RLE Replacement

Source code in one programming language is written with uppercase English letters only, and the same letter often repeats many times in a row. Very large source files in that language are stored compressed by run length encoding.

Run length encoding (RLE) rewrites every maximal block of identical letters as a pair of that letter and the length of the block. For example, the string RRRRLEEE becomes R4L1E3 under RLE.

Three strings $A$, $B$, and $C$ are given in RLE form. Write a program that replaces the first occurrence of $B$ as a substring of $A$ with $C$ and prints the resulting string in RLE form. If $B$ does not occur in $A$, print $A$ in RLE form unchanged.

The input consists of three lines.

A
B
C

The three lines hold the RLE encodings of $A$, $B$, and $C$, in that order. Each line has this format:

c1 l1 c2 l2 ... cn ln $

Each $c_i$ ($1 \le i \le n$) is an uppercase English letter (A through Z), and each $l_i$ ($1 \le i \le n$, $1 \le l_i \le 10^8$) is an integer, the length of the repetition of $c_i$. The number of pairs $n$ satisfies $1 \le n \le 10^3$. The letters and the integers are separated by a single space, and the terminal symbol $ ends the line. For every $i$ with $1 \le i \le n-1$, $c_i \neq c_{i+1}$ holds.

If $B$ occurs in $A$, replace the first occurrence of $B$ with $C$; otherwise keep $A$ as it is. Print the result in RLE form on one line, in this format:

c1 l1 c2 l2 ... cm lm $

The output must satisfy $c_i \neq c_{i+1}$ for $1 \le i \le m-1$ and $l_i > 0$ for $1 \le i \le m$.