Implemented minimal fix for shift > 32 issue of 4GB+ data

[LordKiRon]

This is the fix for only one issue that can happens when processing 4GB+ data .
It does not fix working with such files but it does removes one of the barriers.
Since doing it otherwise is defined problematic up to impossible, let's solve the problem by teaspoon steps, one after another.

In case of my database (67GB , contains 4GB+ versioned files) it result in clone/fetch still fail but much later :)

[dscho]

@LordKiRon I would like to upstream this fix, but upstream Git requires a real name in the Signed-off-by line. Would you mind providing that?

[LordKiRon]

@LordKiRon I would like to upstream this fix, but upstream Git requires a real name in the Signed-off-by line. Would you mind providing that?

Sorry , but this creates ties between my real life and public activity I would like to avoid. Not that I am doing something illigal or unpropriete in either area :) and I guess after some digging on the net you might even connect my nickname to my real name, but its a different than creating real connection.
I preffer to keep "internet life" and personal life separate.

[dscho]

@LordKiRon I would like to upstream this fix, but upstream Git requires a real name in the Signed-off-by line. Would you mind providing that?

Sorry , but this creates ties between my real life and public activity I would like to avoid. Not that I am doing something illigal or unpropriete in either area :) and I guess after some digging on the net you might even connect my nickname to my real name, but its a different than creating real connection. I preffer to keep "internet life" and personal life separate.

Understood. I'll take ownership, publicly, then, documenting that you are the actual author but prefer to stay pseudonymous.