* language.h: Add comment about current_language.

diff --git a/gdb/ChangeLog b/gdb/ChangeLog
index bb820e759617269a5200be681a5c533877891c1c..03ec13711a3487fb2db5d96dc2b8fd2b49633526 100644 (file)
--- a/gdb/ChangeLog
+++ b/gdb/ChangeLog
@@ -1,5 +1,7 @@
 Thu Sep  2 00:07:36 1993  Jim Kingdon  (kingdon@lioth.cygnus.com)
 
+       * language.h: Add comment about current_language.
+
        * mips-tdep.c (_initialize_mips_tdep): Change heuristic-fence-post
        from var_uinteger to var_zinteger.
 

diff --git a/gdb/language.h b/gdb/language.h
index ac32290cb741bcf910f64cbf406813f8fb0fe89b..8bb76a0ca9ebaab91eaf7a714117d0777d35d2d3 100644 (file)
--- a/gdb/language.h
+++ b/gdb/language.h
@@ -185,7 +185,19 @@ struct language_defn
 
 /* Pointer to the language_defn for our current language.  This pointer
    always points to *some* valid struct; it can be used without checking
-   it for validity.  */
+   it for validity.
+
+   The current language affects expression parsing and evaluation
+   (FIXME: it might be cleaner to make the evaluation-related stuff
+   separate exp_opcodes for each different set of semantics.  We
+   should at least think this through more clearly with respect to
+   what happens if the language is changed between parsing and
+   evaluation) and printing of things like types and arrays.  It does
+   *not* affect symbol-reading-- each source file in a symbol-file has
+   its own language and we should keep track of that regardless of the
+   language when symbols are read.  If we want some manual setting for
+   the language of symbol files (e.g. detecting when ".c" files are
+   C++), it should be a seprate setting from the current_language.  */
 
 extern const struct language_defn *current_language;