Stage1 of clang compiler warning patches. #108
Conversation
ethoms
changed the title
| @@ -94,7 +94,7 @@ - (id) initWithName: (NSString *) newName | |||
| { | |||
| if (![newDisplayName length]) | |||
| newDisplayName = newName; | |||
| ASSIGN (displayName, newDisplayName); | |||
| ASSIGN (displayName, [newDisplayName mutableCopy]); | |||
|
The reason I used mutableCopy is to (I thought) safely get a mutable instance. Since the compiler warns that it is the wrong data type being received. E.g. an NSMutableString is expected, but a NSString is given. Initially I just did a simple cast (NSMutableString *), but after reading that it's better to cast properly using mutableCopy I changed the patches. Is it safe to use a simple cast (NSMutableString *) etc.? Or do have any better ideas to get the data types corrected so to avoid compiler warnings? I'm not sure what you mean by leaking? I guess it should be passed by reference, not by copy? |
|
mutableCopy creates a new instance with a retain count of 1, so it's not autoreleased and thus, leaks the newly created instance. Just cast it, it'll be good! |
|
OK, noted. So I'll just change these and add another commit to this PR? |
|
Actually, I'll just use the github editor. I'll post back after build testing on FreeBSD and RHEL 5.11. |
|
OK, all done. Converted all uses of mutableCopy to a cast in this pull request. |
|
NOTE: on line 168 of UI/Scheduler/UIxCalDayView.m there is a use of mutableCopy. It's not one of my changes, but maybe you want to check it out. |
|
Yep, but notice it's returned as autoreleased at the end of the method. |
|
Ah, OK, I didn't read on. |
|
I've successfully build tested on both FreeBSD 10.1 and RHEL 5.11 after correcting the mutableCopy issue. Is there any other issue with this PR? |
|
I don't think so but I'll wait until our nightly builds + integration tests pass in a couple of hours before merging new stuff. Thanks! |
Stage1 of clang compiler warning patches.
|
Can you please adapt this PR so we can merge it also in the v2 branch? I guess it'll be mostly code remove. Open the new PR just for the v2 branch - since the patch was merged in master. |
|
No problem, I'll make a new PR to the v2 branch as soon as I've done looking into the other issues. |
Stage 1:
Fairly rudimentary fixes. Mostly fixing data type and casting inconsistencies.
Fixes for:
(i) Data type casting
(ii) Data type initialization
(iii) String formatting (data types)
(iv) Missing curly brackets in ambiguous if statements / for loops.