This fixes access to indirect values within Java destination #1732
Conversation
|
This user does not have permission to start the build. Can one of the admins verify this patch and start the build? (admin please type: ok to test) |
|
@kira-syslogng ok to test |
|
Build SUCCESS, the tests were executed on test branch: master and test suite: functions |
|
|
||
| (*env)->ReleaseStringUTFChars(env, name, name_str); | ||
|
|
||
| if (value) | ||
| { | ||
| return (*env)->NewStringUTF(env, value); | ||
| gchar stripped_value[value_len + 1]; |
There was a problem hiding this comment.
I would avoid the allocation of string on the stack, the value_len could be big.
You could use the g_strndup instead, with that the \0 termination is guaranteed, although you must g_free it after thew NewStringUTF call.
Something like that:
gchar *stripped_value = g_strndup(value, value_len);
jstring java_string = (*env)->NewStringUTF(env, stripped_value);
g_free(stripped_value);
return java_string;Note: with this the #include <string.h> not needed.
There was a problem hiding this comment.
strndup() can be slow, and this might be a fastpath, although I am not sure how "fast" it is considering it is called from Java.
Anyway, there's a macro called APPEND_ZERO() which checks if the specific name-value pair is zero terminated already, and avoids the copying in that case. Still, that would allocate memory on the stack, which might be limited to couple of megabytes. but if that's the case, the remaining call sites of APPEND_ZERO would be affected as well.
terminated. stick to stack based allocation as macro is used in various modules which would also fail.
|
Build SUCCESS, the tests were executed on test branch: master and test suite: functions |
Related issue: #1714