[Omake] Re: Set of changed files

[Luis O'Shea]

>> Isn't one of the complications that there isn't necessarily a fixpoint
>> for latex?  I believe that you can get ping-pong effects since the
>> layout uses information from the last run only.

You are correct.  See http://broken-latex.jottit.com/ for example. 
Presumably omake, if it were to try to tackle fixed points, would need 
to place an upper bound on how many times it would traverse any one cycle.

> Theoretically - yes, you can have ping-pong. But practically speaking I 
> do not think that this actually ever happens in any real-world documents.

This is my experience also.  However, what *can* happen, is that it can 
take more than two latex runs to stabilize.  It can also take more than 
one bibtex runs (cross references can cause this).  In addition, since 
the lack of stability can be the result of the activities of a number of 
different latex packages, it can be hard to know when you have run latex 
(or bibtex, or makeindex) sufficient times.  In fact, I think the only 
systematic way to know when you are done is to look at the MD5s of the 
inputs to latex and friends and wait for them to stabilize.

One wrinkle to seeking a fixed point for latex documents is that it can 
take a little longer to stabilze depending on the order in which you run 
latex and bibtex.  Furthermore, you don't need perverse documents for 
this to happen.

Luis