aboutsummaryrefslogtreecommitdiff
path: root/src
Commit message (Collapse)AuthorAge
...
* Tests: remove the legacy wss transport test and framework support for Convey ↵Garrett D'Amore2025-04-27
| | | | | | | transport tests This has been needed for some time; the convey framework is not reliable or debuggable, and will ultimately be removed. Only the http client test remains using it.
* tests: Add wss_test following pattern of TLS test.Garrett D'Amore2025-04-27
| | | | This replaces the convey style test.
* Fix typos in TLS test names.Garrett D'Amore2025-04-27
|
* TLS: break up the TLS layer a bit to refactor for DTLS.Garrett D'Amore2025-04-27
| | | | | | | | | | | This allows us to break the assumption that the bottom half is TCP, or even an nng_stream, since the DTLS layer will use a totally different layer. Only nng_stream neeeds to support dial and listen. Also: UDP: Make the sockaddr arguments to open const. Also: Align the IPv6 address in the sockaddr (this allows for efficient 64-bit or even 128-bit operations on these values.)
* fix aio_reap_list use of nni_aio_free() pointer to incorrect function type ↵shikokuchuo2025-04-27
| | | | UBSAN warning
* Make pointers to data that is not modified const. (#2117)Søren Holm2025-04-27
| | | | | | This allow to pass constrant string i particular and constant data in general to nng functions. Co-authored-by: Garrett D'Amore <garrett@damore.org>
* * FIX [HTTP] fix a blocking case when freeing http clientJaylin2025-04-27
| | | basically when aio timeout before the connaio of dialer, and users try to free the http_client obj will end in infinite blocking at nni_http_client_fini. Possibly at nni_aio_free(c->aio); or nng_stream_dialer_free(c->dialer); Both racing case is due to the ingnored aio aborting here. Because the aio_begin is called before it is put into the nni_list. I assume you shall abort it no matter if it is in the dialing list.
* Set server name for TLS test.Garrett D'Amore2025-04-19
|
* Fix typo in NNG_TLS_MAX_RECV_SIZEGarrett D'Amore2025-04-13
|
* fix source formattingshikokuchuo2025-03-16
|
* req protocol: free msg early if no retriesshikokuchuo2025-03-16
|
* posix_pollq_epoll.c:add EPOLLHUP flag to aovid code error continues ↵wangchen2025-03-15
| | | | | | | | executing in tcp_cb the reason of the modification refers to https://github.com/nanomsg/nng/issues/2100 Signed-off-by: wangchen <wangchen41@xiaomi.com>
* Windows: Fix regression where no connections after 5 seconds can occurGarrett D'Amore2025-03-09
|
* Windows: Bring back hEvent -- needed for real IPC.Garrett D'Amore2025-02-18
|
* fixes #2092 nng_close may hang on Windowsv2.0.0-alpha.4Garrett D'Amore2025-01-26
|
* tests: add some tcp dialer local address testsGarrett D'Amore2025-01-14
|
* pipe: nng_pipe_notify should use nng_errGarrett D'Amore2025-01-14
|
* socketpair: use nng_errGarrett D'Amore2025-01-14
|
* init: use nng_errGarrett D'Amore2025-01-14
|
* aio: more nng_err updatesGarrett D'Amore2025-01-14
|
* api: pipes should use nng_errGarrett D'Amore2025-01-14
|
* tests: start using nng_err in NUTSGarrett D'Amore2025-01-12
|
* http: use nng_err throughoutGarrett D'Amore2025-01-12
|
* http: status is an enumGarrett D'Amore2025-01-12
| | | | | This is slightly less efficient, but it provides for better debugging and type safety.
* api: extend usage of nng_errGarrett D'Amore2025-01-12
| | | | | | This replaces the int, and we will expand this further, as this makes it clear that the int is actually an error code and helps in debuggers that can provide symbolic values.
* http: fix mishandling of very long headers or URIs, and mishandling of unicodeGarrett D'Amore2025-01-12
| | | | Also, nng_err is now a distinct type which might be nicer in debuggers.
* http: remove unused nng_http_handler_get_dataGarrett D'Amore2025-01-12
| | | | The data is now passed directly to the handler function.
* http: server error handling improvements and testsGarrett D'Amore2025-01-12
| | | | | | | We want to consume the request properly on an error, so that we can give a reasonable response. We were prematurely closing the connection for certain failure modes. We still have to fix overly long URIs and headers, but thats next!
