nano untagged files

[gngrossi]

Will untagged files be supported?
thanks

[gngrossi]

bash-5.2$ /bin/touch OP

bash-5.2$ ls -lT OP

• untagged T=off -rw-r--r-- 1 @02858 @ISSTC00 0 Mar 13 10:10 OP

bash-5.2$ /bin/echo hello >> OP

bash-5.2$ ls -lrtT OP

• untagged T=off -rw-r--r-- 1 @02858 @ISSTC00 6 Mar 13 10:10 OP

bash-5.2$ /bin/cat OP
hello

bash-5.2$ nano OP
GNU nano 7.2 OP
^h^e^s^s^v^U

[mfsysprog]

Nano relies on the heuristic in zoslib to decide if untagged files contain ascii or ebcdic. This requires a minimum of 8 positions of data to be able to see if it is ebcdic. The only solution at the moment is to have 8 characters or more of ebcdic data in the file or tag the file as ibm-1047 before editing.
Also, at the moment untagged files are always saved as ascii. I'm planning to fix this, but it is really difficult to do that from zoslib where it would benifit all tools.

[gngrossi]

Thanks for the update.
Would you be taking the same approach as with vim? (using export __UNTAGGED_READ_MODE=ASCII)

[IgorTodorovskiIBM]

Thanks for the update. Would you be taking the same approach as with vim? (using export __UNTAGGED_READ_MODE=ASCII)

Nano should respect that environment variable since it's also built with zoslib.

We could probably improve the heuristic though. The cnt > 8 check (https://github.com/ibmruntimes/zoslib/blob/main/src/zos-char-util.cc#L541) is an arbitrary check.

[gngrossi]

Do you plan to first have nano work like vim and then improve the heuristic?
thanks