[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: PSPP-BUG: PSPP 1.2.0 - A few bugs in the document facility

From: Ben Pfaff
Subject: Re: PSPP-BUG: PSPP 1.2.0 - A few bugs in the document facility
Date: Fri, 22 Feb 2019 18:54:51 -0800
User-agent: Mutt/1.10.1 (2018-07-13)

On Tue, Feb 19, 2019 at 09:11:14AM +0000, Nolan Void wrote:
> Manual section "16.1 ADD DOCUMENT" says that document text saved to
> system files can be shown with SYSFILE INFO. However, SYSFILE INFO
> does not show document text, nor does "9.9 SYSFILE INFO" indicate
> any way to make it do so.

I've fixed SYSFILE INFO so that it now displays documents.

> Manual section "16.5 DOCUMENT" says that you can terminate document
> entry "with a dot or a blank line". This is vague terminology for
> *nix systems, where there's a convention, going back to "ed", for
> terminating appended text with a line containing _only_ a dot. What's
> actually going on in PSPP is that entering a blank line does _not_
> terminate input, whereas any line that ends with a dot _does_
> terminate input. These is actually the way that SPSS behaves, so the
> behavior is right, but the manual is wrong and incomplete.

I fixed the documentation for DOCUMENT:

        * It is more specific about the dot.

        * It no longer claims that a blank line ends the document.

> For some reason, every document line is being displayed by DISPLAY
> DOCUMENTS as a separate paragraph, regardless of whether or not one
> enters a blank line (which in SPSS is the documented way to start a
> new paragraph). This happens with both DOCUMENT and with ADD

I think I changed this behavior a while back.  If you find that it's
still the case, then I'll need a better description of the problem.

> Finally, I don't know as this is a bug, per se, but I notice that
> the 80-byte line restriction on document text doesn't seem to exist
> in SPSS, as of Version 25.0.0. Is there some reason for hanging on
> to this limitation in PSPP?

The SPSS system file format represents documents as a number N followed
by 80*N bytes.  That is, every line in a set of documents is represented
by exactly 80 bytes, padded on the right with spaces, and there is no
way to indicate longer (or shorter) lines.

I don't know what SPSS does if a line is too long to fit: maybe it
truncates the long line, maybe it wraps it to a second line (and leaves
a ragged margin), possibly it even ignores line boundaries and fills
paragraphs across 80-column lines.  Maybe someone will eventually do an
experiment to enlighten me.  If they do, I'll probably implement the
same behavior in PSPP.  Until then, it seems reasonable to me to just
keep the 80-byte limit.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]