RFR: 8362889: [GCC static analyzer] leak in libstringPlatformChars.c [v2]

[David Holmes]

On Mon, 21 Jul 2025 15:38:36 GMT, Matthias Baesken <mbaesken at openjdk.org> wrote:

>> test/jdk/java/lang/String/nativeEncoding/libstringPlatformChars.c line 74:
>> 
>>> 72:     (*env)->ReleasePrimitiveArrayCritical(env, bytes, (void*)jbytes, 0);
>>> 73: 
>>> 74:     jstring res = JNU_NewStringPlatform(env, str);
>> 
>> At line 66: where it returns null, it seem like it should also be freeing `str`.  (The analyzer didn't catch that?)
>
> It catches that too 
> 
> /jdk/test/jdk/java/lang/String/nativeEncoding/libstringPlatformChars.c:65:8: warning: leak of 'str' [CWE-401] [-Wanalyzer-malloc-leak]
>    65 |     if (jbytes == NULL) {
> 
> 
> and also
> 
> 
> /jdk/test/jdk/java/lang/String/nativeEncoding/libstringPlatformChars.c:69:16: warning: dereference of possibly-NULL 'str' [CWE-690] [-Wanalyzer-possible-null-dereference]
>    69 |         str[i] = (char)jbytes[i];
>       |         ~~~~~~~^~~~~~~~~~~~~~~~~
> 
> 
> I just overlooked it because a  log of a full build  with '-fanalyzer'  enabled contains quite a lot of output  (the event explanations of every finding are rather long).

Seems better to defer the `malloc` until after you know `jbytes` is not null; and also check the `malloc` result for null.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/26415#discussion_r2221292734