Add DragonFlyBSD support for druntime #1999

dkgroot · 2017-12-18T13:59:25Z

Related: DMD Pull Request
Bootstrapped via 'dmd-cxx' branch: See

Related PR:

Notes:

Fixed getAddrInfo issue in phobos/std/socket.d by giving addrinfo.ai_socktype a default value of SOCK_STREAM.
(Without it getAddrInfo() will generate an error:'servname not supported for ai_socktype' when hints.ai_socktype is not set.)
1 minor dragonfly related FIXME entries in src/core/sync/mutex.d (Caused by the implementation of pthread implementation on DragonFlyBSD): trylock after destroying a mutex does not return an error. (similar to 'Use after free')

dlang-bot · 2017-12-18T13:59:25Z

Thanks for your pull request and interest in making D better, @dkgroot! We are looking forward to reviewing it, and you should be hearing from a maintainer soon.
Please verify that your PR follows this checklist:

My PR is fully covered with tests (you can see the annotated coverage diff directly on GitHub with CodeCov's browser extension
My PR is as minimal as possible (smaller, focused PRs are easier to review than big ones)
My PR follows the DStyle
I have provided a detailed rationale explaining my changes
New or modified functions have Ddoc comments (with Params: and Returns:)

Please see CONTRIBUTING.md for more information.

If you have addressed all reviews or aren't sure how to proceed, don't hesitate to ping us with a simple comment.

Bugzilla references

Your PR doesn't reference any Bugzilla issue.

If your PR contains non-trivial changes, please reference a Bugzilla issue or create a manual changelog.

nemanja-boric-sociomantic · 2017-12-18T14:05:21Z

It would be very nice if you could split this into several commits (as per your checklist in the PR), and amend it to say "Add DragonFlyBSD support for druntime" instead of "Port druntime to ..."

dkgroot · 2017-12-18T14:28:46Z

src/core/sys/posix/netdb.d

+    {
+        int         ai_flags;
+        int         ai_family;
+        int         ai_socktype = SOCK_STREAM;           /* socktype default value required to be able to perform getAddrInfo on DragonFlyBSD


Fixed getAddrInfo issue in phobos/std/socket.d by giving addrinfo.ai_socktype a default value of SOCK_STREAM.
Without a default value for socktype, getAddrInfo() will generate an error:'servname not supported for ai_socktype' when hints.ai_socktype is not set on DragonFlyBSD.

Would it not be better to change std.socket instead?

You mean to set the default value for ai_socktype, right ?

I did try to find a good location in std.socket first (When the unittest failed and pointed out the issue). I tried this, and it worked, but i did not like this solution very much. Maybe mostly caused by the location i choose to 'fix' the issue.

diff --git a/std/socket.d b/std/socket.d index 361ec73da..ba68e1738 100644 --- a/std/socket.d +++ b/std/socket.d @@ -963,6 +963,13 @@ AddressInfo[] getAddressInfo(T...)(in char[] node, T options) else static assert(0, "Unknown getAddressInfo option type: " ~ typeof(option).stringof); } + version (DragonFlyBSD) + { + if (!hints.ai_socktype) + { + hints.ai_socktype = SOCK_STREAM; + } + } return () @trusted { return getAddressInfoImpl(node, service, &hints); }(); }

Also doing this in runtime, 'fixes' the issue even when/if running without using phobos. And by just setting the default value it does not prevent anyone from overwriting the value or even initializing it with void.

dkgroot · 2017-12-18T14:30:36Z

src/rt/util/typeinfo.d


-            assert(f1  == f2);
+            assert(f1 == f2);
+            version (DragonFlyBSD) {} else   // FIXME:  //core.exception.AssertError@src/rt/util/typeinfo.d(249): unittest failure


// FIXME:
core.exception.AssertError@src/rt/util/typeinfo.d(249): unittest failure
Have not figured out what is going on here, nor what this unittest is trying to check.

You can see the F array there,it has four elements, that they all have the same value in the complex field, and first assert is checking if their value comparison is sane (they all should equal each other, as they should all be in the (0, 0) point). The next assert (I assume this is the failing one, you forgot braces around else) f1 !is f2 checks if these elements are all different (because they are different members of the array). Not sure what's going on in here.

