[tlbuild] dvi test?

[luigi scarso]

On Mon, Jan 31, 2022 at 7:38 PM Richard Koch <koch at uoregon.edu> wrote:

> Folks,
>
> Yesterday I simplified John's source file. The file I produced works if
> the source ends with "e", but crashes using R if the source ends with "el".
>
> Later John reported that this experiment failed for him. He sent me his
> version of my source and his dvi output.
>
> I typeset his version of the source, which behaved for me exactly as I
> predicted. However, his dvi file indeed did not crash. It turns out that
> our source files are exactly the same, but our dvi files are different. I
> enclose below our common source files, and the two different dvi files.
>
> Using cmp, the two dvi files differ at the top in an anticipated way,
> because the dvi contains the date and time of creation. After that they
> agree as draftcopy is processed. But at 00000330, which looks like the very
> end of draftcopy to an amateur like me, they diverge slightly.  By the end,
> one file is 4 bytes longer than the other.
>
>
You can also use dvitype .

-- 
luigi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://tug.org/pipermail/tlbuild/attachments/20220131/63d9ea09/attachment.html>