Logs are written by several servers and all of them must be finished
writing before the test results can be determined. This means each
server must have its own lock file rather than sharing a single one,
which is how it was done up to now. Previously, the first server to
complete a test would clear the lock before the other server was done,
which caused flaky tests.
Lock files are now all found in their own directory, so counting locks
equals counting the files in that directory. The result is that the
proxy logs are now reliably written which actually changes the expected
output for two tests.
Fixes#11231Closes#11259
Test files in subdirectories were not created after parallel test log
directories were moved down a level due to a now-bad comparison.
Follow-up to 92d7dd39
Ref #11264Closes#11267
Aka "jumbo" or "amalgamation" builds. It means to compile all sources
per target as a single C source. This is experimental.
You can enable it by passing `-DCMAKE_UNITY_BUILD=ON` to cmake.
It requires CMake 3.16 or newer.
It makes builds (much) faster, allows for better optimizations and tends
to promote less ambiguous code.
Also add a new AppVeyor CI job and convert an existing one to use
"unity" mode (one MSVC, one MinGW), and enable it for one macOS CI job.
Fix related issues:
- add missing include guard to `easy_lock.h`.
- rename static variables and functions (and a macro) with names reused
across sources, or shadowed by local variables.
- add an `#undef` after use.
- add a missing `#undef` before use.
- move internal definitions from `ftp.h` to `ftp.c`.
- `curl_memory.h` fixes to make it work when included repeatedly.
- stop building/linking curlx bits twice for a static-mode curl tool.
These caused doubly defined symbols in unity builds.
- silence missing extern declarations compiler warning for ` _CRT_glob`.
- fix extern declarations for `tool_freq` and `tool_isVistaOrGreater`.
- fix colliding static symbols in debug mode: `debugtime()` and
`statename`.
- rename `ssl_backend_data` structure to unique names for each
TLS-backend, along with the `ssl_connect_data` struct member
referencing them. This required adding casts for each access.
- add workaround for missing `[P]UNICODE_STRING` types in certain Windows
builds when compiling `lib/ldap.c`. To support "unity" builds, we had
to enable `SCHANNEL_USE_BLACKLISTS` for Schannel (a Windows
`schannel.h` option) _globally_. This caused an indirect inclusion of
Windows `schannel.h` from `ldap.c` via `winldap.h` to have it enabled
as well. This requires `[P]UNICODE_STRING` types, which is apperantly
not defined automatically (as seen with both MSVS and mingw-w64).
This patch includes `<subauth.h>` to fix it.
Ref: https://github.com/curl/curl/runs/13987772013
Ref: https://dev.azure.com/daniel0244/curl/_build/results?buildId=15827&view=logs&jobId=2c9f582d-e278-56b6-4354-f38a4d851906&j=2c9f582d-e278-56b6-4354-f38a4d851906&t=90509b00-34fa-5a81-35d7-5ed9569d331c
- tweak unity builds to compile `lib/memdebug.c` separately in memory
trace builds to avoid PP confusion.
- force-disable unity for test programs.
- do not compile and link libcurl sources to libtests _twice_ when libcurl
is built in static mode.
KNOWN ISSUES:
- running tests with unity builds may fail in cases.
- some build configurations/env may not compile in unity mode. E.g.:
https://ci.appveyor.com/project/curlorg/curl/builds/47230972/job/51wfesgnfuauwl8q#L250
Ref: https://github.com/libssh2/libssh2/issues/1034
Ref: https://cmake.org/cmake/help/latest/prop_tgt/UNITY_BUILD.html
Ref: https://en.wikipedia.org/wiki/Unity_buildCloses#11095
Parallel testing is enabled by using a nonzero value for the -j option
to runtests.pl. Performant values seem to be about 7*num CPU cores, or
1.3*num CPU cores if Valgrind is in use.
Flaky tests due to improper log locking (bug #11231) are exacerbated
while parallel testing, so it is not enabled by default yet.
Fixes#10818Closes#11246
Such as what happens with the --repeat option. Some functions are
changed to pass the runner ID instead of relying on the non-unique test
number.
Ref: #10818
Each runner needs a unique random seed to reduce the chance of port
number collisions. The new scheme uses a consistent per-runner source of
randomness which results in deterministic behaviour, as it did before.
Ref: #10818
The main test loop is now able to handle multiple runners, or no
additional runner processes at all. At most one process is still
created, however.
Ref: #10818
These are two boolean options to ask curl to use the native OS's CA
store when verifying TLS servers. For peers and for proxies
respectively.
They currently only have an effect for curl on Windows when built to use
OpenSSL for TLS.
Closes#11049
... instead of putting them in the regular pid directories because
systems generally have strict length requirements for the path name to
be shorter than 107 bytes and we easily hit that boundary otherwise.
The new concept generates two random names: one for the socks daemon and
one for http.
Reported-by: Andy Fiddaman
Fixes#11152Closes#11166
- leave transfer loop when --limit-rate is in effect and has
been received
- adjust stream window size to --limit-rate plus some slack
to make the server observe the pacing we want
- add test case to confirm behaviour
Closes#11115
Because of how libssh2_userauth_keyboard_interactive_ex() works: the
libcurl callback allocates memory that is later free()d by libssh2, we
must set the custom memory functions.
Reverts 8b5f100db3
Ref: https://github.com/libssh2/libssh2/issues/1078Closes#11235
... and have curl trim the end when it reaches the expected total amount
of bytes instead of over-sending.
Reported-by: JustAnotherArchivist on github
Closes#11223
This change replaces the previous method of picking a port number at
random to try to start servers on, then retrying up to ten times with
new random numbers each time, with a function that creates a server
socket on port zero, thereby getting a suitable random port set by the
kernel. That server socket is then closed and that port number is used
to setup the actual test server on.
There is a risk that *another* server can be started on the machine in
the time gap, but the server verification feature will detect that.
Closes#11220
If you try to assign and export on the same line on some older /bin/sh
implementations, it complains:
```
$ export "NAME=value"
NAME=value: is not an identifier
```
This commit rewrites run-compiler's assignments and exports to work with
old /bin/sh, splitting assignment and export into two separate
statements, and only quote the value. So now we have:
```
NAME="value"
export NAME
```
While we're here, make the same change to the two supporting
assign+export lines preceeding the script to be consistent with how
exports work throughout the rest of configure.ac.
Closes#11228
