Possible bug when writing to a full partition

[Bosanek]

I have discovered a very nasty bug in Avidemux v2.5.x (it is present in v2.5.3 and v2.5.6)! I have not tried it with Avidemux v2.6.8, because my nerves have already got irritated by it. I suggest that one of the testers find it out.


In essence: Avidemux does not check if there is enough free space on the destination drive when writing an output file, and will not react even after it fills the last free byte on the target drive - it will simply continue writing the output file to the destination drive, into God knows which sectors ....

I was using direct stream copy with A-B selection. The issue is repetable.

[Jan Gruuthuse]

There are no testers, only users.
And there is nothing to test with is it?
Also getting very irritated. And I never found any nasty bugs in avidemux, just some issues that were solved.

[AQUAR]

An avidemux user here recently implied with an LOL that I am a twit - now that's irritating!

Programs usually have limited systems disaster proofing, as that resides in the domain of the OS.
Does not the OS provide a warning when a HDD partition is nearing full capacity?

[Bosanek]

This bug directly leads to data loss.
I had produced two output files, and erased the source files after that, only to find out later that the two output files were garbled, because they could not fit on the destination drive, but Avidemux kept writing them anyway (in a loop) after the partition got full.

I suppose that it would be reasonable to be a bit unnerved after you get in this kind of situation.

But all emotions and subjective perceptions aside, my goal of writing this information here was not to blame or attack anybody. It was to inform the public, and especially the programmers. I hope that I made my contribution to the common goal of achieving a quality software.

OS might or might not provide a warning. That does not redeem the application for its part of responsibility. Every application which writes files to a storage medium has to stop as soon as the destination drive is full, and inform the user that the write operation has failed due to output drive being full. That is the minimum level of robustness which is always required. This does not need any special explanation or elaboration.


Kind regards to all!

[mean]

It depends on the output format
Some are detecting the disk full issue, some do not

[AQUAR]

@ Bosanek

Best protection is to be savy in knowing the status of your hardware.

I've written to HDD's with bad sectors and nearly lost data.
The program I used to write this data could have prevented that, based on information of the HDD's SMART's.
To do that you need a standardised fault level model and associated handshaking between the OS, the HDD and the Program.
Such standards don't exist, especially for multiplatforms.

In practise, programs write indirectly to HDD via the OS.
And the OS and controller software do a lot of damage control in the background (eg no lost data due to reallocated sectors).

All thing aside - I too would like your minimum level of robustness - ideally a better OS.