I think I've found a bug in libev backend, in function lws_libev_io(). I'm using latest version from master branch.
When deleting a context with active connections via lws_context_destroy(), context->being_destroyed is set to 1 early in the function, before the loop calling lws_close_free_wsi() on each active connection.
lws_close_free_wsi() calls remove_wsi_socket_from_fds(), which calls lws_libev_io(), and here is my problem :
lws_libev_io() returns without doing anything if context->being_destroyed is set, so libev callbacks for deleted connections file descriptors stay registered after context is destroyed, which may lead to segfault/undefined behaviour if these file descriptors get reused later (which would trigger the callbacks).
I think the "if (!pt->io_loop_ev || context->being_destroyed) return;" statement should be replaced with " if (!pt->io_loop_ev) return;"
This fixes the problem for me and I have not seen any side effect yet. Moreover, libuv backend does not have such a test.
Also introduce CMake LWS_WITH_ZIP_FOPS defaulting to ON that builds junzip.c and
make sure this is exported in lws_config.h (included by libwebsockets.h)
Improve lws handling of stdint.h import or simulate it.
- if protocol set, allocate own user_space
If the child wsi wants the parent wsi user_space, it can use
lws_wsi_user(lws_get_parent(child_wsi))
- raw file close processing handles parent-child relationship
1) There's now a .fops pointer that can be set in the context creation info. If set, the array of
fops it points to (terminated by an entry with .open = NULL) is walked to find out the best vfs filesystem
path match (comparing the vfs path to fops.path_prefix) for which fops to use.
If none given (.fops is NULL in info) then behaviour is as before, fops are the platform-provided one only.
2) The built in fileserving now walks any array of fops looking for the best fops match automatically.
3) lws_plat_file_... apis are renamed to lws_vfs_file_...
