GNU Make's wildcard function does not recurse into subdirectories when
passed the '**' glob, which results in adding a dependency only to .go
files in the first level of subdirectories under the source root.
We shell out to 'find' instead, which catches all .go files in the
given directory.
Running make -jN would result in the tests starting to execute
before the tests binaries were built, resulting in the error:
"Cannot find the tests' local ipfs tool"
Each test now depends on the deps. They also depend on a new
target for cleaning the test results, so that the tests can
write new clean results.
The aggregate target also needs to depend on the same test
results clean target, as well as the tests themselves, so
that the aggregation happens when all tests have finished
running.
By introducing a separate target for cleaning test results we
also ensure that we don't end up removing and rebuilding
the binary on each test run.
The result is that the tests *can* be run with with -jN > 1,
but individual tests may still not supports this, so to get
stable test results it's still recommended to run them in
sequence.
The GOFLAGS variable makes it possible to run all sharness
tests with go binaries built with some special flags.
The "race" target makes it easy run the sharness tests
with go binaries built with the -race flag.
License: MIT
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
...and remove bin/* stuff from the .gitignore
as /test/bin is already in the root .gitignore.
License: MIT
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
This makes it possible to build binaries with
different flags.
The content of the GOFLAGS variable is stored
in a IPFS-BUILD-OPTIONS file, so that if GOFLAGS
changes a rebuild of the binaries with the new
flags is forced.
License: MIT
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
This script can be used in a Makefile to detect flag changes
and to save the new flags in a file.
License: MIT
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
As `ipfs version` and `ipfs help` are basic commands
it is ok to test them online in t0060.
License: MIT
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
It looks like some tests that were expected failures
are now always succeeding, so let's mark them as such.
License: MIT
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
To do that the current tests are move inside a
test_object_cmd() function. This function is then first
called before running the daemon for offline tests
and then after starting the daemon for online tests.
License: MIT
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
To do that the current tests are move inside a
test_ls_cmd() function. This function is then first
called before running the daemon for offline tests
and then after starting the daemon for online tests.
License: MIT
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
To do that the current tests are move inside a
test_get_cmd() function. This function is then first
called before running the daemon for offline tests
and then after starting the daemon for online tests.
License: MIT
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
It looks like there were two problems:
- "read" command probably needs the name of the variable
it reads into
- [[ didn't work
This might be because on Linux /bin/sh is sometimes not
bash. For example on Ubuntu it is a symlink to dash.
License: MIT
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
Tests in this script had been deactivated because
there was no good way to wait until the daemon was
ready. Now that we can use pollEndpoint for that
it's time to reactivate the tests.
License: MIT
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
Add information from the discussion of PR #891.
While at it, suggest using -v and -i options to debug one test.
License: MIT
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
This makes the following changes:
- remove some superfluous \ at the end of some lines
- use `cat <<-\EOF >expected` to ignore the starting tabs
in the here document
- add missing && after some `cat`
- use test_cmp to compare files
License: MIT
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
we have a problem where initializing daemons with the same api port
often fails-- it hangs indefinitely. The proper solution is to make
ipfs pick an unused port for the api on startup, and then use that.
Unfortunately, ipfs doesnt yet know how to do this-- the api port
must be specified. Until ipfs learns how to do this, we must use
specific port numbers, which may still fail but less frequently
if we at least use different ones.
