| Age | Commit message (Collapse) | Author |
|---|
|
|
|
This is not a proper OTP application.
|
|
Hut has both a Makefile and erlang.mk but the default target
does not build the project. To fix a few packages we have a
special patch HUT_PATCH that will call the appropriate target.
|
|
|
|
|
|
We expect users to fetch these projects from hex.pm now
and the packages no longer worked.
|
|
|
|
|
|
Test broke when git did a security fix related to file://
that we are using for the submodule in this test.
|
|
Thanks to David Ansari for the report.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This is the case in erlang-systemd [1]. As far as I can tell, the format
is the same for the bits Erlang.mk is interested in.
The testsuite is expanded to use version 0.6.0 of erlang-systemd.
[1] https://github.com/rabbitmq/erlang-systemd
|
|
Apparently it was renamed.
|
|
|
|
source file is modified. Without a sleep the source file and the
target may get the same modification time, on some machines.
|
|
|
|
Somehow the develop and master branches of LFE have different
commit hashes and because LFE switched its default branch to
develop the hash we use for testing became incorrect. To fix
I just switched to the develop hash of the equivalent commit.
|
|
There are no popular Mercurial providers anymore. We can
enable it again when things change. Gitlab might add it
in the future.
|
|
|
|
In commit c55d0dcd6182983521d4ab34478cfe4b093edb12, the recipe was
improved to only rebuild out-of-date test files (not everything as it
was the case before).
However in the process, the exit status of the Erlang compiler was not
checked anymore, leading to build failures in the test directory to be
ignored. This patch fixes this bug.
|
|
|
|
https://concuerror.com/
|
|
The rebar in the lager repository is not compatible with OTP-23,
causing this test case to fail.
|
|
... instead of always rebuilding all of them each time a testsuite is
executed.
The only exception is when a Makefile was modified: like for main
source files, test modules are all recompiled in this case.
|
|
Doesn't work on CentOS, too new.
|
|
|
|
|
|
|
|
This is a waste of time and if they were to fail, this does not serve
the Erlang.mk testsuite.
|
|
It was previously "pinned" to its `master` branch. Unforunately, the
testcase started to fail after `farwest` changed its own pinning of
`cowlib`.
By pinning to a specific version (or commit in this case) of `farwest`,
we make sure this won't happen again.
|
|
The query-deps, query-doc-deps, query-rel-deps, query-test-deps
and query-shell-deps targets go through dependencies and print
some information about them:
$ make query-deps
...
cowboy: cowlib git https://github.com/ninenines/cowlib 2.7.3
cowboy: ranch git https://github.com/ninenines/ranch 1.7.1
$ make query-deps QUERY="name repo version absolute_path"
...
cowboy: cowlib https://github.com/ninenines/cowlib 2.7.3 /path/to/deps/cowlib
cowboy: ranch https://github.com/ninenines/ranch 1.7.1 /path/to/deps/ranch
The query-deps target is recursive; the others aren't. This
mirrors the behavior when fetching and building dependencies.
The full list of options is:
QUERY="fetch_method name repo version extra absolute_path"
When an option has no corresponding value, "-" will be printed.
For example most fetch methods do not have extra information to
print. When there is extra information, the value will be prefixed
with a descriptive name. For example: package-name=uuid_erl
This commit also introduces changes to deps handling: some of
the functions for querying will now be used directly. In the
future the goal is to remove the old dep_name, dep_commit, etc.
functions and replace their usage with the new query functions.
Custom fetch methods should implement query functions in order
to have information about the relevant dependencies printed
properly.
|
|
|
|
|
|
Seems it doesn't passs on my machine at the moment.
It gets stuck. Maybe it doesn't work with OTP 22.1?
|
|
|
|
|
|
Internally, strings were stored using Unicode code points. However, when
being written to disk with the `file:write_file()` function, there were
converted to ISO-8859-1. According to the documentation, that is because
the file module is bytewise-oriented: the conversion to another encoding
than ISO-8859-1 is the responsibility of the caller.
Using unicode:character_to_binary() permits the script to convert the
Unicode string to an UTF-8-encoded binary.
Without this patch, the added testcase would fail with the following
error:
gmake[3]: *** No rule to make target '(...)/erlang.mk/test/h��test_core_makedep_non_usascii_paths/deps/test_core_makedep_non_usascii_paths_dep/include/hello.hrl', needed by 'src/hello.erl'. Stop.
In this case, the path (passed from the Makefile to the `makedep.erl`
script) contains UTF-8-encoded `é` characters but it was converted when
doing the final file I/O.
|
|
|
|
Getting:
Nothing to be done for `ct'
Instead of, in later versions:
Nothing to be done for 'ct'
Only check for the start of the sentence instead.
|