(The {} are not missing, it's only excluding a single line, without having to change it's surroundings, Wanted to keep the 'DragonFlyBSD' string and 'FIXME' on a single line, so it's easy to find using grep. I noticed this 'unittest exclusion' pattern in other commits.).
It's just a little unclear based on the line i choose to attach this comment to. I should have include the next time.

Actual assert Issue reraised below.

dkgroot · 2017-12-18T14:32:58Z

src/core/sync/mutex.d

    // For example, Bionic doesn't appear to do so, so this test is
    // not run on Android.
    version (CRuntime_Bionic) {} else
+    version (DragonFlyBSD) {} else               // FIXME : Will have to write a test program to see what is going on here.


// FIXME :
Will have to write a little c-program to check if DragonFlyBSD pthread does allow trylock after the mutex has been destroyed and does return 0 in this case.

Have not found a good solution for this. But trying to lock after destroying the lock does not really make sense anyway. Reading up on some of the documentation regarding destroying a lock, it looks like it is undefined behaviour to try and lock the mutex after destruction. Even though it seems to "work" in some implementations (e.g.Linux).

Leaving it like this for now.
Would require changing pthread implementation on dragonfly, or adding a check to "class Mutex" to verify if it has been initialized/constructed. Both don't really seem reasonable.

dkgroot · 2017-12-18T14:35:10Z

posix.mak

+#       $(GREP) -n -U -P "([ \t]$$|\r)" $(CWS_MAKEFILES) ; test "$$?" -ne 0
+#       $(GREP) -n -U -P "( $$|\r|\t)" $(NOT_MAKEFILES) ; test "$$?" -ne 0
+#endif
+checkwhitespace: $(INSTALL_DIR)/${OS}/bin${MODEL}/dmd $(TOOLS_DIR)/checkwhitespace.d 


Instead of using platform dependent grep
we should use the tools/checkwhitespace.d D-version
Note: This not directly related to supporting DragonFlyBSD, and should maybe have been in a separate commit.

It's not as easy as that. There's currently no logic in place that ensures that tools is actually checked out and there have been long discussion of tools being an option or required dependency for building DMD + Druntime + Phobos.
The current status quo is that it's okay to use use tools in non essential targets.
Anyhow as you said yourself, please put this in a separate PR.

dkgroot · 2017-12-18T14:36:04Z

@nemanja-boric-sociomantic
I had a little mishap during rebasing on top of master. Before this these were seperate commits. It it helps i added some comments (See above) to clarify the locations where the issues arose).

dkgroot · 2017-12-18T14:39:32Z

src/core/memory.d

    assert(errno == fakePureErrno()); // errno shouldn't change
+
+    version (DragonFlyBSD) {} else    // FIXME : z seems to be allocated and not null. Why would size_t.max - 2 not be allowed to allocate ? or is the assertion wrong
    assert(z is null);


z seems to be allocated and not null.
Why would size_t.max - 2 not be allowed to allocate ? or is the assertion wrong

I can't find any documented reference, but I believe this is testing if the glibc's malloc would fail because of the internal block size and the additional alignment (and any other internal metadata) that should be added to the internal allocation, which while adding with the values close to size_t.max would lead to the overflow.

Wrote a little test program in ansi-c

#include <stdio.h> #include <stdlib.h> #include <stdint.h> #include <errno.h> // On Linux #if !defined(SIZE_MAX) #define SIZE_MAX __SIZE_MAX__ // ((size_t)-1) = 9223372036854775807 = LONG_MAX #endif // On DragonFlyBSD //#define SIZE_MAX UINT64_MAX // from /usr/include/machine/int_limits.h int main () { printf("SIZE_MAX=%lu\n", SIZE_MAX); unsigned long tests[] = { 9223372036854775807, 223372036854775808, 23372036854775808, 3372036854775808, 372036854775808, 72036854775808, 2036854775808, 36854775808, INTPTR_MAX, UINTPTR_MAX, SIZE_MAX - 2, SIZE_MAX, SIZE_MAX + 'a', SIZE_MAX + SIZE_MAX, SIZE_MAX * 20}; for (int test = 0; test < 14; test++) { printf("test%d: malloc size:%lu, ", test, tests[test]); uint *ptr = (int *) malloc(tests[test]); if (ptr == NULL) { printf("malloc failed, ptr == NULL, errno:%d\n", errno); } else { printf(" ptr == %p\n", ptr); free(ptr); ptr = NULL; } } return 0; }

