Help recovering a recording (missing .aup)

Hi all,

I took about an hour of stereo recording today, as i’ve made many times before.
But today, while saving the project, an error message occurred and audacity crashed, without saving the .aup project file.
All the .au files are there, but it’s folders (d00 to d06) with a total 1560 files.
I couldn’t find a tool that can rebuild the .aup file, which is just an xml it seems.
If such a tool doesn’t exist, I would be able to write a tool which recreates the xml, but i need the exact sequence of the .au pieces.
for some reason, the file naming it’s not just a numerical counter, but it looks like a random hexadecimal code.
I hope there’s a meaning in that hex code and hope that i can use this for reconstructing the correct sequence.
Can anyone help me?

Thanks in advance

Additional info:
I’m on audacity 2.1.0 and noticed that every project saves .au files with the same sequence:

	<wavetrack name="Traccia audio" channel="0" linked="1" mute="0" solo="0" height="150" minimized="0" isSelected="0" rate="44100" gain="1.0" pan="0.0">
		<waveclip offset="0.00000000">
			<sequence maxsamples="262144" sampleformat="262159" numsamples="120094720">
				<waveblock start="0">
					<simpleblockfile filename="" len="262144" min="-0.679169" max="0.993652" rms="0.102954"/>
				<waveblock start="262144">
					<simpleblockfile filename="" len="262144" min="-0.771606" max="0.996765" rms="0.098826"/>
				<waveblock start="524288">
					<simpleblockfile filename="" len="262144" min="-0.870911" max="0.996002" rms="0.117586"/>
				<waveblock start="786432">
					<simpleblockfile filename="" len="262144" min="-0.808533" max="0.874542" rms="0.102034"/>

so I’m thinking of taking the project file from another project and rename it in order to recover my recording.
I guess that file names are actually not randomly generated but “procedural”

Most people should be using the latest 2.1.2 version from which has the latest bug fixes.

Someone has already written a script but we don’t usually mention it because it requires installing python and using the command-line, which most Windows users are not prepared to do.

Most users should simply follow this page

The AU files are numbered randomly. An important consideration is that there are separate AU files for the left and right channels. For the first block of audio, the left is written first, then the right, then the left for the next block and so on.

Using the link in the Manual above, there is a method that lets you access a sufficiently accurate timestamp to sort the AU files correctly so that the left channel of a stereo pair will always come before the right. Without some method to do that, the recovered recording will have some sections in which left and right are transposed.

I don’t know if there is a cross-platform method to get finest possible timestamps from the system. I think you can do it with Powershell on Windows. Unfortunately Linux EXT3 and Mac HFS+ file systems only have 1 second timestamp resolution, so on those systems it is impossible to recover AU files without having some channels transposed.