Text differences (Was: The VAX is running)

[Ethan Dicks]

On Mon, Apr 6, 2009 at 1:20 PM, David Coolbear <david at thecoolbears.org> wrote:
>>> Source files should not be giving you a 'record problem' (since they're
>>> text) - but there is another potential issue.
>>> ... discovered that there was a 'spurious' CR at the end of each line.  I
>>> removed those and everything compiled and linked.  Take a look at the
>>> line-level translation that may be happening.  -- Ian
>
> This is almost exactly what I was seeing. I was able to resolve the problem
> by NOT transferring the source files with FTP's binary image mode turned on.

Ah... there you go... text files must be transmitted between
architectures and platforms in text mode to allow the tool to Do The
Right Thing(tm).  Both ftp servers and clients can communicate in an
RFC-described manner, but each end is ultimately responsible for
knowing what a text file is on their own systems and translating as
required.

VMS itself doesn't represent files internally (i.e., on the
filesystem) as streams of bytes with CR or LF or a combination at the
end of a line.  An RMS text file is a series of variable length
records that are converted to things like null terminated strings,
etc., by, IIRC, a combination of RMS and the run-time-library for your
application's language (VAXCRTL.EXE, etc.)

This is conceptually quite different from how DOS and UNIX and
AmigaDOS, etc., etc., do things.

> I was then able to compile the source and I'm now in a maze of twisty
> passages all alike.

Excellent.  Congrats!

-ethan