-
-
Notifications
You must be signed in to change notification settings - Fork 59
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Switch testing from ad-hoc scripts to OUnit #34
Conversation
The OS X build failed, but it looks like it has been failing since https://travis-ci.org/rleonid/bisect_ppx/builds/78852923 (4 months ago), just not noticed by Travis due to ignored exit status. I'm guessing you will have more insight into the cause than I. By the way, I develop on OS X, and testing works fine locally. |
I don't mind assuming
Those are good reasons,
I'd like to have a one-to-one mapping from Issue's to PR's, with the caveat that you can always add more issues to keep track of things that need to get done. So if this changes the infrastructure and there is another issue about checking for space, or another for property driven tests, that would be fine.
I don't know. I tend to favor the loosest open source licenses possible as those reflect the amount of time I'm willing to worry about those issues. If you're not opposed, I would keep it GPL, out of respect to @xclerc original wishes (with your name on it).
True.
No objection, redundant usage of |
Thanks for the replies. Don't spend time reading the existing diff please, because I am about to push a new one. Just have to spend some time separating the rename changes into a separate commit... |
Fine.
That's fine. I noticed that it was failing when you started submitting your patches. It is failing because of this bug in
The original intent was to make sure that both order of rewriters worked correctly:
great. #36
great. |
Cool. Let me know when you're ready for a review. It looks great so far! |
Ok, should be ready for review now. There is still the Travis OS X issue, but I don't think it's caused by these commits.
I don't know if you want to look into the OS X Travis build issue before merging this. I was going to take a look, but I was hoping you could offer some insight as to what may have caused it to appear back when it did. |
Dang it, I can't seem to get |
Scratch that, was using a BSD-specific option in the |
- Most ppx- prefixes became redundant when camlp4 support was dropped. - ppx-deriving includes tests for rewriters other than ppx_deriving, so it is now called ppx-integration. - ounit_integration renamed for consistency's sake.
The code is largely complete, but I intend to amend the commit before it's ready for merging. I have a bunch of questions/notes first, though.
I intend to:
tests/_scratch
, but the best place to explain that is in the code..gitignore
for obsolete entries.tests/Makefile
to have features like the old one, and comment usage.ppx_blob
aren't installed. We can make these fatal in Travis to make sure we are testing everything.Before merging:
ocamlfind
now? Some of the older code relies on./configure
and doesn't assumeocamlfind
, but the newer code makes the opposite choice. I went with the former.ppx_tools
requires it, but ppx was supported experimentally in 4.00 (or was it?). I'm in favor of assuming 4.02, it makes it possible to assume|>
and@@
:) Perhaps we should add a constraint toopam
for clarity.ppx-deriving
directory seems like it should be renamed, because it contains tests with various ppx rewriters, not justppx_deriving
.ppx-*
directories are named that way from the time when ppx was an alternative in this code base to camlp4. The prefix is confusing withppx-deriving
. Let me know if you object to dropping the prefix from these.Other notes:
ppx_blob
tests are failing due to Warnings about compiler-libs when using ppx_blob johnwhitington/ppx_blob#1. I skipped them.ppx_deriving
test has been failing without us noticing, because the testing framework ignored the exit code. For example: https://travis-ci.org/rleonid/bisect_ppx/jobs/102869628#L507. The translated version is still failing now, for the same reason, but I marked it skipped, and we can fix it in a separate change.ppx-deriving
for testingppx_env
and such. I deleted this. We can recover and adapt it later from the history.combine-expr
tests 1 and 2 have the same output. Might want to avoid that :)The output of
make tests
(aftermake all
) now looks something like this:In case of failure, you get a non-zero exit code and output like this: