Only run -dip1000 checks on CircleCi on not on the auto-tester#6443
Only run -dip1000 checks on CircleCi on not on the auto-tester#6443dlang-bot merged 1 commit intodlang:masterfrom
Conversation
|
Thanks for your pull request, @wilzbach! Bugzilla referencesYour PR doesn't reference any Bugzilla issue. If your PR contains non-trivial changes, please reference a Bugzilla issue or create a manual changelog. Testing this PR locallyIf you don't have a local development environment setup, you can use Digger to test this PR: dub fetch digger
dub run digger -- build "master + phobos#6443" |
posix.mak
Outdated
| $(UT_D_OBJS): $(ROOT)/unittest/%.o: %.d | ||
| @mkdir -p $(dir $@) | ||
| $(DMD) $(DFLAGS) $(UDFLAGS) $(aa[$(subst /,.,$(basename $<))]) -c -of$@ $< | ||
| $(DMD) $(DFLAGS) $(UDFLAGS) $(basename $<) -c -of$@ $< |
There was a problem hiding this comment.
That $(basename $<) seems to cause issues.
There was a problem hiding this comment.
Yeah, remove $(basename $<) completely
|
@wilzbach did you see our comments? |
|
@marler8997 Just realized I could edit it. We'll see if that helps. |
|
How does editing other people's PRs work? Did you just push to the PR branch? |
|
You have to be a member. I just edit it directly in github (I know there's other ways, but I'm not familiar enough with how git works). |
|
It looks like Seb's branch has your commit on top of his. So I guess when you make a PR, you are granting push rights to the repo you are pushing to. Very interesting use case... |
Yes, it's a box that's checked by default. See here: https://help.github.com/articles/allowing-changes-to-a-pull-request-branch-created-from-a-fork/ Seb knows how to do it without going "on top" of the original, but I have no clue how to do that :) |
The command I sent you should do the trick. Just amend his commit and do a force push to his branch. |
|
FWIW I have a simple script that automates checking out and pushing to any PR: So it looks like the error is unrelated... great :/ |
|
Your tools are in bash :( If I decide to use them maybe I'll convert them to D :) |
|
P.S. I have one custom git tool that I use for pull requests https://github.com/marler8997/utils It fetches the branch from |
wilzbach
force-pushed
the
dip1000-unblock
branch
from
8e9aa6a to
34f2a0a
Compare
|
Regarding the DAutoTest failure -> dlang/dlang.org#2345 (I also squashed the commits as discussed here and to restart DAutoTest)
Yes, but the new uni semester started this week and I was rather busy. Sorry. |
|
School...bleh. Majoring in compter science I assume? How's that going? |
My major actually is computational biology, so well not entirely on track ;-) Anyhow this is passing now, so I assume you still want this for your dmd PR. |
|
There you go. Should merge soon. |
|
@schveiguy auto-merge isn't working because this PR doesn't have an approval yet. |
|
|
Oops! |
|
@marler8997 what @wilzbach meant is that someone with commit rights must approve. |
What's the bugzilla issue for that? It seems to be mis-categorized, as I can't find it. |
|
I added one: https://issues.dlang.org/show_bug.cgi?id=18820 |
5 participants
It seems that there are still many issues with -dip1000 and for the time being it's apparentely better to only test the separate
.testtargets with -dip1000.This is not ideal as -dip1000 is now only checked on Linux 64-bit, but before it wasn't run on Windows neither and -dip1000 is not very platform dependent.
This should hopefully unblock dlang/dmd#8124
The only downside to this is that now a PR at dmd will no longer run Phobos with -dip1000 checks. We can either
*.test*.testtargets too