[thornik]

Hello, Rainer and everybody!
Lately tried to assemble VD from sources (just taken from SVN) and it failed. Hope somebody have a time to fix it or at least point where to dig. Additionally I wanna share another ideas.

1. It could be nice to put directly in README or whatever, which variables in Makefile(s) must be changed. Mine is:
   
   
   
   • For Makefile: DMD2, COFFIMPLIB, WINSDK, VSISDK
     
   • For sdk\Makefile: D, IMPLIB, COFFIMPLIB, WINSDK
     
   
   (Cannot they use the same common vars?)
   
2. NMake goes OK, but these warnings soon will be issues:
   

   Quote:

   c:\dmd2\windows\bin\dmd.exe -map bin\Release\vsi2d.exe.map -ofbin\Release\vsi2d.exe -version=vsi c2d\idl2d.d c2d\tokenizer.d c2d\tokutil.d c2d\dgutil.d c2d\dlist.d Warning: As of Phobos 2.054, std.ctype has been scheduled for deprecation in January 2012. Please use std.ascii instead. Warning: As of Phobos 2.054, std.regexp has been scheduled for deprecation in August 2011. Please use std.regex instead. Warning: As of Phobos 2.054, std.string.toupper has been scheduled for deprecation in January 2012. Please use std.string.toUpper instead. Warning: As of Phobos 2.054, std.string.tolower has been scheduled for deprecation in January 2012. Please use std.string.toLower instead. Warning: As of Phobos 2.054, std.string.stripl has been scheduled for deprecation in January 2012. Please use std.string.stripLeft instead. Warning: As of Phobos 2.052, std.string.repeat has been scheduled for deprecation in August 2011. Please use std.array.replicate instead.

   
   Fortunately, it's just deprecations, i.e. definitely have a solution.
   
3. After "Converting files" stage we fail:
   

   Quote:

   NMAKE : fatal error U1073: don't know how to make 'rg\d\vs2010_addin\sdk\vsi\dte80.d'

   
   Problem can be quickly solved by removing last 3 lined from 'VS2010_addin\sdk\vsi_sources', but what they do in public repository? :(
   
4. Following re-NMake doesn't go so far:
   

   Quote:

   ..\sdk\vsi\dte80a.d(3): ';' expected following module declaration instead of ..

   
   Now we have to fix mentioned dte80a.d: Just replace two dots to one.
   
5. And finally, make procedure falls to knee without any hope:
   

   Quote:

   "c:\dmd2\windows\bin\dmd.exe" -of..\bin\Release\vdc.lib -O -inline -release -lib -g -d -L/DELEXECUTABLE -L/LINENUMBERS -I.. -J.. -Jresources @C:\WINDOWS\TEMP\nm83.tmp Internal error: ..\ztc\el.c 904

   
   Mentioned 'el.c' can be easy found in D sources, system falls on this strange assert:
   

   Quote:

   switch (tybasic(t)) { case TYfloat: case TYifloat: e->EV.Vfloat = val; break; case TYdouble: case TYidouble: e->EV.Vdouble = val; break; case TYldouble: case TYildouble: e->EV.Vldouble = val; break; case TYcfloat: case TYcdouble: case TYcldouble: assert(0); break; default: e->EV.Vllong = val; break; }

   
   Cannot offer any idea here, but we again have no smooth building of VisualD.
   

My main PIA is the last error, which doesn't point anywhere. May be even Walter required here?

[sagitario]

Hi,

I just happen to have committed the files for a new release, which builds for dmd 2.054 and dmd 2.055beta (the latest dmd from github).
You could give them a try, but it seems that the problems you are running into are different.