Compiled using gcc test_malloc.c -o test_malloc

Linux x86_64 Results:

SIZE_MAX=18446744073709551615 test0: malloc size:9223372036854775807, malloc failed, ptr == NULL, errno:12 test1: malloc size:223372036854775808, malloc failed, ptr == NULL, errno:12 test2: malloc size:23372036854775808, malloc failed, ptr == NULL, errno:12 test3: malloc size:3372036854775808, malloc failed, ptr == NULL, errno:12 test4: malloc size:372036854775808, malloc failed, ptr == NULL, errno:12 test5: malloc size:72036854775808, malloc failed, ptr == NULL, errno:12 test6: malloc size:2036854775808, malloc failed, ptr == NULL, errno:12 test7: malloc size:36854775808, malloc failed, ptr == NULL, errno:12 test8: malloc size:9223372036854775807, malloc failed, ptr == NULL, errno:12 test9: malloc size:18446744073709551615, malloc failed, ptr == NULL, errno:12 test10: malloc size:18446744073709551613, malloc failed, ptr == NULL, errno:12 test11: malloc size:18446744073709551615, malloc failed, ptr == NULL, errno:12 test12: malloc size:96, ptr == 0x7f6cd80008c0 test13: malloc size:18446744073709551614, malloc failed, ptr == NULL, errno:12

On Linux the destination pointer seems to be checked as well.

DragonFlyBSD (x86_64) Results:

SIZE_MAX=18446744073709551615 test0: malloc size:9223372036854775807, malloc failed, ptr == NULL, errno:12 test1: malloc size:223372036854775808, malloc failed, ptr == NULL, errno:12 test2: malloc size:23372036854775808, malloc failed, ptr == NULL, errno:12 test3: malloc size:3372036854775808, malloc failed, ptr == NULL, errno:12 test4: malloc size:372036854775808, malloc failed, ptr == NULL, errno:12 test5: malloc size:72036854775808, ptr == 0x800800000 test6: malloc size:2036854775808, ptr == 0x800800000 test7: malloc size:36854775808, ptr == 0x800800000 test8: malloc size:9223372036854775807, malloc failed, ptr == NULL, errno:12 test9: malloc size:18446744073709551615, ptr == 0x800455000 test10: malloc size:18446744073709551613, ptr == 0x800455000 test11: malloc size:18446744073709551615, ptr == 0x800455000 test12: malloc size:96, ptr == 0x8006200c0 test13: malloc size:18446744073709551614, ptr == 0x800455000

On DragonFly it seems to be ok, as long as the upperbound of '9223372036854775807' is observed. Using SIZE_MAX however results faulty behaviour.

Perhaps is worth raising the bug report in the DragoFlyBSD tracker?

After creating the above 'proof' i think that makes sense.
DragonFlyBSD Bug Report

Found a potential solution to the malloc(SIZE_MAX) issue. Send a diff to upstream dragonfly issue manager (Issue 3114), seems to resolve the case reported above.

Issue has been fixed in dragonfly master
Removing this unittest escape

nemanja-boric-sociomantic · 2017-12-18T14:40:18Z

It it helps i added some comments (See above) to clarify the locations where the issues arose).

Comments do really help, thanks and thank you for the work!

dkgroot · 2017-12-18T14:47:11Z

@nemanja-boric-sociomantic Thanks and Thanks for the review.
I could use a little help with these three open FIXME's. It is not always clear what the intent of a unittest is. Some of them could use a couple of comments as well.

wilzbach · 2017-12-18T14:59:27Z

posix.mak

 #    pkg_add -r gmake
 # and then run as gmake rather than make.

+#QUIET:=@


What's the reason for this addition?

wilzbach · 2017-12-18T14:59:46Z

posix.mak


 # build with shared library support
 # (defaults to true on supported platforms, can be overridden w/ make SHARED=0)
-SHARED=$(if $(findstring $(OS),linux freebsd),1,)


Reverted all the posix.mak changes i made (They should not have been committed).

