On Tue, Jan 6, 2015 at 11:09 AM, Michael Paquier
<michael.paquier@gmail.com> wrote:
> On Mon, Jan 5, 2015 at 10:29 PM, Fujii Masao <masao.fujii@gmail.com> wrote:
>> On Sun, Dec 28, 2014 at 10:57 PM, Michael Paquier wrote:
>> The patch 1 cannot be applied to the master successfully because of
>> recent change.
> Yes, that's caused by ccb161b. Attached are rebased versions.
>
>>> - The real stuff comes with patch 2, that implements the removal of
>>> PGLZ_Header, changing the APIs of compression and decompression to pglz to
>>> not have anymore toast metadata, this metadata being now localized in
>>> tuptoaster.c. Note that this patch protects the on-disk format (tested with
>>> pg_upgrade from 9.4 to a patched HEAD server). Here is how the APIs of
>>> compression and decompression look like with this patch, simply performing
>>> operations from a source to a destination:
>>> extern int32 pglz_compress(const char *source, int32 slen, char *dest,
>>> const PGLZ_Strategy *strategy);
>>> extern int32 pglz_decompress(const char *source, char *dest,
>>> int32 compressed_size, int32 raw_size);
>>> The return value of those functions is the number of bytes written in the
>>> destination buffer, and 0 if operation failed.
>>
>> So it's guaranteed that 0 is never returned in success case? I'm not sure
>> if that case can really happen, though.
> This is an inspiration from lz4 APIs. Wouldn't it be buggy for a
> compression algorithm to return a size of 0 bytes as compressed or
> decompressed length btw? We could as well make it return a negative
> value when a failure occurs if you feel more comfortable with it.
I feel that's better. Attached is the updated version of the patch.
I changed the pg_lzcompress and pg_lzdecompress so that they return -1
when failure happens. Also I applied some cosmetic changes to the patch
(e.g., shorten the long name of the newly-added macros).
Barring any objection, I will commit this.
Regards,
--
Fujii Masao