[
https://issues.apache.org/jira/browse/TS-14?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12884525#action_12884525
]
Leif Hedstrom commented on TS-14:
---------------------------------
In addition to INKfread() mentioned above, these are also candidates for
getting their prototypes fixed:
inkapi void *_INKmalloc(unsigned int size, const char *path);
inkapi void *_INKrealloc(void *ptr, unsigned int size, const char *path);
inkapi char *_INKstrdup(const char *str, int length, const char *path);
inkapi void _INKfree(void *ptr);
inkapi int INKfwrite(INKFile filep, const void *buf, int length);
inkapi char *INKfgets(INKFile filep, char *buf, int length);
inkapi INKVIO INKVConnRead(INKVConn connp, INKCont contp, INKIOBuffer bufp,
int nbytes);
inkapi INKVIO INKVConnWrite(INKVConn connp, INKCont contp, INKIOBufferReader
readerp, int nbytes);
inkapi INKReturnCode INKVConnCacheObjectSizeGet(INKVConn connp, int
*obj_size);
inkapi int INKVIONBytesGet(INKVIO viop);
inkapi INKReturnCode INKVIONBytesSet(INKVIO viop, int nbytes);
inkapi int INKVIONDoneGet(INKVIO viop);
inkapi INKReturnCode INKVIONDoneSet(INKVIO viop, int ndone);
inkapi int INKVIONTodoGet(INKVIO viop);
inkapi INKReturnCode INKIOBufferWaterMarkGet(INKIOBuffer bufp, int
*water_mark);
inkapi INKReturnCode INKIOBufferWaterMarkSet(INKIOBuffer bufp, int
water_mark);
inkapi int INKIOBufferCopy(INKIOBuffer bufp, INKIOBufferReader readerp, int
length, int offset);
inkapi int INKIOBufferWrite(INKIOBuffer bufp, const void *buf, int length);
inkapi INKReturnCode INKIOBufferProduce(INKIOBuffer bufp, int nbytes);
inkapi const char *INKIOBufferBlockReadStart(INKIOBufferBlock blockp,
INKIOBufferReader readerp, int *avail);
inkapi char *INKIOBufferBlockWriteStart(INKIOBufferBlock blockp, int *avail);
inkapi int INKIOBufferReaderAvail(INKIOBufferReader readerp);
inkapi int INKAIONBytesGet(void* data);
There were also some deprecated APIs that are 'wrong', but we should not modify
those (ever).
> InkAPI is not 64-bit clean
> --------------------------
>
> Key: TS-14
> URL: https://issues.apache.org/jira/browse/TS-14
> Project: Traffic Server
> Issue Type: Bug
> Components: InkAPI
> Affects Versions: 2.0.0a
> Environment: x86_64 linux
> Reporter: John Plevyak
> Fix For: 2.1.2
>
>
> This is different than the 64-bit port in that it effects the public API. In
> particular there are functions which use 'int' where they should be using
> other types for things like sizes which could exceed 2^32, the 'int' size
> even on 64-bit linux. Specifically:
> api/include/InkAPI.h: inkapi int INKfread(INKFile filep, void *buf, int
> length);
> But a full review of the API should be conducted and the types updated. In
> this case inku64, ink_off_t or most likely size_t would
> be more appropriate. All of these are 64-bit on 64-bit linux.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
