
<#398> should be in, but I haven't
read all the comments over there in detail.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#404 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AACOFSPTASD2U6W3JT2OSGLR3NZFHANCNFSM4MKFOT2A>
.
|
| @@ -547,7 +551,7 @@ | |||
| 2FA28006D043CBDBBAEF6E3F /* OCMMacroState.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = OCMMacroState.h; sourceTree = "<group>"; }; | |||
| 2FA280987F4EA8A4D79000D0 /* OCMMacroState.m */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.objc; path = OCMMacroState.m; sourceTree = "<group>"; }; | |||
| 2FA280EB5E8CDEEAE76861F7 /* OCMNonRetainingObjectReturnValueProvider.m */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.objc; path = OCMNonRetainingObjectReturnValueProvider.m; sourceTree = "<group>"; }; | |||
| 2FA2813F93050582D83E1499 /* OCMockObjectRuntimeTests.m */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.objc; path = OCMockObjectRuntimeTests.m; sourceTree = "<group>"; }; | |||
| 2FA2813F93050582D83E1499 /* OCMockObjectRuntimeTests.m */ = {isa = PBXFileReference; explicitFileType = sourcecode.cpp.objcpp; fileEncoding = 4; path = OCMockObjectRuntimeTests.m; sourceTree = "<group>"; }; | |||
There was a problem hiding this comment.
This change breaks the build for me. When I change the file type back to Objective-C it works. The error is:
/Users/erik/Projects/OCMock/Repositories/ocmock/Source/OCMockTests/OCMockObjectRuntimeTests.m:219:54: error: cannot initialize a parameter of type 'void *' with an lvalue of type 'const char [4]'
XCTAssertNoThrow([[myMock expect] aSpecialMethod:"foo"], @"Should not complain about method with type qualifiers.");
There was a problem hiding this comment.
ahh.. that is probably an accidental mod when I changed the file to objc++ before I broke the tests out into their own files. No need to pull that across. Apologies.
| XCTAssertTrue([boxed isMethodReturnType:type2 compatibleWithValueType:type3]); | ||
| XCTAssertTrue([boxed isMethodReturnType:type3 compatibleWithValueType:type1]); | ||
| XCTAssertTrue([boxed isMethodReturnType:type3 compatibleWithValueType:type2]); | ||
| XCTAssertTrue([boxed isMethodReturnType:type1 compatibleWithValueType:type2 value:&value valueSize:valueSize]); |
There was a problem hiding this comment.
I changed these tests to pass NULL and 0 for the last two arguments. Couldn't see a reason why not, and it saved creating the value in the test.
Fix for #403.
This improves the special case handling of passing nil in return values to handle
C++ cases that previously weren't working.
Adds tests for C++11 and C++98 variants. To test fully tests need to be executed in both
32bit and 64 bit modes.