Re: Reading from the IFS

[Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx>]

> When calling the _C_IFS_fgets external procedure in RPGLE, is the data
> being read in ascii?
I'm not sure I understand this question.  Are you trying to find out if 
the fgets() API can translate ASCII to EBCDIC?  Or are you tring to find 
out if it'll fail when the data is not ASCII?  Or what are you trying to 
find out?
Basically, the IFS stores the data of a file as a collection of bytes. 
Just one big long stream of bytes.  Then, it also stores attributes of the 
file, such as the file's name, authorities, last change date, whether it 
needs to be archived, etc.  One of those attributes is a CCSID that you're 
supposed to set to tell the system what sort of data is stored in the 
file.  It might be a CCSID that corresponds to an ASCII character set.  It 
might be a CCSID that corresponds to an EBCDIC character set.  Or Unicode. 
Or whatever.
When you read the data from the file, the system is reading bytes.  It's 
just copying data from disk to memory as a bunch of bytes.
Then, depending on the options you used when you opened the file, it'll 
provide other services for you, such as translating from the file's CCSID 
(which could be ASCII or EBCDIC or Unicode) to your job's CCSID (which is 
probably EBCDIC, but could be any of a number of different flavors of 
EBCDIC).
I don't know if that means that the file is "read in ASCII".  I guess I 
don't know what the phrase "read in ASCII" means.  OBviously, if the data 
in the file is in ASCII, then bytes that happen to contain ASCII values 
are what's read, but if the file contains non-ASCII data, then that's what 
ends up being read...
Also remember, the terms "ASCII", "EBCDIC" and "Unicode" only apply to 
human-readable text.  Not all data stored in a file is text, so not 
everything can have the label ASCII, EBCDIC or Unicode applied to it/ 
(Though, text is probably the only thing you'd read with the fgets() API)

> If so, does it get automatically translated into ebcdic during the read 
> (in the RPG) or should a translate operation be performed after the read 
> to ensure it is in ebcdic before subsequent processing?
You control that by specifying whether the file is text or binary by 
passing flags on the fopen() API. (_C_IFS_fopen, that is)  You can also 
control the CCSIDs used in translation with flags you pass to the fopen() 
API.
> Second question - can the IFS store data in ebcdic or is it always in
> ascii?
Data is stored in bytes.  Again, not everything fits into the catgories of 
EBCDIC or ASCII or Unicode.  You can store any bytes you like in the IFS. 
If you tag the file with the appropriate CCSID, the system can help you 
translate it.  Otherwise, you have to handle the trnaslation yourself.
You can store any bytes you want.  ASCII, EBCDIC, anything you like.