d02c94090c Fix error reporting in process-async-https-with-delay
9a4862a973 * doc/misc/org.org: Remove spurious markup.
768ed1476a Make Tramp version check more robust
7f778c6943 Fix debugging with GDB when a breakpoint has multiple loca...
25e53e9391 ; * lisp/files.el (file-expand-wildcards): Doc fix.
3ea9357d10 Update documentation of 'aset' and 'store-substring'
# Conflicts:
# lisp/files.el
* test/src/process-tests.el (process-async-https-with-delay): Use
'plist-get' instead of 'assq' in testing for a connection error.
The 'status' variable is a plist, not an alist. (Bug#55849)
This addresses a FIXME comment in lisp/emacs-lisp/comp.el,
relating to the number of subsidiary processes used by
comp-run-async-workers in native compilation.
* admin/merge-gnulib (GNULIB_MODULES): Add nproc.
* doc/lispref/processes.texi (Process Information), etc/NEWS:
Document num-processors.
* lib/gnulib.mk.in, m4/gnulib-comp.m4: Regenerate.
* lib/nproc.c, lib/nproc.h, m4/nproc.m4:
New files, copied from Gnulib by admin/merge-gnulib.
* lisp/emacs-lisp/comp.el (w32-get-nproc): Remove decl.
(comp-effective-async-max-jobs): Use num-processors.
* src/process.c: Include nproc.h.
(Fnum_processors): New function.
(syms_of_process): Define ‘all’, ‘current’, ‘num-processors’.
* src/w32proc.c (Fw32_get_nproc): Add FIXME comment.
* test/src/process-tests.el (process-num-processors): New test.
If a TLS handshake is attempted before the completion of an
asynchronous TCP connection has been ascertained, our local state will
not be set up correctly for further progress and the sentinel "open"
event will never be sent. This can occur if sufficient time passes
after the initiation of an async TCP connection so that by the time
`wait_reading_process_output` is called, the connection has already
been established on the TCP level.
This somewhat timing-sensitive bug has plagued HTTPS connections on
some platforms, notably macOS, for a long time (bug#49449).
* src/process.c (wait_reading_process_output): Gate the TLS handshake
by the NON_BLOCKING_CONNECT_FD flag. The flag will be cleared as soon
as the TCP socket is found to be writable.
* test/src/process-tests.el (process-async-https-with-delay):
New test.
* test/src/process-tests.el
(process-tests/fd-setsize-no-crash/make-network-process): Skip
test on Cygwin to avoid hang due to connect/accept handshake.
(Bug#49496)
Bug#46641
The network-stream-tests actually work fine when the local machine has
no IP at all, but cause a crash in the GnuTLS library when there is an
IP configured but the interface is down.
* test/lisp/net/network-stream-tests.el (internet-is-working): New
defvar, checks if we can resolve "google.com".
(connect-to-tls-ipv4-nowait, connect-to-tls-ipv6-nowait)
(open-network-stream-tls-nowait, open-gnutls-stream-new-api-nowait)
(open-gnutls-stream-old-api-nowait): Use it to check for working
Internet access.
* test/src/process-tests.el (internet-is-working): New defvar, checks
if we can resolve "google.com".
(lookup-family-specification, lookup-unicode-domains)
(unibyte-domain-name, lookup-google, non-existent-lookup-failure): Use
it to check for working Internet access.
* test/src/process-tests.el (process-tests/multiple-threads-waiting):
Also check that 'thread-join' and 'thread-last-error' return the
expected errors.
This unit test tests that we can call 'accept-process-output' in
parallel from multiple threads.
* test/src/process-tests.el (process-tests/multiple-threads-waiting):
New unit test.
That way, the tests only depend on Emacs, and not on utilities that
might not be available during test time.
* test/src/process-tests.el (process-tests--eval)
(process-tests--emacs-command, process-tests--emacs-binary)
(process-tests--dump-file)
(process-tests--usable-file-for-reinvoke): New helper functions.
(process-tests/sentinel-called)
(process-tests/sentinel-with-multiple-processes): Use them.
When we're trying to notify a process about a status change, we need
to ignore the SIGCHLD pipe temporarily, otherwise the code would
likely not run into the timeout case that's necessary for a status
change to happen.
* src/process.c (wait_reading_process_output): Ignore the SIGCHLD pipe
when notifying a process about a status change.
* test/src/process-tests.el (process-tests/sentinel-called)
(process-tests/sentinel-with-multiple-processes): New unit tests.
If we receive and handle a SIGCHLD signal for a process while waiting
for that process, 'pselect' might never return. Instead, we have to
explicitly 'pselect' that the process status has changed. We do this
by writing to a pipe in the SIGCHLD handler and having
'wait_reading_process_output' select on it.
* src/process.c (child_signal_init): New helper function to create a
pipe for SIGCHLD notifications.
(child_signal_read, child_signal_notify): New helper functions to
read from/write to the child signal pipe.
(create_process): Initialize the child signal pipe on first use.