wilzbach · 2017-12-18T15:02:36Z

posix.mak

+#       $(GREP) -n -U -P "([ \t]$$|\r)" $(CWS_MAKEFILES) ; test "$$?" -ne 0
+#       $(GREP) -n -U -P "( $$|\r|\t)" $(NOT_MAKEFILES) ; test "$$?" -ne 0
+#endif
+checkwhitespace: $(INSTALL_DIR)/${OS}/bin${MODEL}/dmd $(TOOLS_DIR)/checkwhitespace.d 


It's not as easy as that. There's currently no logic in place that ensures that tools is actually checked out and there have been long discussion of tools being an option or required dependency for building DMD + Druntime + Phobos.
The current status quo is that it's okay to use use tools in non essential targets.
Anyhow as you said yourself, please put this in a separate PR.

dkgroot · 2017-12-18T17:10:09Z

src/rt/util/typeinfo.d

-            assert(f1  == f2);
+            assert(f1 == f2);
+            version (DragonFlyBSD) {} else   // FIXME:  //core.exception.AssertError@src/rt/util/typeinfo.d(249): unittest failure
            assert(f1 !is f2);


This is the failing assertion.
I think i might need some help from one of the 'src/rt/typeinfo/*' code owners on how to debug this.

@mathias-lang-sociomantic @Geod24 @WalterBright @redstar
Can one of you help me figure out why/how 'assert(f1 !is f2);' could be failing.
Could it be that comparing complex numbers is causing trouble. Does one of you have a good way of tracing ?

ibuclaw · 2017-12-18T18:55:16Z

src/core/stdc/errno.c

    errno = val;
    return val;
 }
+#endif


Just a nit, but use an editor that adds newline at bottom of files.

ibuclaw · 2017-12-18T18:58:09Z

src/core/stdc/fenv.d

+        };
+        _x87 x87;
+
+        uint mxcsr;


Not related to this patch, but I've just noticed that the OpenBSD branch is wrong.

n8sh · 2017-12-19T04:53:09Z

src/core/stdc/stdio.d

+else version (Solaris)
+{
+  // No unsafe pointer manipulation.
+  @trusted


It looks like this version(Solaris) block has been unintentionally duplicated.

joakim-noah · 2018-01-11T10:03:13Z

src/core/stdc/errno.c

+int * __error1(void) {
+    return __error();
+}
+#else


Thinking about this some more, the idea behind the block in core.stdc.errno below is to avoid needing this C file from druntime, as implemented in #1917. Would something like this work to replace the D declaration and alias below, so you could drop this C patch altogether?

pragma(mangle, "errno") extern int __errno; ref int errno() { return __errno;}

@joakim-noah Thanks for this much nicer solution !

Does it work, ie all the stdlib tests pass?

Yes it does. I run all the unittests before each/any commit. Thanks for this much nicer solution.

ibuclaw · 2018-01-12T23:35:21Z

src/core/stdc/errno.c


 #include <errno.h>

-


Unrelated change here.

dkgroot · 2018-01-15T22:13:01Z

Hi guys, would it help if this PR were be split by file, so that it would make a little more progress ?

joakim-noah · 2018-01-16T03:06:10Z

No, somebody just needs to make sure all review comments have been addressed and make a decision.

wilzbach · 2018-01-16T03:31:10Z

Hi guys, would it help if this PR were be split by file, so that it would make a little more progress ?
No, somebody just needs to make sure all review comments have been addressed and make a decision.

In my experience it does help. However, splitting it up by file might be a bit to granular.
Did you see @nemanja-boric-sociomantic's comment at the top:

It would be very nice if you could split this into several commits (as per your checklist in the PR)

For example, gc + rtcould one "package" PR:

this is more work entirely on your side (for reviewers smaller PRs are less "mental" work)
as @joakim-noah pointed out it's not a requirement

joakim-noah · 2018-01-16T03:58:27Z

He already split this pull into three after Nemenja said that. Don't split again, I will round up the review and try to get it going here.

dkgroot · 2018-01-24T14:38:08Z

Ping @ZombineDev @CyberShadow @klickverbot @MartinNowak
Are there any issues with this PR ? Thanks in advance for you review !

