Prev: Clipboard internals
Next: Non blocking FIFO list
From: Corinna Vinschen on 31 Jan 2010 08:17 Aaron Gray wrote: > Thanks for looking. No I have a X86 COFF backend already and need debugging > information in order to debug it. If I cannot get hold of the spec I will > have to reverse engineer it :( If I understand you correctly, you're just implementing the tool which produces the debug output. So, what speaks against emitting an open, well-documented format like stabs or dwarf2? Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat
From: Pavel A. on 31 Jan 2010 21:01 Also, maybe Phoenix stuff can be of help for you http://social.msdn.microsoft.com/forums/en-US/phoenix/thread/4eb6a89c-cb57-4fa7-9cb0-6f7b2d161f7e --pa "Aaron Gray" <ang.usenet(a)gmail.com> wrote in message news:#y#BVdfoKHA.1548(a)TK2MSFTNGP06.phx.gbl... > "Pavel A." <pavel_a(a)12fastmail34.fm> wrote in message > news:OzZf0JfoKHA.1544(a)TK2MSFTNGP06.phx.gbl... >> Yep. They move things around a lot. >> I could not find CV debug info under published documents >> ( http://msdn.microsoft.com/en-us/library/dd208104(PROT.10).aspx ) >> >> Maybe, you'd better use .NET IL for your code generation, rather than >> native code. >> Or generate C or asm code and then compile with supported MS compilers. >> >> Good luck, > > Thanks for looking. No I have a X86 COFF backend already and need > debugging information in order to debug it. If I cannot get hold of the > spec I will have to reverse engineer it :( > > Aaron > > >> "Aaron Gray" <ang.usenet(a)gmail.com> wrote in message >> news:O3mSF1eoKHA.1552(a)TK2MSFTNGP04.phx.gbl... >>> "Pavel A." <pavel_a(a)12fastmail34.fm> wrote in message >>> news:56C3F9AE-FF73-4D7D-A052-91E97FCF968E(a)microsoft.com... >>>> "Aaron Gray" <ang.usenet(a)gmail.com> wrote in message >>>> news:OdJkO3doKHA.1552(a)TK2MSFTNGP04.phx.gbl... >>>>> Hi, >>>>> >>>>> For want of a better place to ask. I am after the CodeView data format >>>>> specifications 4 and/or 5. I am writing an object module generator and >>>>> need to generate CodeView debugging information. >>>>> >>>> >>>> Look on MSDN. They had documented quite a lot of formats and interfaces >>>> recently. >>>> Also, try microsoft.public.windbg >>> >>> Hi Pavel, >>> >>> I could not find anything on it on MSDN :( >>> >>> Could you have a look for me as I never seem to have any luck finding >>> anything on MSDN, they just keep moving things around. >>> >>> Many thanks, >>> >>> Aaron >>> >
From: Aaron Gray on 6 Feb 2010 19:32 "Corinna Vinschen" <corinna(a)community.nospam> wrote in message news:hk3vt0$3e6$1(a)perth.hirmke.de... > Aaron Gray wrote: >> Thanks for looking. No I have a X86 COFF backend already and need >> debugging >> information in order to debug it. If I cannot get hold of the spec I will >> have to reverse engineer it :( > > If I understand you correctly, you're just implementing the tool which > produces the debug output. So, what speaks against emitting an open, > well-documented format like stabs or dwarf2? Hi Corinna, Yes I have been thinking about doing several formats rather than just one, but most Windeows debuggers work with CV so would like to get that working first. Aaron
From: Aaron Gray on 6 Feb 2010 19:38 > For want of a better place to ask. I am after the CodeView data format > specifications 4 and/or 5. I am writing an object module generator and > need to generate CodeView debugging information. I am trying to get a copy of "Visual C++ 5.0 Symbolic Debug Information Specification" from 2001 MSDN days. Many thanks in advance, Aaron
From: Pavel A. on 6 Feb 2010 21:48
"Aaron Gray" <ang.usenet(a)gmail.com> wrote in message news:#sLar34pKHA.2352(a)TK2MSFTNGP02.phx.gbl... >> For want of a better place to ask. I am after the CodeView data format >> specifications 4 and/or 5. I am writing an object module generator and >> need to generate CodeView debugging information. > > I am trying to get a copy of "Visual C++ 5.0 Symbolic Debug Information > Specification" from 2001 MSDN days. > If you wish to be compatible with _current_ MS toolchains, perhaps documentation from VC++ 5.0 days is way too old. Even VC++ 6.0 is too old. Have you looked in the Phoenix & MS Research resources? --pa |