-t and -i options usage about afsctool HOT 10 OPEN

The simple attempt to try to list only txt files compression status:

afsctool -t txt .

leads to program entering never ending loop and printing forever below info:

File content type: public.plain-text
File extension(s): txt
Number of HFS+/APFS compressed files: 1

Number of HFS+/APFS compressed files: 4
/disk/txt: No such file or directory
/disk/.:

It looks like arguments parser get it wrong.

from afsctool.

from afsctool.

Thank you for quick reply but not sure I understand.

I have just tried:

afsctool -c -t txt .

my understanding is that this should compress all txt files in current folder.

It does not - it enters never ending loop dispalying again and again:

/disk/-t: No such file or directory
/disk/txt: No such file or directory

from afsctool.

To be precise I think it works (txt file was compressed) but looks that -t is parsed correctly but then as you allow to specify multiple folders your code parses -t and txt as they belong to list of folders

from afsctool.

The same when I try to invert selection:

afsctool -c -i -t txt .

result - looping with message:

Totals of file content types
Number of HFS+/APFS compressed files: 0

Number of HFS+/APFS compressed files: 1
/disk/-i: No such file or directory
/disk/-t: No such file or directory
/disk/txt: No such file or directory
/disk/.:

Bottom line is that -i and -t options parsing is broken

from afsctool.

from afsctool.

This is why I put in plenty of conditional code to be able to test everything but the (de)compression on Linux... I can reproduce your issue, and will try to fix it soonish.

This is great news! Thank you.

-t takes a single argument only; it has to be specified multiple times if you want to declare multiple file types/extensions.

This is fine. As long as documented. But this is detail.

In the meantime, please use an external solution (find?). Or trust that the kind of files you want to exclude will be rejected quickly because a chunk that "compresses" to a larger size is encountered very quickly (so don't use the -L option nor LZVN compression).

yes I could do this with 'find' but then I lose parallel processing - so it becomes real slow. In my case I want to process many TB dataset which can contain hundreds thousands of files. I know that many of these files cant be compressed anyway so I want to speed up all process by excluding them. Obvious suspects like jpg, zip, gz etc.

from afsctool.

and BTW afsctool is fantastic - extremely useful tool

from afsctool.

from afsctool.

yes I could do this with 'find' but then I lose parallel processing - so it becomes real slow.
No! If you activate parallel processing with -j or -J, the files specified on the commandline will be added to a queue (regardless of whether you specify the folders they are in, or individual files). That queue is than emptied by worker threads.

ok so it is some workaround - thank you for clarification

but clearly you suggest something else than:

find pattern -exec afsctool

??

from afsctool.