March 22, 2012 Re: [D-runtime] Arghhh, DLLs broken on Win32 *again*! | ||||
---|---|---|---|---|
| ||||
Posted in reply to Martin Nowak Attachments:
| http://d.puremagic.com/issues/show_bug.cgi?id=7749 Added a couple days ago. -Steve >________________________________ > From: Martin Nowak <dawg@dawgfoto.de> >To: D's runtime library developers list <d-runtime@puremagic.com> >Sent: Wednesday, March 21, 2012 7:40 PM >Subject: Re: [D-runtime] Arghhh, DLLs broken on Win32 *again*! > >> That may be, but until a better approach is made we must not break existing code. > >Yes, and the best approach is to nail it into the test suite. >_______________________________________________ >D-runtime mailing list >D-runtime@puremagic.com >http://lists.puremagic.com/mailman/listinfo/d-runtime > > > |
March 22, 2012 Re: [D-runtime] Arghhh, DLLs broken on Win32 *again*! | ||||
---|---|---|---|---|
| ||||
Posted in reply to Steve Schveighoffer Attachments:
| What needs to be done to get the current DLL tests to pass?
On 3/22/2012 12:52 PM, Steve Schveighoffer wrote:
> http://d.puremagic.com/issues/show_bug.cgi?id=7749
>
> Added a couple days ago.
>
> -Steve
>
> --------------------------------------------------------------------------------
> *From:* Martin Nowak <dawg@dawgfoto.de>
> *To:* D's runtime library developers list <d-runtime@puremagic.com>
> *Sent:* Wednesday, March 21, 2012 7:40 PM
> *Subject:* Re: [D-runtime] Arghhh, DLLs broken on Win32 *again*!
>
> > That may be, but until a better approach is made we must not break
> existing code.
>
> Yes, and the best approach is to nail it into the test suite.
> _______________________________________________
> D-runtime mailing list
> D-runtime@puremagic.com <mailto:D-runtime@puremagic.com>
> http://lists.puremagic.com/mailman/listinfo/d-runtime
>
>
>
> _______________________________________________
> D-runtime mailing list
> D-runtime@puremagic.com
> http://lists.puremagic.com/mailman/listinfo/d-runtime
|
March 23, 2012 Re: [D-runtime] Arghhh, DLLs broken on Win32 *again*! | ||||
---|---|---|---|---|
| ||||
Posted in reply to Walter Bright Attachments:
| From what I understand, it's an issue with the gcstub library not building or not being rebuilt. None of the changes in Martin's patch should affect runtime, they are implementation details. Except he removes the isRunning function from the public API. So it's really an API change that is causing it not to build. I remember when I added array append caching (and safety to array slice appending), I needed a function which not only allocated memory, it returned the block info. So I added that to druntime's normal GC. But I didn't update the gcstub, and it broke the DLL build. See here: http://forum.dlang.org/post/4B8352FC.2030408@digitalmars.com -Steve >________________________________ > From: Walter Bright <walter@digitalmars.com> >To: d-runtime@puremagic.com >Sent: Thursday, March 22, 2012 4:24 PM >Subject: Re: [D-runtime] Arghhh, DLLs broken on Win32 *again*! > > >What needs to be done to get the current DLL tests to pass? > >On 3/22/2012 12:52 PM, Steve Schveighoffer wrote: http://d.puremagic.com/issues/show_bug.cgi?id=7749 >> >> >>Added a couple days ago. >> >> >>-Steve >> >> >> >>>________________________________ >>> From: Martin Nowak <dawg@dawgfoto.de> >>>To: D's runtime library developers list <d-runtime@puremagic.com> >>>Sent: Wednesday, March 21, 2012 7:40 PM >>>Subject: Re: [D-runtime] Arghhh, DLLs broken on Win32 *again*! >>> >>>> That may be, but until a better approach is made we must not break existing code. >>> >>>Yes, and the best approach is to nail it into the test suite. >>>_______________________________________________ >>>D-runtime mailing list >>>D-runtime@puremagic.com >>>http://lists.puremagic.com/mailman/listinfo/d-runtime >>> >>> >>> >> _______________________________________________ D-runtime mailing list D-runtime@puremagic.com http://lists.puremagic.com/mailman/listinfo/d-runtime >_______________________________________________ >D-runtime mailing list >D-runtime@puremagic.com >http://lists.puremagic.com/mailman/listinfo/d-runtime > > > |
March 23, 2012 Re: [D-runtime] Arghhh, DLLs broken on Win32 *again*! | ||||
---|---|---|---|---|
| ||||
Posted in reply to Steve Schveighoffer Attachments:
| Forgot to complete my point... What I was saying is, we do not need to be on Windows to test that gcstub still builds and links. I think part of druntime's unit tests should test that gcstub is up to date. -Steve >________________________________ > From: Steve Schveighoffer <schveiguy@yahoo.com> >To: D's runtime library developers list <d-runtime@puremagic.com> >Sent: Friday, March 23, 2012 11:34 AM >Subject: Re: [D-runtime] Arghhh, DLLs broken on Win32 *again*! > > >From what I understand, it's an issue with the gcstub library not building or not being rebuilt. > > >None of the changes in Martin's patch should affect runtime, they are implementation details. Except he removes the isRunning function from the public API. So it's really an API change that is causing it not to build. > > >I remember when I added array append caching (and safety to array slice appending), I needed a function which not only allocated memory, it returned the block info. So I added that to druntime's normal GC. But I didn't update the gcstub, and it broke the DLL build. See here: http://forum.dlang.org/post/4B8352FC.2030408@digitalmars.com > > >-Steve > > > > > > > > >>________________________________ >> From: Walter Bright <walter@digitalmars.com> >>To: d-runtime@puremagic.com >>Sent: Thursday, March 22, 2012 4:24 PM >>Subject: Re: [D-runtime] Arghhh, DLLs broken on Win32 *again*! >> >> >>What needs to be done to get the current DLL tests to pass? >> >>On 3/22/2012 12:52 PM, Steve Schveighoffer wrote: http://d.puremagic.com/issues/show_bug.cgi?id=7749 >>> >>> >>>Added a couple days ago. >>> >>> >>>-Steve >>> >>> >>> >>>>________________________________ >>>> From: Martin Nowak <dawg@dawgfoto.de> >>>>To: D's runtime library developers list <d-runtime@puremagic.com> >>>>Sent: Wednesday, March 21, 2012 7:40 PM >>>>Subject: Re: [D-runtime] Arghhh, DLLs broken on Win32 *again*! >>>> >>>>> That may be, but until a better approach is made we must not break existing code. >>>> >>>>Yes, and the best approach is to nail it into the test suite. >>>>_______________________________________________ >>>>D-runtime mailing list >>>>D-runtime@puremagic.com >>>>http://lists.puremagic.com/mailman/listinfo/d-runtime >>>> >>>> >>>> >>> _______________________________________________ D-runtime mailing list D-runtime@puremagic.com http://lists.puremagic.com/mailman/listinfo/d-runtime >>_______________________________________________ >>D-runtime mailing list >>D-runtime@puremagic.com >>http://lists.puremagic.com/mailman/listinfo/d-runtime >> >> >> > > |
March 25, 2012 Re: [D-runtime] Arghhh, DLLs broken on Win32 *again*! | ||||
---|---|---|---|---|
| ||||
Posted in reply to Walter Bright | On 3/21/2012 6:29 PM, Walter Bright wrote: > Nobody liked my .sh version of the win32 tests :-) I guess that's because you are using your home-grown version of shell.exe. I have made a few changes to the tests: - static linking of DLLs work if you export the module info symbol (GDC seems to do this by default) - dynamic linking, but still importing the d file, can sometimes work if you use a di file instead without further dependencies and without static constructor/destructor. Here's a pull request to fix the building and running of the tests: https://github.com/D-Programming-Language/dmd/pull/841 but - you still need the digital mars shell - you have to add implib.exe to the build server (I would not mind adding it together with coffimplib.exe to the dmd distrbution). The optlink /IMPLIB is too buggy to be used. - you should change the paths at the top of runtests.sh > The documentation here: > > http://dlang.org/dll.html > > must work. The test cases, druntime, and the documentation must all be > in sync. As said some times before: the current examples of sharing the GC between different DLLs are very limited, they don't support any other sharing (threads, files, C-heap) and do not support multi-threading. I would suggest to either nuke that part of the documentation or put a big warning to this effect at the top of the respective paragraph. Rainer _______________________________________________ D-runtime mailing list D-runtime@puremagic.com http://lists.puremagic.com/mailman/listinfo/d-runtime |
Copyright © 1999-2021 by the D Language Foundation