DragonFly Based CI

wilzbach

OK everything looks properly scoped in version blocks to me. Anything else blocking this PR?

dkgroot · 2018-02-09T18:55:27Z

@wilzbach Thanks a lot !
Thanks everyone for reviewing !

joakim-noah

A few nits left.

joakim-noah · 2018-02-11T09:23:47Z

posix.mak


 $(DMD):
-	make -C $(DMD_DIR)/src -f posix.mak BUILD=$(BUILD) OS=$(OS) MODEL=$(MODEL)
+	$(MAKE) -C $(DMD_DIR)/src -f posix.mak BUILD=$(BUILD) OS=$(OS) MODEL=$(MODEL)


I see one more on line 541, don't know if you care about that one.

I just moved the posix.mak file to the other PR (#2002), because it was mostly needed to compile the dragonflybsd specific files, so a better fit (and i hope it will fix the autotester issues when compiling on win32).
I did checkout the posix.mak file, but it only has 374 lines.

Whoops, checked the Phobos posix.mak for make by mistake.

Moving the make files to the other PR, seems to have been a good idea. It fixed the autotester win32 issue.
I will check the phobos posix.mak file, to see if it needs correcting (And create a PR if necessary).

joakim-noah · 2018-02-11T09:25:51Z

src/core/stdc/errno.d

 }
+else version( DragonFlyBSD )
+{
+    enum EPERM              = 1;        /// Operation not permitted


I'm guessing you didn't write these comments yourself but copied them from somewhere? If so, take out all the comments that you copied, see the linux block for an example. We don't really want to deal with copyright issues.

WIll do !
Funny how to most 'copyright bothersome os-ses' (ie: windows and osx) do have similar comments :-) Might be a good idea, to to the same for them as well.

joakim-noah · 2018-02-11T09:27:55Z

src/core/sync/mutex.d

    // not run on Android.
    version (CRuntime_Bionic) {} else
    version (CRuntime_Musl) {} else
+    version (DragonFlyBSD) {} else               // FIXME : Will have to write a test program to see what is going on here.


Still need to test this?

Tested / Checked with dragonflybsd developers / Will not be corrected
FIXME: Comment removed
Comment above the test extended/updated.

joakim-noah · 2018-02-11T09:30:37Z

src/rt/backtrace/dwarf.d


 version(CRuntime_Glibc) version = linux_or_freebsd;
 else version(FreeBSD) version = linux_or_freebsd;
+else version(DragonFlyBSD) version = linux_or_freebsd;


As long as we're here, can you change this to glibc_or_bsd? This name has annoyed me for awhile.

Used "glibc_or_bsdlibc" instead (thought that would be a better fit, and also usable by netbsd/openbsd when needed), if that is ok with you.

joakim-noah · 2018-02-11T09:31:42Z

src/rt/backtrace/elf.d


 version(linux) version = linux_or_freebsd;
 else version(FreeBSD) version = linux_or_freebsd;
+else version(DragonFlyBSD) version = linux_or_freebsd;


Same here, linux_or_bsd.

joakim-noah

Bit more to do.

joakim-noah · 2018-02-11T16:03:50Z

src/core/sync/mutex.d

-    // For example, Bionic doesn't appear to do so, so this test is
-    // not run on Android.
+    // BTW: Bionic/Musl and DragonFlyBSD don't (appear) to do so, so this test is
+    // not skipped on/for them.


Change to 'The Bionic and Musl C runtimes and DragonFly don't appear to do so, so skip this test.'

joakim-noah · 2018-02-11T16:07:41Z

src/rt/backtrace/elf.d


-version(linux) version = linux_or_freebsd;
-else version(FreeBSD) version = linux_or_freebsd;
+version(linux) version = glibc_or_bsdlibc;


This won't actually enable Glibc, as linux refers to kernel APIs. I don't know where these declarations are from, so either change the version selected to CRuntime_Glibc or the new version to linux_or_bsd.

dkgroot · 2018-02-11T20:42:06Z

@joakim-noah Thanks for your remarks/comments/review !

