Comintern

Wildly ambitious maybe.

Right - you'd need the PE headers, etc. The exports table would be the one I'd worry was missing.

It's worth a shot I guess - I'm guessing it's already linked there though.

Maybe, but I'm guessing it would be fairly difficult - aren't they stored as pcode in the documents?

And if you do an old-school ctrl-F, there's only one, right?

Oh yeah. It's one you opened - #4848. :-P

@FreeMan There's an open issue for that IIR.

It would at least narrow down whether or not the linker is missing something.

This is a good starting place: x64dbg.com/#start

Yeah, this is link attempt 2.

I may need to actually try to load it from a VBA project instead of trying to just load the .dll in a dummy process.

At least the functions appear to line up with the symbols AFAICT.

Well, looking at it, the base address seems right.

The failing offset is 00007FFD5579C933

00007FFD5579C92C                                | E8 FFE4FCFF              | call <ntdll.ZwQueryInformationThread>                    |
00007FFD5579C931                                | 85C0                     | test eax,eax                                             |
00007FFD5579C933                                | 78 0A                    | js ntdll.7FFD5579C93F                                    |
00007FFD5579C935                                | 807C24 40 00             | cmp byte ptr ss:[rsp+40],0                               |

It looks like it might have something to do with the base addressing. It stops here:

I doubt that the COFF headers have changed for a long time.

It's failing in ntdll on LdrpInitializeProcess, but I can't tell if it's a linking problem or not - the last error that was set was STATUS_OBJECT_NAME_NOT_FOUND.

@mansellan Looks like the 64-bit edit confused the linker.

Worked perfectly.

@mansellan Yep. Pulled it this morning on my way out the door for work. I was going to start poking at it when I got home.

(or you can build your own by tacking #L42 on the end, where 42 is the line of code).

@PeterMTaylor You can get an anchor on GH by clicking on the line number.

It basically just removes them all and re-adds them in the selected order.

Probably during. The code isn't horribly sophisticated ATM.

> Usage sample provided by Rubberduck VBA contributors. Written by the VBExperts.

> Rubberduck is the premier VBE add-in. Click here to install.

I thought we did listen to ReferenceEvents (for the benefit of the parser). Could be wrong though.

> UtterAccess is the premier Microsoft Access wiki and help forum. Click here to join.

WTF? MS is letting people tag ads on docs now?

Access.References? Is that different than the VBIDE.ReferencesEvents?

I can't tell from the log if it was the first reference it tried to remove that threw or a subsequent one.

@this It is running class code apparently (from the screenshot). I was thinking that something might have been holding a WithEvents reference.

I should submit a PR to remove all the superfluous statement separators in VBA-Web. Tim should know better than that crap.

VBE seems to think it was. I wonder if it's related to that other bizarro Access form thing.

Oh wait, I'm not even with next at work...

If utc_NegativeOffset Then: utc_Offset = -utc_Offset

Well crap - apparently I missed a case with the last parser update. This is failing:

2019-03-18 09:16:18.0913;ERROR-2.4.0.4613;Rubberduck.UI.AddRemoveReferences.AddRemoveReferencesViewModel;System.Runtime.InteropServices.COMException (0x800ADF09): Can't execute code in break mode
   at Microsoft.Vbe.Interop._References.Remove(Reference Reference)
   at Rubberduck.AddRemoveReferences.ReferenceReconciler.ReconcileReferences(IAddRemoveReferencesModel model, List`1 allReferences) in C:\projects\rubberduck\Rubberduck.Core\AddRemoveReferences\ReferenceReconciler.cs:line 75
   at Rubberduck.UI.AddRemoveReferences.AddRemoveReferencesViewModel.ExecuteApplyCommand(Object parameter) i…

This seems to think you were in break mode:

Yeah, I'm checking the log on that one. I'm assuming the reference counts are correct on the RD toolbar?

I'm mainly curious why the parser throws on it.

ParseType: Main parse;Rubberduck.Parsing.VBA.Parsing.ParsingExceptions.MainParseSyntaxErrorException: mismatched input 'End Function' expecting {ABS, ANY, ARRAY, CBOOL, CBYTE, CCUR, CDATE, CDBL, CDEC, CINT, CLNG, CLNGLNG, CLNGPTR, CSNG, CSTR, CURRENCY, CVAR, CVERR, DEBUG, DOEVENTS, FIX, INPUTB, INT, LBOUND, LEN, LENB, LONGLONG, LONGPTR, MIDB, PSET, SGN, UBOUND, ACCESS, ADDRESSOF, ALIAS, AND, ATTRIBUTE, APPEND, AS, BEGINPROPERTY, BEGIN, BINARY, BOOLEAN, BYVAL, BYREF, BYTE, CLASS, CLOSE, DATABASE, DATE, DOUBLE, ELSE, ELSEIF, END_IF, ENDPROPERTY, END, EQV, ERROR, FALSE, GET, IMP, IN, INPUT, IS…

What's in your WebHelpers component at line 2947? There's a parse error in the log for that function:

FWIW, that's pretty much all the VBE OM - the references dialog just calls References.Remove there.

@FreeMan I'm completely stumped. Only thing that I can think of is that there was a running VBA process, and forcing a recompile triggered it to discard an object or something.

That sounds more like a Windows problem than a Linux one.

No, Mint - Arch just has better online docs.

$ mkdir Testing
$ mkdir testing
mkdir: cannot create directory ‘testing’: File exists

@FreeMan It won't let you do that if the underlying file system doesn't support it.