There are a few literal non-ASCII characters in the SGML documentation,
namely in
isn.sgml
release-7.4.sgml
release-8.4.sgml
Also, there are some encoded (&foo;) non-ASCII characters in
release-8.0.sgml
release-8.1.sgml
release-8.2.sgml
unaccent.sgml
These all work fine, because they are all LATIN1, and DocBook SGML uses
LATIN1.
But I notice that the contributor names in the 9.1 release notes have
been carefully ASCII-fied, presumably from the Git UTF-8 commit
messages.
For additional amusement, when creating the HISTORY file, lynx recodes
the HTML into the encoding specified by your LC_CTYPE environment
setting.
Also, the following source files contain non-ASCII characters in
comments:
src/backend/port/dynloader/darwin.c (LATIN1)
src/backend/storage/lmgr/predicate.c (UTF8)
src/backend/storage/lmgr/README-SSI (UTF8)
The last two are new in 9.1.
So, some questions:
* Should we consistently use entities for encoding non-ASCII
characters in SGML? Or use LATIN1 freely?
* Should we allow/use non-ASCII characters in the release notes?
* What encoding should the HISTORY file have?
* Should we allow non-ASCII characters in general source files?
* If so, what should the encoding be?