what is the format of the graphedit .grf file ?

[vhelp]

i am working on a project dealing with graphedit (.grf) files. and i recently learned how to load graphedit .grf files inside a delphi app. and now i want to extend on that by opening and parsing it, to search for the term for the output filename so that i can change it just before starting a new capture. so, say the .grf file is video.hdpvr.001.ts, i want to increment the index part at every capture, thus, 001.ts 002.ts 003.ts and so on. i would of course, only be touching the "001" part inside the .grf file. or, i would prefer to give a completely *new* output filename of any length while keeping the integrity of the graph file. this will need to be done using delphi code.

in searching around, i found this resource but it is beyond me:
http://msdn.microsoft.com/en-us/library/windows/desktop/dd388788%28v=vs.85%29.aspx

here is the output (in hex) of my current hauppauge hdpvr capture graph:

[El Heggunte]

to search for the term for the output filename so that i can change it just before starting a new capture. so, say the .grf file is video.hdpvr.001.ts, i want to increment the index part at every capture, thus, 001.ts 002.ts 003.ts and so on. i would of course, only be touching the "001" part inside the .grf file.

Start by searching for the output filename IN "Unicode Latin (UTF16-Little-Endian)"
Not sure if that weird hex-editor can do it, though

Hope this helps

[vhelp]

i made additional searches but couldn't find what i was looking for, so i guess i will look into writing a custom parser for graphedit (.grf) files for myself so that i can peek inside for certain values to make changes to and then add that to my custom capture app.

not sure what you mean by " Unicode Latin (utf16-little-endian) " but i do recall doing work in another project and little/big endian giving me some trouble.

[jagabo]

Little endian = least significant byte stored first in memory
Big endian = most significant byte store first in memory

For example the 32 bit hex value 0x12345678 (305,419,896 decimal):

stored as big endian bytes: 0x78 0x56 0x34 0x12
stored as little endian bytes: 0x12 0x34 0x56 0x78

Unicode characters are 16 bits so there are only two bytes. Intel architecture is natively little endian. Motorola (68K processors, like in old Macs) were natively big endian.

[vhelp]

ty for the explanations. i believe i found the location(s) where i need to begin my search.

but let me ask.. how / where do i determine which endian to use and ?

is it, when i determine a given location is a variable that holds a value ? and, does the type (numerical vs text vs binary) make the difference ?