@Gábor: Thanks, I didnt know that one, thats nice and elegant.
Everyday there's something new to learn in programming, thats cool.
Edited the example to do it this way, and removed the two public
properties from moduledefinition since they dont make much sense.
public string GetBuildIndependentHash()
{
var byteArray = File.ReadAllBytes(Image.FileName);
//overwrite parts of the assembly file content byte array
//(only in memory of course) that change on every build
//with an empty bytearray and thus make it compareable
var guidSection = Image.GuidHeap.Section;
var empty = new byte[guidSection.SizeOfRawData];
Buffer.BlockCopy(empty, 0, byteArray,
Image.TimeDateStampPosition, 4);
Buffer.BlockCopy(empty, 0, byteArray,
Image.FileChecksumPosition, 4);
//MVID
Buffer.BlockCopy(empty, 0, byteArray,
(int) guidSection.PointerToRawData, (int)
guidSection.SizeOfRawData);
//Create Md5 Hash
var md5 = new MD5CryptoServiceProvider();
var result = md5.ComputeHash(byteArray);
//Stringify hash for human readability
var hashString = BitConverter.ToString(result);
return hashString;
}
Well, anyone else using this might propably prefer returning the hash
using a byte[] instead of the string, would make it more low level and
faster and stuff... I just did it this way because its how it fits my
needs best at the moment.
Best regars
Florian
On 18 Jun., 23:39, Gábor Kozár <[email protected]> wrote:
> Well done! :)
>
> Just as a side note, you could use BitConverter.ToString() when stringifying
> the md5 hash instead of that ugly loop using StringBuilder.
>
> 2011/6/18 Florian <[email protected]>
>
>
>
> > Hi everybody,
>
> > well I got it working - thanks to JBs for commenting the field
> > positions in ImageReader.
> > In case anybody needs the same really urgently it may be helpfull so I
> > am posting it here:
>
> > Addtional code in class ModuleDefinition:
>
> > public int TimeDateStampPosition { get { return
> > Image.TimeDateStampPosition; } }
> > public int FileChecksumPosition { get { return
> > Image.FileChecksumPosition; } }
>
> > public string GetBuildIndependentHash()
> > {
> > var byteArray = File.ReadAllBytes(Image.FileName);
>
> > //overwrite parts of the assembly file content byte array
> > //(only in memory of course) that change on every build
> > with
> > //an empty bytearray and thus make it compareable
> > var guidSection = Image.GuidHeap.Section;
> > var empty = new byte[guidSection.SizeOfRawData];
> > Buffer.BlockCopy(empty, 0, byteArray,
> > TimeDateStampPosition, 4);
> > Buffer.BlockCopy(empty, 0, byteArray,
> > FileChecksumPosition, 4);
> > //MVID
> > Buffer.BlockCopy(empty, 0, byteArray,
> > (int) guidSection.PointerToRawData, (int)
> > guidSection.SizeOfRawData);
>
> > var md5 = new MD5CryptoServiceProvider();
> > var result = md5.ComputeHash(byteArray);
>
> > var sb = new StringBuilder();
> > for (var i = 0; i < result.Length; i++)
> > sb.Append(result[i].ToString("X2"));
> > return sb.ToString();
> > }
>
> > The int field TimeDateStampPosition was added to the Image class and
> > filled in the ReadImage() method of ImageReader (right above the //
> > TimeDateStamp comment) using the following line:
> > image.TimeDateStampPosition = (int) BaseStream.Position;
>
> > Also the int field FileChecksumPosition was added to the Image class
> > and filled in method ReadOptionalHeaders() of class ImageReader. The
> > line stands under under the Advance(66); statement (which follows
> > the // FileChecksum comment) and looks like this:
> > image.FileChecksumPosition = (int) (BaseStream.Position-4);
>
> > For my requirements this works good enough but I would of course be
> > interested about your thoughts about this, JB.
>
> > Best regards
>
> > Florian
>
> > On 18 Jun., 23:07, Greg Young <[email protected]> wrote:
> > > I could use this code as well
>
> > > On 18 Jun 2011 16:30, "Florian" <[email protected]> wrote:
> > > Well as it seems, the locations of TimeDateStamp and FileChecksum are
> > > quite easy to find. Their positions can be found in the DOSHeader and
> > > in PEOptionalHeader, the first two headers read from a assembly file
> > > (looked the code up in Mono.Cecil.PE.ImageReader.ReadImage()).
> > > Unfortunately locating the position of MVID is more complicated.
>
> > > I am not quite sure what to do. Changing the sourcecode of ImageReader
> > > on my local machine to save the byte positions of those three fields
> > > would be the easiest thing to do. This way I could later on build a
> > > file hash that ignores these fields and reuse the filehash on
> > > different dlls / exes built on the same sources to get them compared.
>
> > > But of course every time the next Mono.Cecil version is released I
> > > would have to do the same changes again and again.
>
> > > @JB: Would you bother including such a hashing method in the
> > > Mono.Cecil sources if I can provide you a proper implementation that
> > > meets your quality expectations? I guess the requirement I have there
> > > would also be present for other people, for example for working with
> > > continuous integration build systems.
>
> > > Best Regards Florian
>
> > > On 18 Jun., 15:57, Florian <[email protected]> wrote:
>
> > > > Hi Alex,
>
> > > > thanks for the pointer, I didnt k...
> > > > > - Zitierten Text anzeigen -- Zitierten Text ausblenden -
>
> > > > - Zitierten Text anzeigen -
>
> > > --
> > > --
> > > mono-cecil
>
> > --
> > --
> > mono-cecil- Zitierten Text ausblenden -
>
> - Zitierten Text anzeigen -
--
--
mono-cecil
