Thanks, Mike. Here's what I'm doing. (Pardon the length of this - want to be complete.)
In a VFP8 table I'm using, there are fields of different types - specifically, numeric, date, character, and memo. The routine I use converts all of these into strings, each stored in a separate memvar. When all have been collected for each record, the '?' command outputs the memvars to an 'alternate' text file as a single line. In a sense, the strings are not concatenated in VFP, but rather output collectively as a single text line to the text file.
Example: ? cMemVar1 + cMemVar2 + cMemVar3 + cMemVar4, etc.
The '?' routine also adds some control characters at the beginning, which are used to format the string eventually (see below).
If I open the text file (in Notepad or something like it), I can see that a CR/LF is inserted into the line at a point around 240 'columns' (characters plus spaces). I'm assuming that the control characters must add some length in themselves, which may internally expand the actual byte length of the string that the file sees as totalling 255 (though this is an uneducated guess on my part). Or maybe it's just 'wrapping' the line at a space.
Now, the VFP Memo field can be short or quite long. The CR/LF seems to occur within the memo string portion of the text line, but this may be coincidence. (On a side note, if I set Memowidth to something very large, and display ('?') a long Memo field to the VFP window, it also seems to add a CR/LF at about 200 characters.)
The next step in the process is for a word processor (in this case, WordPerfect) to convert the file to its own format, read the command characters as formatting commands, and create an entry in the list this routine is building. Where the CR/LF occurred in the text file, WordPerfect inserts a Hard Return, thus causing a separate line in the WP file. This occurs whether or not the text file has been previously opened in Notepad. WP also has no apparent limit on line length, so I can only assume it's not the culprit.
So ... the offending CR/LF appears to be added in the process of exporting from VFP to the text file. This is what I'm trying to eliminate.
If it's any help, this table was originally created in FoxPro for DOS. VFP seemed to open it okay, and it operates correctly in VFP. But maybe its origin in FP for DOS has something to do with it?
Any help is greatly appreciated.
Roger