I had this code:
99 \placeformula\startformula
100 \eqalignno{
101 >> H(I,J) &= H(I) + H(J|I)\qquad&\hbox{\ineq[eq:equivoc]}\cr
102 W(I,J) &= H(J) - H(J|I)\qquad&\hbox{\ineq[eq:wedinfo]}\cr
103 }
104 \stopformula
It is better to use the construction such as:
\placeformula
\startformula
\startalign
\NC H(I,J) \NC = H(I) + H(J|I) \NR[eq:equivoc]
\NC W(I,J) \NC = H(J) - H(J|I) \NR[eq:wedinfo]
\stopalign
\stopformula
Below I illustrate why this is not better, and I did not find a way how to remedy its effect.
This is the original produced by the \eqalignno. See the active links to previous equations in green, produced by my \ineq-macro.
The equation numbers here are from the 7th chapter, in the latter (shortened) examples the chapter number will be 1.
The suggested code gives:
Wrong because the numbers should refer to equations 1.21 and 1.28 instead of numbering forward.
Putting my \ineq-macro there
\NC W(I,J) \NC H(J) - H(J|I)\NR\hbox{\ineq[eq:wedinfo]}
gives an equally unacceptable result:
Hi Hans,
Since a few weeks (or months?…) ago the use of \eqno and \eqalignno has been changed in mkiv.
A remark afterwards I really regret making. But I cannot understand why \eqalignno and friends have been changed from their original behaviour. It breaks older products of mine without warning. In fact I wouldn't have spotted this, had not a colleague drawn
my attention to it. Is this the only math macro behaving differently, giving a TeX error? Are there others?
I cannot but ask most strongly not to let mkiv deviate in this manner from basic TeX's behaviour. I am doing a lot in ConTeXt but I mix it with parts of basic TeX. When I can no longer trust TeX doing like TeX I really don't know how to go further.
Please return the original behaviour.
Thanks for looking into it.
Hans van der Meer