On Wed, 31 May 2017 09:46:35 -0500, Elardus Engelbrecht <[email protected]> wrote: > >>Is it possible to provide codepage information to UTOL or I have to use >>something like >>FINDREP=(INOUT=(C'a',C'A' for every character? > >>With vim on Linux, I tried: (ąćęłńóśżź). >>... then the command: >>~~~~~~~~~~~~ >>... changed it to: (ĄĆĘŁŃÓŚŻŹ). > >Is this the right translation for Radoslaw? If so, he would be glad. > On Linux, not on z/OS.
>Or you can easily use REXX function TRANSLATE to do the same [1] , but if the >input is large, then you may have some long overhead. > Does that work for MBCS? >>BTW, why is the collating sequence for En_US just plain wrong? > >En_US is not wrong, but Radoslaw needs to use his country's code page. > In fact, my Linux setting was: 572 $ locale -a C C.UTF-8 en_US.utf8 POSIX 573 $ ... evidently vim understands polish characters even in the U.S.A. But my complaint is that for LC_COLLATE=En_US, SORT on z/OS gives results very simlar to LC_COLLATE=C, whereas LC_COLLATE=En_CA gives more reasonable results. Problem does not occur on Linux. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
