1 files changed, 209 insertions, 0 deletions
diff --git a/personal/itches/emacs/narrow-lighter.md b/personal/itches/emacs/narrow-lighter.md
new file mode 100644
index 0000000..e5987b8
--- /dev/null
+++ b/personal/itches/emacs/narrow-lighter.md
@@ -0,0 +1,209 @@
+# Make " Narrow" lighter customizable
+
+The " Narrow" string comes from `src/xdisp.c:decode_mode_spec`:
+
+``` c
+case 'n':
+  if (BUF_BEGV (b) > BUF_BEG (b) || BUF_ZV (b) < BUF_Z (b))
+    return " Narrow";
+```
+
+This is probably just a matter of returning the contents of a Lisp
+variable instead of this constant string.
+
+TODO:
+
+1. get the string value of a variable in C
+2. define a customizable string variable
+3. write a news entry
+4. write a patch
+5. extra credits: display string properties
+
+## Get the string value of a variable in C
+
+`decode_mode_spec` has some relevant snippets:
+
+- Given a `Lisp_Object obj`, `SSDATA(obj)` gives the string value as a
+  `char*`.
+
+- How to get a variable's `Lisp_Object`?
+    - `BVAR` works for buffer-local variables
+    - `V${lispname//-/_}`
+
+## Define a customizable string variable
+
+### Defining variables visible to C code
+
+The C macro `DEFVAR_LISP(string-name, field-name)` does the following:
+
+    define a static `Lisp_Objfwd` variable v
+    get the address of globals._f##field-name &f
+
+    defvar_lisp(v, string-name, &f)
+
+As explained in the comments above `DEFVAR_LISP`, `globals` is a
+global variable defined in `globals.h`, which is "auto-generated by
+make-docfile" and exposes fields, `#define`s and `Lisp_Object`s for
+every global variable.
+
+make-docfile (`lib-src/make-docfile.c`) takes C files as input and
+searches all occurences of `^ +DEFSYM[ \t(]`, `^ +DEFVAR_[ILB]` or
+`^DEFU`, analyses what comes after and generates appropriate
+definitions for `globals.h`.
+
+`defvar_lisp` allocates a symbol using `Fmake_symbol`.
+
+### Making it customizable
+
+`lisp/cus-start.el` defines customizable properties of symbols defined
+by C code.
+
+AFAICT, there is no need to assign the default value right after
+defining the variable with `DEFVAR_LISP`: e.g. `shell-file-name` is
+`DEFVAR_LISP`ed in `src/callproc.c` and its default value is set in…
+Mmm.  Not in `cus-start.el`.  There is this snippet in
+`callproc.c:init_callproc`:
+
+  ``` c
+sh = getenv ("SHELL");
+Vshell_file_name = build_string (sh ? sh : "/bin/sh");
+  ```
+
+But when starting with `SHELL=rofl emacs -Q`, Custom says that the
+value "has been changed outside Customize".  Changed from what to
+what?
+
+`cus-start.el` may contain a hint:
+
+``` elisp
+;; Elements of this list have the form:
+;; …
+;; REST is a set of :KEYWORD VALUE pairs.  Accepted :KEYWORDs are:
+;; :standard - standard value for SYMBOL (else use current value)
+;; …
+```
+
+Except that nope, this does not work.  Giving `:standard " Narrow"`
+and looking at the variable in Custom yields
+
+    narrow-lighter: nil
+        [State]: CHANGED outside Customize. (mismatch)
+
+A better example might be `overlay-arrow-string`, whose default value
+is set right after `DEFVAR_LISP` by calling `build_pure_c_string`.
+
+Why `build_pure_c_string` and not `build_string`?  From "(elisp) Pure
+Storage":
+
+> Emacs Lisp uses two kinds of storage for user-created Lisp objects:
+> “normal storage” and “pure storage”.  Normal storage is where all
+> the new data created during an Emacs session are kept (see Garbage
+> Collection).  Pure storage is used for certain data in the preloaded
+> standard Lisp files—data that should never change during actual use
+> of Emacs.
+>
+> Pure storage is allocated only while ‘temacs’ is loading the
+> standard preloaded Lisp libraries.  In the file ‘emacs’, it is
+> marked as read-only (on operating systems that permit this), so that
+> the memory space can be shared by all the Emacs jobs running on the
+> machine at once.
+
+"(elisp) Building Emacs" explains that "temacs" is the minimal Elisp
+interpreter built by compiling all C files in `src/`; temacs then
+loads Elisp sources and creates the "emacs" executable by dumping its
+current state into a file.
+
+## Debug stuff
+
+### Unicode characters represented as octal sequences
+
+Trying to customize the new variable to any string with non-ASCII
+characters fails: they show up as sequences of backslash-octal codes.
+For some reason they show up fine in the Help and Custom buffers.
+
+Things to investigate:
+
+1. Should the `Lisp_Object` be created with something other than
+   `build_pure_c_string`? 🙅
+2. What does the code calling `decode_mode_spec` do with the returned
+   string? **🎉**
+3. (Does `SSDATA` make some transformation before returning the
+   string? 🤷)
+4. (Should a specialized Custom setter be defined? 🤷)
+
+#### Should the `Lisp_Object` be created with something other than    `build_pure_c_string`?
+
+Maybe this would work?
+
+``` c
+Vnarrow_lighter = make_multibyte_string(" Narrow", strlen(" Narrow"),
+                                        strlen(" Narrow)");
+```
+
+That looks too ugly though, let's try something else.
+
+Maybe `STRING_SET_MULTIBYTE(Vnarrow_lighter)` would help?
+
+*compiles and tries*
+
+… Nope, it does not.
+
+#### What does the code calling `decode_mode_spec` do with the returned string?
+
+``` c
+spec = decode_mode_spec (it->w, c, field, &string);
+multibyte = STRINGP (string) && STRING_MULTIBYTE (string);
+```
+
+*slowly turns around*
+
+*finds `string` standing right there with a blank stare*
+
+Gah!  How long have you been there?
+
+``` c
+/* Return a string for the output of a mode line %-spec for window W,
+   generated by character C.  […]  Return a Lisp string in
+   *STRING if the resulting string is taken from that Lisp string.
+   […] */
+static const char *
+decode_mode_spec (struct window *w, register int c, int field_width,
+                  Lisp_Object *string)
+{
+  Lisp_Object obj;
+  /* … */
+  obj = Qnil;
+  *string = Qnil;
+
+  switch (c)
+    {
+    /* … */
+    }
+
+  if (STRINGP (obj))
+    {
+      *string = obj;
+      return SSDATA (obj);
+    }
+  else
+    return "";
+}
+```
+
+Alright then:
+
+``` c
+case 'n':
+  if (BUF_BEGV (b) > BUF_BEG (b) || BUF_ZV (b) < BUF_Z (b))
+      obj = Vnarrow_lighter;
+  break;
+```
+
+### Why do string properties not show up?
+
+🤷
+
+## Extra credit
+
+Maybe it would be simpler to have the narrowing lighter work like the
+" Compiling" lighter (cf. `compilation-in-progress` variable), i.e. adding an entry to `minor-mode-alist`.