| Summary: | String::createCFString() returns a non-retained empty string | ||||||
|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | Alexey Proskuryakov <ap> | ||||
| Component: | Platform | Assignee: | Alexey Proskuryakov <ap> | ||||
| Status: | RESOLVED FIXED | ||||||
| Severity: | Minor | ||||||
| Priority: | P4 | ||||||
| Version: | 528+ (Nightly build) | ||||||
| Hardware: | Mac | ||||||
| OS: | OS X 10.5 | ||||||
| Attachments: |
|
||||||
Created attachment 35051 [details]
proposed patch
Comment on attachment 35051 [details] proposed patch > + * platform/text/cf/StringCF.cpp: (WebCore::String::createCFString): Rather than retain > + CFSTR result, we can just return a new string, this doesn't seem to be a hot code path Even though it's not a hot code path it seems like we should avoid the extra memory allocation that CFStringCreateWithCharacters does compared to CFSTR. Comment on attachment 35051 [details]
proposed patch
If either solution is fine speed-wise, then why not call CFRetain to save memory in case a lot of these strings are kept around?
Committed revision 47449.
> If either solution is fine speed-wise, then why not call CFRetain to save
> memory in case a lot of these strings are kept around?
OK, changed. /me doesn't like to share objects across threads.
|
If String is null, createCFString() returns CFSTR(""). According to CFString.h, 'CFSTR(), not being a "Copy" or "Create" function, does not return a new reference for you. So, you should not release the return value.' I am not aware of any practical issues caused by this, but we should probably comply.