[no subject]

[Patrice Dumas]

branch: master
commit e0985097c6f62b741dfa15c35ff1dc758732b23f
Author: Patrice Dumas <pertusus@free.fr>
AuthorDate: Sat Jun 1 09:46:32 2024 +0200

    * tp/Texinfo/XS/main/element_types.txt: add more text flags for text
    elements.
    
    * tp/Texinfo/XS/main/manipulate_tree.c (copy_tree_internal): text
    elements may have inserted info set, so copy info.
    
    * tp/Texinfo/XS/main/build_perl_info.c (element_to_perl_hash):
    reorganize to take into account that text element do not args nor
    contents nor extras.
    
    * tp/Texinfo/XS/parsetexi/parser.c (isolate_trailing_space)
    (isolate_last_space, process_remaining_on_line): pass last contents
    element as isolate_trailing_space argument.
    
    * tp/Texinfo/XS/main/build_perl_info.c (element_to_perl_hash),
    tp/Texinfo/XS/parsetexi/parser.c (check_space_element)
    (text_contents_to_plain_text, merge_text, isolate_trailing_space),
    (process_remaining_on_line), tp/Texinfo/XS/parsetexi/source_marks.c
    (place_source_mark): use type flags to determine if an element
    is a text element.
---
 ChangeLog                              | 23 ++++++++++
 tp/TODO                                | 21 +--------
 tp/Texinfo/XS/main/build_perl_info.c   | 21 +++++----
 tp/Texinfo/XS/main/element_types.txt   | 16 +++----
 tp/Texinfo/XS/main/manipulate_tree.c   |  1 +
 tp/Texinfo/XS/parsetexi/parser.c       | 78 +++++++++++++++-------------------
 tp/Texinfo/XS/parsetexi/source_marks.c |  4 +-
 7 files changed, 84 insertions(+), 80 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index 3f314629d3..85f1dda97d 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -11,6 +11,29 @@
 
        * NEWS, doc/info-stnd.texi (Custom Key Bindings): update.
 
+2024-05-31  Patrice Dumas  <pertusus@free.fr>
+
+       * tp/Texinfo/XS/main/element_types.txt: add more text flags for text
+       elements.
+
+       * tp/Texinfo/XS/main/manipulate_tree.c (copy_tree_internal): text
+       elements may have inserted info set, so copy info.
+
+       * tp/Texinfo/XS/main/build_perl_info.c (element_to_perl_hash):
+       reorganize to take into account that text element do not args nor
+       contents nor extras.
+
+       * tp/Texinfo/XS/parsetexi/parser.c (isolate_trailing_space)
+       (isolate_last_space, process_remaining_on_line): pass last contents
+       element as isolate_trailing_space argument.
+
+       * tp/Texinfo/XS/main/build_perl_info.c (element_to_perl_hash),
+       tp/Texinfo/XS/parsetexi/parser.c (check_space_element)
+       (text_contents_to_plain_text, merge_text, isolate_trailing_space),
+       (process_remaining_on_line), tp/Texinfo/XS/parsetexi/source_marks.c
+       (place_source_mark): use type flags to determine if an element
+       is a text element.
+
 2024-05-31  Patrice Dumas  <pertusus@free.fr>
 
        * tp/Texinfo/XS/main/element_types.awk: order correctly the types.
diff --git a/tp/TODO b/tp/TODO
index d918eb4708..0786caee87 100644
--- a/tp/TODO
+++ b/tp/TODO
@@ -91,28 +91,11 @@ ELEMENT *elements[3]
 4) separate the structure of text elements and other elements, as text elements
    never have contents/args/info/extra, have cmd != 0.
     EXCEPTION: needed transient _copy and _counter in copy_tree_internal in 
extra
+    EXCEPTION: info inserted for spaces in @def* for definition aliases
   
