digitalmars.D.bugs - [Issue 8328] New: Mind Blowing error (possibly in DMD or std.variant)
- d-bugmail puremagic.com (79/79) Jun 30 2012 http://d.puremagic.com/issues/show_bug.cgi?id=8328
- d-bugmail puremagic.com (10/10) Jun 30 2012 http://d.puremagic.com/issues/show_bug.cgi?id=8328
- d-bugmail puremagic.com (8/9) Jun 30 2012 http://d.puremagic.com/issues/show_bug.cgi?id=8328
- d-bugmail puremagic.com (22/22) Jun 30 2012 http://d.puremagic.com/issues/show_bug.cgi?id=8328
- d-bugmail puremagic.com (30/30) Jun 30 2012 http://d.puremagic.com/issues/show_bug.cgi?id=8328
- d-bugmail puremagic.com (9/9) Jun 30 2012 http://d.puremagic.com/issues/show_bug.cgi?id=8328
- d-bugmail puremagic.com (11/11) Jun 30 2012 http://d.puremagic.com/issues/show_bug.cgi?id=8328
http://d.puremagic.com/issues/show_bug.cgi?id=8328
Summary: Mind Blowing error (possibly in DMD or std.variant)
Product: D
Version: D2
Platform: x86_64
OS/Version: Linux
Status: NEW
Severity: critical
Priority: P2
Component: DMD
AssignedTo: nobody puremagic.com
ReportedBy: IcePapih gmail.com
---
Created an attachment (id=1122)
The minimal source tree that is giving the error
Hello lovely D folks!
I have been dealing with this incredibly strange error that I don't really even
know how to describe because I can't understand what's causing it. I can really
just show you.
Here's the quick scoop if you don't want to download stuff etc:
https://gist.github.com/3025946
Anyways here's everything I can say,
Assuming the following tree:
./
├── bin
├── lib
│ ├── import
│ │ └── gl3n
│ │ ├── color
│ │ │ └── hsv.di
│ │ ├── interpolate.di
│ │ ├── linalg.di
│ │ ├── math.di
│ │ └── util.di
│ ├── libgl3n-dmd.a
|
├── reducedt.reduced
│ ├── build (HERE IS WHERE YOU RUN THE COMPILE COMMAND..)
│ │ └── empty
│ └── torment
│ ├── component
│ │ ├── all.d
│ │ ├── component.d
│ │ └── render.d
│ ├── entity
│ │ └── camera.d
│ ├── math
│ │ └── vector.d
│ └── scene
│ └── scene.d
compiling with "dmd -g -debug -property -w -wi -I../../lib/import
-L-L../../lib/ -L-lgl3n-dmd -of../../bin/execute_me ../torment/component/all.d
../torment/component/component.d ../torment/component/render.d
../torment/scene/scene.d ../torment/math/vector.d ../torment/entity/camera"
outputs the following error:
../torment/scene/scene.d(3): Error: cannot compare
const(VariantN!(1LU,RenderComponent)[]) and
const(VariantN!(1LU,RenderComponent)[])
The line it's complaining about (3) looks like:
struct Scene {
And I don't know what to make of that. I've tried reducing (in fact the tree
you are looking at is *very* reduced) so I have the minimal case of failure
here. I have attached the tree (including lib imports and the .a file compiled
for 64bit linux) so you can try compiling for yourself. I've tried it on two
computers - both 64bit linux though so I don't know if that's got anything to
do with the problem.
At this point I don't know if this is my fault or dmd's fault or phobos'. I
assume it's the compiler but I can't know for sure. I marked this bug as
critical because it has completely stalled the development of my project for
two days now. And that does not give me a good feeling at all. I really *need*
to get on with it!
I just feel like this either is out of my hands, or just me doing something
obviously wrong and is not getting it. But nobody in irc seems to know what's
wrong either. So I am submitting this here.
Thank you!
-- Isak Andersson
--
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Jun 30 2012
d-bugmail puremagic.com
http://d.puremagic.com/issues/show_bug.cgi?id=8328
Jonathan M Davis <jmdavisProg gmx.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jmdavisProg gmx.com
PDT ---
Have you tried the latest master from github?
--
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Jun 30 2012
http://d.puremagic.com/issues/show_bug.cgi?id=8328 ---Have you tried the latest master from github?You mean DMD? No I have not. But I've considered since I really need to keep up development! Should I? Or are there any errors visible from the code I posted? -- Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email ------- You are receiving this mail because: -------
Jun 30 2012
http://d.puremagic.com/issues/show_bug.cgi?id=8328 PDT --- You should just try the latest master. I honestly don't want to spend the time digging through your code to try and figure out what's wrong with it. It's much simpler if you can present an example which exhibits the problem. But for all I know, your problem has been fixed in git already, and it'll be easier for you to test that than anyone else. You'll need to grab dmd, druntime, and Phobos from here: https://github.com/D-Programming-Language build them all, and then use those, which would involve building dmd, then making sure that your system is using the new dmd, then building druntime and Phobos and making sure that your system is using them (which involves altering your dmd.conf or sc.ini file). It's not hard to deal with (particularly in Linux thanks to symlinks) once you've got it set up, but it might be a bit of a pain to figure out exactly Alternatively, you could use dvm ( https://github.com/jacob-carlborg/dvm ) to download and build it using its compile command. I've never used dvm, but it's supposed to be quite easy to use. It's a tool for managing dmd, and will allow you to install a specific compiler version or build the latest from github. -- Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email ------- You are receiving this mail because: -------
Jun 30 2012
http://d.puremagic.com/issues/show_bug.cgi?id=8328
Vladimir Panteleev <thecybershadow gmail.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |rejects-valid
CC| |thecybershadow gmail.com
Platform|x86_64 |All
Summary|Mind Blowing error |Wrong constness of
|(possibly in DMD or |autogenerated opEquals when
|std.variant) |fields' opEquals have mixed
| |constness
OS/Version|Linux |All
Severity|critical |normal
18:07:42 PDT ---
Minimal testcase:
---
struct S1 { bool opEquals(S1 o) { return true; } }
struct S2 { bool opEquals(S2 o) const { return true; } }
struct X { S1 s1; S2 s2; }
---
The problem seems to be that DMD decides that the autogenerated opEquals should
be "const", and then fails when it tries to call S1's opEquals.
The solution for this particular problem would be to reverse the
common-denominator logic, although this won't work if immutable is thrown into
the mix. A better error message would also be an improvement.
Workaround:
Write your own non-const opEquals in X.
--
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Jun 30 2012
http://d.puremagic.com/issues/show_bug.cgi?id=8328 18:16:47 PDT --- Still happens in DMD head (7307e7190d9). The submitter's issue could also be resolved by making Variant const-correct. I don't know what that would involve; I assume the issue is non-trivial, seeing that this hasn't been done yet. -- Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email ------- You are receiving this mail because: -------
Jun 30 2012
http://d.puremagic.com/issues/show_bug.cgi?id=8328 --- Huge thanks to Vladimir for figuring this out! This means I'll be able to go to sleep without worrying about being concerned about solving this. I hope this can be resolved so that others don't fall into the same trap, because it was definitely not easy to get out of for me who is new to D! Thank you :) -- Isak Andersson -- Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email ------- You are receiving this mail because: -------
Jun 30 2012