I should have noticed this when reviewing the module API patch: MinidumpModule returns the debug_identifier as an uppercase hex GUID on Windows, but the symupload tool sends a lowercase GUID (guid_string.cc). They should be the same, though I don't have a particular preference.

There's actually more... it seems like symupload should return an uppercase GUID, no hyphens, plus the age. That way it's consistent with the debug_identifier() in the processor.

I had to do some perl-munging to work around this in my Mozilla work. dump_syms uses lowercase with hyphens, but the symbol store uses uppercase, no hyphens. If you're going to standardize, I'd go with the symbol store representation, since that's compatible with the MS symbol server.

Right...I agree...

I chose the uppercase dashless-delimiterless format in the processor for the reason Ted gave.

We should probably always pass these strings around in the exact format that we use for identifiers in the processor library. This probably means getting rid of the "guid" and "age" parameters in symupload and just passing "identifier". We'd merge the guid and age fields that we output as the first MODULE line for .sym files in the same way. As a bonus, we can kill a parameter from GetModuleInfo.

For symupload, how about we rename "module" to "debug_file" and merge "guid" and "age" into "debug_identifier", to match the names we're using in the processor?

Fixed in revision [77].