# [texhax] could someone give me an explanation as to why fontsize fails with an overfull box

Peter Flynn peter at silmaril.ie
Sat Jan 26 23:48:48 CET 2019

On 26/01/2019 13:19, Carlos wrote:
>
> Peter, thank you for your timely response in breaking down the
> explanation for me... I was going to reply earlier, but I got caught up
> in a few errands and now I'm getting an error with the same file but
> running under plain TeX- admittedly thankfully to the DVI previewer, or
> else I wouldn't have noticed it. The error came  with the following lines
>
> xdvi: File corrupted (will try to reload) ...

I haven't used xdvi for many years but I think that is purely
incidental, caused by xdvi trying to re-read the document to update the
display, and accidentally accessing the file while it was being written.
I think you can ignore it.

> The two plain TeX files have the same contents.

Then they will produce identical results. If the results are not
identical, then the files are not the same. You can test this with the
diff and file commands.

> but even then the results were entirely different.

So the files are not the same.

> Why did I want plain TeX to handle it instead? Because someone asked me
> why if it may very well the case that the spacing issue is the way to
> loosen it, the first part of the log wasn't reflecting the other overfull
> boxes.

I am still not clear what the original problem was. As far as I
remember, it was a narrow setting that (naturally) had spacing problems.
I think several people have suggested solutions for this.

[...]
> But putting aside the above, and going by the plain TeX file in
> question, what I currently have is
>
> \tracingall
> \hsize=-2in

What is a negative \hsize for?

> The reason is that upon restarting the computer, there are also
> other programs that start running on the background, among them
> Antivirus programs and whatever other programs Windows deems
> necessary for the functionality of the system.

Those should not affect a TeX system, as it does not share resources
with any of them.

> And yet the shipped out page's results are so distant from one another,
> that I couldn't help but wondering if it is the most current
> distribution the underlying factor. I still find it troublesome that a
> newly edited file would provoked such pronounced differences.

As I don't actually understand the problem, can you post a minimal
example that shows it not working (ie two files), with all the
irrelevant material removed.

To do this, take your existing files and remove all the irrelevant stuff
(\tracingall, \hsize, \font, etc) one by one until you have the BARE
MINIMUM that illustrates the problem, and nothing more.

Then perhaps we can see what the problem really is. Guessing doesn't work.

///Peter