aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAge
* Work on endpoints. More C99 & type cleanups.Garrett D'Amore2016-12-22
|
* Inline locks (fewer allocs), simpler absolute times for wakeups. nn_sock_recv.Garrett D'Amore2016-12-22
|
* Synchronization enhancements - inproc & msgqueue. Absolute waits...Garrett D'Amore2016-12-22
|
* Start of work to inline mutexes and condition variables.Garrett D'Amore2016-12-22
|
* Use C99 structure initializers FTW. Various other changes.Garrett D'Amore2016-12-22
|
* Logic for socket shutdown, cleanup, and draining figured out.Garrett D'Amore2016-12-22
| | | | | There's work to do still, but I've left clear indications of the design in comments. Some ugly mysteries are now solved.
* Older Darwin lacks clock_gettime(); we need poll.h to use poll().Garrett D'Amore2016-12-21
|
* Ditch our own snprintf/vsnprintf (C99). Symbol naming fixes for inproc.Garrett D'Amore2016-12-21
|
* Use C99.Garrett D'Amore2016-12-21
|
* Provide Sublime Text settings, and a lot more detail about coding style, etc.Garrett D'Amore2016-12-21
|
* Use trusty -- hopefully get a newer version of uncrustify.Garrett D'Amore2016-12-21
|
* Add a README for the etc/ directory..Garrett D'Amore2016-12-21
|
* Uncrustify clean now.Garrett D'Amore2016-12-21
|
* Another attempt at uncrustify checks.Garrett D'Amore2016-12-21
|
* More tweaks to script.Garrett D'Amore2016-12-21
|
* fix compile errors.Garrett D'Amore2016-12-21
|
* Travis stuff for uncrustify.Garrett D'Amore2016-12-21
|
* Uncrustify configuration, and shorter copyright banners, plus reformatGarrett D'Amore2016-12-21
| | | | code with uncrustify. (Minor adjustments.) No more arguments!
* Updates to reflect new external convey framework.Garrett D'Amore2016-12-20
|
* Fix the definition of Main() in convey.hGarrett D'Amore2016-12-19
|
* fix windows compilation bug (colon instead of semicolon)Garrett D'Amore2016-12-19
|
* test_main in terms of test_main_group and test_group.Garrett D'Amore2016-12-18
|
* Minor bug fixes to test.h and co.Garrett D'Amore2016-12-18
|
* Nicer convey API - more closely mirrors the Go one.Garrett D'Amore2016-12-18
|
* Remove stale clean up handler bits. Attempt to provide Win32 API support.Garrett D'Amore2016-12-18
|
* Added support for test_reset(), better than cleanup or other hacks.Garrett D'Amore2016-12-18
|
* Suppress exposure of contexts to code under test.Garrett D'Amore2016-12-18
| | | | Fix timing display, and other output bugs.
* Add test groups, free memory used by tests.Garrett D'Amore2016-12-18
|
* Added linked list test suite.Garrett D'Amore2016-12-18
|
* So "test" target now does something useful.Garrett D'Amore2016-12-17
|
* Test framework is dev-complete. More functionality may be added later.Garrett D'Amore2016-12-17
| | | | Also, the current code is not Win32 safe.
* Nicer test framework -- more internals in the .C, nicer and more correctGarrett D'Amore2016-12-17
| | | | test result output.
* Early test framework, modeled on GoConvey.Garrett D'Amore2016-12-15
|
* Endpoint structure defined.Garrett D'Amore2016-12-14
|
* nni_socket_add_pipe and nni_socket_remove_pipe implementation.Garrett D'Amore2016-12-14
|
* More robust platform definition support.Garrett D'Amore2016-12-14
| | | | | | | The idea is that someday it will be possible to just concatenate the entire set of source files into a single giant source file, for systems that want to work this way. As a result, the build system now compiles every file, although some of them will not have any definitions.
* Socket sendfilter called.Garrett D'Amore2016-12-14
|
* Better separate public vs. private names.Garrett D'Amore2016-12-14
|
* Oops... Pipe list initialization was not complete.Garrett D'Amore2016-12-14
|
* Socket pipes list.Garrett D'Amore2016-12-14
|
* nn_socket_create() implemented.Garrett D'Amore2016-12-14
|
* Some endpoint work started.Garrett D'Amore2016-12-14
|
* Protocols are 16-bit.Garrett D'Amore2016-12-14
|
* Add ep_destroy implemenation to inproc.Garrett D'Amore2016-12-14
|
* Initial swing at pair protocol.Garrett D'Amore2016-12-13
|
* More comments, and detection of fork-reentrancy. Much effort was spentGarrett D'Amore2016-12-13
| | | | trying to come to a fork-safe solution, but ultimately we gave up.
* Make an interruptible version of message queues.Garrett D'Amore2016-12-12
|
* More pipe details, and clarified locking / call order comments in transport.h.Garrett D'Amore2016-12-12
|
* New pipe stuff.Garrett D'Amore2016-12-12
|
* Initial cmake plumbing (stolen from libnanomsg)Garrett D'Amore2016-12-12
|