On various sites, they recommend not loading the
package under XeLaTeX but to simply use the combination of
xunicode (the latter being loaded automatically
by the former). This recommendation is based upon the fact
textcomp uses the deprecated TS1 encoding.
However, when I try to typeset a symbol
textcomp package, e.g.
(character 108 in the CRM font)
I receive a compilation error.
My question is: Do we still need to load the textcomp package
or is there an automated way, like
unicode-math (for mathematical
symbols) to use these symbols without loading the package?