Custom Query (44 matches)

Passing a null string to @eawrite is supposed to remove the EA. Instead, it keeps the EA and sets the length to 0.

set ret=%@eawrite[file,.SUBJECT,the subject] set ret=%@eawrite[file,.SUBJECT,] <--4OS2 thinks the EA is gone

Rexx shows the whole EA--

set ret=%@rexx[call SysGetEA 'file','.SUBJECT','EAval' ; parse var EAVal EAhdr +2 EAlen +2 EAstr ; say ' [REXX] EA == "'c2x(EAhdr)'" "'c2x(reverse(EAlen))'" "'EAstr'"']

I noticed it because I was getting error messages from FC/2 about invalid EAs when copying files from an NDFS drive. I don't know the 4OS2 EAs were the cause, but FC/2 has not complained since I cleaned out all the empty EAs.

I suspect the fix would be just a matter of passing 0 to DosSetPathInfo? instead of the length of the empty EA header.

A workaround is to use Rexx to delete the EA--

set ret=%@rexx[call SysPutEA "%filename", %EAname,]

Like many people, I do not use spaces consistently. Using @eawrite, it is all too easy to create the wrong EA by putting a space after the comma--

%@eawrite[file, .SUBJECT,text] %@earead[file,.SUBJECT]

won't work, because there is no ".SUBJECT" EA; the EA is " .SUBJECT"

(why did IBM allow leading and trailing spaces in an EA name anyway?)

@words[a] returns 1 @words[] returns 1

Claiming there is one word in an empty string is contrary to common sense, useless, and makes for unnecessary complications in loops. This has been around for years in 4DOS and 4OS2, but I still think its wrong.