Over time, it has cost us more developer cycles than it has gained.
It has hidden intermittant bugs, and allowed cruft to accumulate:
when we eventually tried to figure out what was going wrong, the
actual change which caused it was now stale and forgotten.
This was a particular bane during the connectd rewrite, and I
worked through some issues which had occurred before, but were not
more likely.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
I have no idea why someone else owns the directory suddenly, but all git
commands fail. Workaround as suggested by the error message.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Mostly comments and docs: some places are actually paths, which
I have avoided changing. We may migrate them slowly, particularly
when they're user-visible.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
check-dbstmts was just running the normal pytest, AFAICT:
```
export TEST_CHECK_DBSTMTS=0
+ TEST_CHECK_DBSTMTS=0
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Setting VALGRIND=1 actually does nothing here; reduce it to two cases,
covering gcc and clang, sqlite3 and postgres.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Add an extra run configuration for the proto-tests.
Proto-tests require DEVELOPER=1; enabling EXPERIMENTAL_FEATURES=1
by default ensures that experimental additions are automatically put
under test also.
It's been causing me quite some headache, and I don't see the point in
jumping through the hoops for something that can be trivially fixed by
having the required build tools.
A new target `check-gen-updated` to verify that all derived/generated
files that were modified were also checked in. This is used on CI to
check, and is not added to `check` since it'll complain on dirty
trees, i.e., before the devs check in their changes.
Suggested-by: Rusty Russell <@rustyrussell>
We split into quick smoke-tests, normal-tests, and valgrind-tests. The
first provides a quick-abort preventing overusing resources on what
will be a failed run anyway.
We'd like to have some more structure to the tests, with smoke-tests
going first, followed by valgrind and normal tests running in
parallel. So to cut down on the copy-pasta we extract the scripts from
the yaml, into their own files from where they can be easily invoked.
We don't have a good way of referring to the configuration that
failed, so let's give them a numberic ID. Particularly useful for the
artifacts that'd be overwritten otherwise.