(handle_child_signal): Notify waiters that a process status has
changed.
(wait_reading_process_output): Make sure that we also catch
SIGCHLD/process status changes.
* test/src/process-tests.el
(process-tests/fd-setsize-no-crash/make-process): Remove workaround,
which is no longer needed.
* test/src/process-tests.el (ipv6-is-available): New function for
checking whether we have a globally routable IPv6 prefix assigned.
(lookup-family-specification): Use 'ipv6-is-available' to check for
IPv6. Use 'localhost' instead of 'google.com' to test
'network-lookup-address-info' API.
(lookup-google): Use 'ipv6-is-available' to check for
IPv6.
* test/lisp/net/nsm-tests.el (nsm-ipv6-is-available): Rename to
'ipv6-is-available', make identical to the one in
test/src/process-tests.el.
The test 'process-tests/fd-setsize-no-crash/make-process' used to call
'sleep' to ensure that enough processes are live to trigger a
FD_SETSIZE overflow. However, we can just call 'cat' instead and
close standard input when done. That way, we only wait as long as
needed.
* process-tests.el
(process-tests/fd-setsize-no-crash/make-process): Invoke 'cat' instead
of 'sleep'. Close standard input to exit the 'cat' processes.
The 'process-tty-name' already provides the TTY name, we don't have
interrogate the TTY host.
* test/src/process-tests.el
(process-tests/fd-setsize-no-crash/make-serial-process): Use
'process-tty-name' instead of having the TTY host print its TTY
name. Check whether TTY names are unique.
(process-tests--new-pty, process-tests--with-temp-file): Remove;
no longer used.
This works even if another thread or process resets the resource limit
for open file descriptors, e.g., using 'prlimit' on GNU/Linux.
* src/process.c (create_process, create_pty, Fmake_pipe_process)
(Fmake_serial_process, connect_network_socket)
(server_accept_connection): Limit file descriptors to FD_SETSIZE.
* test/src/process-tests.el (process-tests--with-raised-rlimit): New
helper macro.
(process-tests--fd-setsize-test): Rename from
'process-tests--with-many-pipes'. Increase resource limit during test
if possible.
(process-tests/fd-setsize-no-crash/make-process)
(process-tests/fd-setsize-no-crash/make-pipe-process)
(process-tests/fd-setsize-no-crash/make-network-process)
(process-tests--new-pty): Rename callers.
* test/src/process-tests.el
(process-tests/fd-setsize-no-crash/make-process): Allow special exit
codes that can happen if terminal setup fails in the child process.
Instead of trying to generate the right number of processes,
pre-create lots of unused pipe processes until creation fails. Extend
the tests to the 'pty' connection type and other kinds of process
objects.
* test/src/process-tests.el (process-tests--ignore-EMFILE)
(process-tests--with-buffers, process-tests--with-processes)
(process-tests--with-many-pipes, process-tests--with-temp-file)
(process-tests--with-temp-directory): New helper macros.
(process-tests/fd-setsize-no-crash/make-process): Renamed from
'process-tests/fd-setsize-no-crash'. Fail on timeout. Also test the
'pty' connection type. Pre-create lots of pipe processes so we reach
the FD_SETSIZE limit faster. Ignore EMFILE more precisely, if
possible.
(process-tests/fd-setsize-no-crash/make-pipe-process)
(process-tests/fd-setsize-no-crash/make-network-process)
(process-tests/fd-setsize-no-crash/make-serial-process): New tests
that test FD_SETSIZE limits for other kinds of processes.
(process-tests--EMFILE-message): New helper function and cache
variable.
(process-tests--new-pty): New helper function.
* src/w32.c (pipe2): When forcibly closing pipe handles due to
overflow of FD_SETSIZE, set the handles to -1, to avoid assertion
violations in emacs_close.
* test/src/process-tests.el (process-tests/fd-setsize-no-crash):
No need to skip this test anymore.
* test/src/process-tests.el (process-tests/fd-setsize-no-crash):
Skip this test on windows-nt systems, as we cannot use more than
FD_SETSIZE file descriptors there: if we try, we crash.
* src/process.c (Fmake_network_process) [AF_INET6]: Only build ::1
localhost when IPv6 is supported.
(Fnetwork_lookup_address_info) [AF_INET6]: Move check for Qipv6 inside
ifdef, since its definition depends on AF_INET6. Don't return IPv6
addresses when they're not supported.
* test/src/process-tests.el (lookup-family-specification,
lookup-google): Only do IPv6 lookup if IPv6 is supported.
* test/src/process-tests.el (set-process-filter-t): Don't assume
subprocess output will come in a single chunk, keep waiting for more
data until next "prompt" is read from subprocess.
* test/src/process-tests.el: (lookup-family-specification,
lookup-unicode-domains, unibyte-domain-name, lookup-google,
non-existent-lookup-failure): Skip on Hydra, which doesn't have DNS.
Fix buggy test condition. (Bug#37165)