* http: use common canonify at request parse timeGarrett D'Amore2025-01-12
|
* http: move connection properties to http_conn, use static Location for redirectGarrett D'Amore2025-01-12
|
* http: status and reason fixes (make it match docs)Garrett D'Amore2025-01-12
|
* http: improve buffer reuse for heeaders, and discard unused bodiesGarrett D'Amore2025-01-11
| | | | | | | | | | | | | The body content not being consumed was leading to misparses, where we consumed body data as if it were a request. When mixed with proxies this could lead to a security problem where the following request content submitted from a different client winds up as stolen request body content. This also ensures we actually deliver errors to clients without prematurely closing the connection. (There are still problems where the connection may be closed prematurely for an overlarge header.)
* http: remove public access to req / res structuresGarrett D'Amore2025-01-11
|
* http: drop unused internal functionsGarrett D'Amore2025-01-10
|
* http: drop exclusive tree modeGarrett D'Amore2025-01-10
| | | | | Nothing really needs it -- we kept it in 1.0 to preserve semantics, but there is no requirement for semantic preservation in 2.0.
* http test: test HEAD method in serverGarrett D'Amore2025-01-10
|
* http: The big HTTP API refactoring of January 2025.v2.0.0-alpha.3http-client-transGarrett D'Amore2025-01-09
| | | | | | | | | | | | | | | | | | | | | | This represents a major change in the HTTP code base, consisting of a complete revamp of the HTTP API. The changes here are too numerous to mention, but the end result should be a vastly simpler API for both server and client applications. Many needless allocations were removed by providing fixed buffers for various parameters and headers when possible. A few bugs were fixed. Most especially we have fixed some bugs around very large URIs and headers, and we have also addressed conformance bugs to more closely conform to RFCs 9110 and 9112. As part of this work, the APIs for WebSockets changed slightly as well. In particular the properties available for accessing headers have changed. There is still documentation conversion work to do, and additional functionality (such as proper support for chunked transfers), but this is a big step in the right direction.
* Update CMakeLists.txtVũ Quang Thịnh2025-01-09
| | | Fix typo of filename. Can not find filename
* http: server callback API simplifiedGarrett D'Amore2025-01-06
| | | | | | | | | | This simplified API lets callbacks obtain the response from the connection objection directly, and does not require the aio to carry it as a parameter. Further, the request and response are both stored inline in the connection, reducing allocations. This is at present only for the server; the client will get a similar set of changes.
* http: changing transaction API to inline req and res structuresGarrett D'Amore2025-01-05
| | | | | This is a step towards simplifying this API and ultimately simplifying the HTTP callback API used for the server side.
* posix: move tcpdialer definition out of shared headerGarrett D'Amore2025-01-05
|
* platform: remove reader/writer locksGarrett D'Amore2025-01-05
| | | | | | The only thing using this was the transport lookups, but as those transports are now fully initialized in nng_init, we no longer need to lock that at all.
* platform: eliminate NNI_CV_INITIALIZER altogetherGarrett D'Amore2025-01-05
|
* reap: use sys init and fini instead of NNI_CV_INITIALIZER.Garrett D'Amore2025-01-05
| | | | The CV_INITIALIZER is error prone, as it cannot use cv_until.
* pthreads: avoid double indirection for cv mutexGarrett D'Amore2025-01-05
|
* Liberally apply some UWYI (use what you include) to header filesGarrett D'Amore2025-01-05
|
* args: Convert nng_opts_parse into a header only library using `nng_args_parse`.Garrett D'Amore2025-01-04
| | | | | | | | The API is identical, except that some names have changed, and this is now a header library in `nng/args.h` - so the core library does not need to carry this code in binaries. Being a header library also means it is not necessary to link against NNG, and it does not include any parts of NNG; it only depends on a standard C99 or C11 environment.
* api: Promote idhash supplemental API to coreGarrett D'Amore2025-01-04
|
* api: remove the supplemental platform.h headerGarrett D'Amore2025-01-04
|
* tests: make NUTS_PASS use TEST_ASSERTGarrett D'Amore2025-01-04
| | | | | This will ensure that we fast fail if a test cannot complete, rather than muddling on and reporting failures elsewhere.
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2026-01-21 12:05:14 +0000