> On Aug. 11, 2016, 4:32 p.m., Andrew Stitcher wrote: > > I've not thought through the buffer change properly yet. > > > > As pn_rwbytes_t obsoletes pn_buffer_memory_t it should replace it > > everywhere it is used (not too many places actually)
done - Alan ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/50989/#review145516 ----------------------------------------------------------- On Aug. 11, 2016, 5:26 p.m., Alan Conway wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/50989/ > ----------------------------------------------------------- > > (Updated Aug. 11, 2016, 5:26 p.m.) > > > Review request for qpid, Andrew Stitcher, Cliff Jansen, and Justin Ross. > > > Bugs: proton-1277 > https://issues.apache.org/jira/browse/proton-1277 > > > Repository: qpid-proton-git > > > Description > ------- > > Removed types pn_buf_t and pn_cbuf_t. > > Replaced pn_cbuf_t with existing public type pn_bytes_t. > Replaced pn_buf_t with new public type pn_rwbytes_t. > Replaced internal pn_buffer_memory_t with pn_rwbytes_t. > > PROTON-1277: pn_connection_engine to allow external buffers > > Extended the pn_connection_engine interface with use_read/write_buffer > functions > to allow an external buffer to be used. API allows user-controlled buffer > re-allocation. > > I tried for a "neater" interface but it didn't work out. For async write you > have to separate "get the data" and "all done with the data" so you can't > really > make it simpler than it is, and you need to be able to grow both buffers so > you > can't really do much to simplify the read side. For reads the transport moves > data around to accumulate complete frames, so you still have to ask "where is > the readable bit" even if you supplied the buffer. > > No implementation yet. I feel like we need a clearer idea of what we are > trying > to achieve before we commit. This proposal is backwards compatible which takes > some of the pressure off about deciding when to add it. > > > Diffs > ----- > > proton-c/bindings/cpp/src/io/connection_engine.cpp > f15b019b4637971509d9970024236ead10537778 > proton-c/bindings/go/src/qpid.apache.org/proton/engine.go > 5b9dbf069baf491c4b11459da7a27a438253a396 > proton-c/include/proton/connection_engine.h > 4fdeb85713e89256223efc5e93d27c4cb3e37a6b > proton-c/include/proton/types.h 3f2ebcbcca058fc9c5a2e4b0d7ef9f754bc069f0 > proton-c/src/buffer.h 94d48ac39fd1996208f75d990ac426c4a1a46710 > proton-c/src/buffer.c 64fa61fa4e7de9c4a293f27275444c493ff6f614 > proton-c/src/codec/codec.c ae77b0845c47b766a1870bc8bd96da2e46795a57 > proton-c/src/engine/connection_engine.c > 75a16ae7c09dcaa46525eee59a225868d5897483 > proton-c/src/transport/transport.c 1d7cc87fb8f863d8491a38d8d8546613d56064bb > proton-c/src/types.c 5433e193bbef31f1e426fe4f286dadac9d9dc64c > > Diff: https://reviews.apache.org/r/50989/diff/ > > > Testing > ------- > > ctest -E java > > > Thanks, > > Alan Conway > >