dkgroot requested review from Burgos, CyberShadow, DmitryOlshansky, MartinNowak, PetarKirov, WalterBright, andralex, dnadlinger, ibuclaw, jmdavis, rainers, schveiguy and wilzbach as code owners December 18, 2017 13:59

dkgroot mentioned this pull request Dec 18, 2017

Add DragonFlyBSD support for phobos dlang/phobos#5941

Merged

dkgroot commented Dec 18, 2017

View reviewed changes

dkgroot changed the title ~~Port of druntime to DragonFlyBSD~~ Add DragonFlyBSD support for druntime Dec 18, 2017

dkgroot commented Dec 18, 2017

View reviewed changes

wilzbach reviewed Dec 18, 2017

View reviewed changes

dkgroot commented Dec 18, 2017

View reviewed changes

ibuclaw reviewed Dec 18, 2017

View reviewed changes

n8sh reviewed Dec 19, 2017

View reviewed changes

dkgroot force-pushed the dragonflybsd-master branch 2 times, most recently from e279d63 to 9a82aed Compare January 3, 2018 17:29

wilzbach mentioned this pull request Jan 9, 2018

Add a new DragonFlyBSD build host braddr/d-tester#62

Closed

joakim-noah reviewed Jan 11, 2018

View reviewed changes

dkgroot force-pushed the dragonflybsd-master branch from bf847b3 to 48bb186 Compare January 12, 2018 15:43

ibuclaw reviewed Jan 12, 2018

View reviewed changes

src/core/stdc/errno.c

#include <errno.h>

Copy link

Member

ibuclaw Jan 12, 2018

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Unrelated change here.

dkgroot force-pushed the dragonflybsd-master branch from d253cf1 to ee89791 Compare January 13, 2018 16:02

dkgroot force-pushed the dragonflybsd-master branch from ee89791 to e09d5f7 Compare January 18, 2018 16:43

dkgroot mentioned this pull request Jan 18, 2018

Add DragonFlyBSD support for dmd dlang/dmd#7463

Merged

dkgroot force-pushed the dragonflybsd-master branch from e09d5f7 to 0ee0954 Compare January 21, 2018 22:24

dkgroot force-pushed the dragonflybsd-master branch from 48c5110 to 3454101 Compare January 28, 2018 18:52

dkgroot force-pushed the dragonflybsd-master branch from 3454101 to 8fe321d Compare February 7, 2018 03:08

wilzbach approved these changes Feb 9, 2018

View reviewed changes

wilzbach added the 72h no objection -> merge The PR will be merged if there are no objections raised. label Feb 9, 2018

dkgroot force-pushed the dragonflybsd-master branch 2 times, most recently from 3e5e903 to bf676c5 Compare February 10, 2018 23:36

Port of druntime to DragonFlyBSD

d580191

joakim-noah suggested changes Feb 11, 2018

View reviewed changes

dkgroot force-pushed the dragonflybsd-master branch from 1e118d9 to d580191 Compare February 11, 2018 20:39

joakim-noah approved these changes Feb 12, 2018

View reviewed changes

wilzbach added the auto-merge label Feb 12, 2018

dlang-bot merged commit 2613173 into dlang:master Feb 12, 2018

Uh oh!

Add DragonFlyBSD support for druntime #1999

Add DragonFlyBSD support for druntime #1999

Uh oh!

Conversation

dkgroot commented Dec 18, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dlang-bot commented Dec 18, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Bugzilla references

Uh oh!

nemanja-boric-sociomantic commented Dec 18, 2017

Uh oh!

dkgroot Dec 18, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dkgroot Dec 18, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dkgroot Dec 18, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dkgroot Jan 5, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dkgroot Dec 18, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dkgroot commented Dec 18, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dkgroot Dec 18, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dkgroot Dec 18, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nemanja-boric-sociomantic commented Dec 18, 2017

Uh oh!

dkgroot commented Dec 18, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dkgroot commented Dec 18, 2017 •

edited

Loading

dlang-bot commented Dec 18, 2017 •

edited

Loading

dkgroot Dec 18, 2017 •

edited

Loading

dkgroot Dec 18, 2017 •

edited

Loading

dkgroot Dec 18, 2017 •

edited

Loading

dkgroot Jan 5, 2018 •

edited

Loading

dkgroot Dec 18, 2017 •

edited

Loading

dkgroot Dec 18, 2017 •

edited

Loading

dkgroot Dec 18, 2017 •

edited

Loading

dkgroot Dec 18, 2017 •

edited

Loading

joakim-noah Jan 11, 2018 •

edited

Loading

dkgroot Jan 12, 2018 •

edited

Loading

joakim-noah commented Jan 16, 2018 •

edited

Loading

dkgroot commented Feb 9, 2018 •

edited

Loading