The value of *latin-1-code-pages* is a list of integers, which must be Windows code page identifiers. When the current Code Page is on this list, the default file encoding detection algorithm will cause
(:latin-1 :encoding-error-action 63) to be used for file I/O. Files will be written as Latin-1 with '?' replacing any non-Latin-1 character. This is faster than converting to the code page.
If safe-locale-file-encoding is used for file encoding detection, then the :latin-1-safe external format will be used.
The LispWorks editor binds *latin-1-code-pages* to nil when reading and writing files, in order to ensure that code page characters outside of Latin-1 are handled regardless of the configuration of open.
LispWorks User Guide and Reference Manual - 13 Feb 2015