-   if (content->type != ET_normal_text)
-            fprintf (stderr, "TTT %s\n", print_element_debug (content, 1));
-
-
-parsetexi
-  close.c
-    is_container_empty: (all text types) current->text.end == 0
-  debug_parser.c
-    print_element_debug_parser e->text.end > 0
-  parser.c
-    merge_text last_element->text.space > 0
-    check_space_element e->text.space > 0
-    text_contents_to_plain_text e1->text.end > 0
-    isolate_last_space text_len = last_elt->text.end; text_len <= 0
-    process_remaining_on_line && last_contents_child (current)->text.end > 0
-  source_marks.c 
-    place_source_mark last_child->text.end > 0
-
+type_data[->type].flags & TF_text
 main
   build_perl_info.c
-    element_to_perl_hash e->text.space > 0
     extra_misc_args contents.list[j]->text.end > 0
   convert_to_texinfo.c
     convert_to_texinfo_internal e->text.end > 0
diff --git a/tp/Texinfo/XS/main/build_perl_info.c 
b/tp/Texinfo/XS/main/build_perl_info.c
index ab628aa116..41ad4926e0 100644
--- a/tp/Texinfo/XS/main/build_perl_info.c
+++ b/tp/Texinfo/XS/main/build_perl_info.c
@@ -34,6 +34,7 @@
 
 #include "command_ids.h"
 #include "element_types.h"
+#include "types_data.h"
 #include "tree_types.h"
 #include "global_commands_types.h"
 /* for GLOBAL_INFO ERROR_MESSAGE */
@@ -609,12 +610,23 @@ element_to_perl_hash (ELEMENT *e, int avoid_recursion)
       hv_store (e->hv, "parent", strlen ("parent"), sv, HSH_parent);
     }
 
+  store_source_mark_list (e);
+
   if (e->type && e->type != ET_normal_text && e->type != ET_other_text)
     {
       sv = newSVpv (element_type_names[e->type], 0);
       hv_store (e->hv, "type", strlen ("type"), sv, HSH_type);
     }
 
+  store_additional_info (e, &e->info_info, "info", avoid_recursion);
+
+  if (type_data[e->type].flags & TF_text)
+    {
+      sv = newSVpv_utf8 (e->text.text, e->text.end);
+      hv_store (e->hv, "text", strlen ("text"), sv, HSH_text);
+      return;
+    }
+
   if (e->cmd)
     {
       sv = newSVpv (element_command_name (e), 0);
@@ -668,14 +680,7 @@ element_to_perl_hash (ELEMENT *e, int avoid_recursion)
         }
     }
 
-  if (e->text.space > 0)
-    {
-      sv = newSVpv_utf8 (e->text.text, e->text.end);
-      hv_store (e->hv, "text", strlen ("text"), sv, HSH_text);
-    }
-
   store_additional_info (e, &e->extra_info, "extra", avoid_recursion);
-  store_additional_info (e, &e->info_info, "info", avoid_recursion);
 
   if (e->associated_unit)
     {
@@ -691,8 +696,6 @@ element_to_perl_hash (ELEMENT *e, int avoid_recursion)
         }
     }
 
-  store_source_mark_list (e);
-
   if (e->source_info.line_nr)
     {
 #define STORE(key, sv, hsh) hv_store (hv, key, strlen (key), sv, hsh)
diff --git a/tp/Texinfo/XS/main/element_types.txt 
b/tp/Texinfo/XS/main/element_types.txt
index 5cf9bd65a0..42abd691d9 100644
--- a/tp/Texinfo/XS/main/element_types.txt
+++ b/tp/Texinfo/XS/main/element_types.txt
@@ -54,7 +54,7 @@ brace_command_arg
 brace_command_context
 block_line_arg
 line_arg
-rawline_arg
+rawline_arg                               text
 
 menu_entry
 menu_entry_leading_text                   text
@@ -70,8 +70,8 @@ internal_spaces_before_argument            text
 internal_spaces_before_brace_in_index      text
 internal_spaces_after_cmd_before_arg       text
 
-macro_name
-macro_arg
+macro_name                                 text
+macro_arg                                  text
 before_item
 table_entry
 table_term
@@ -86,7 +86,7 @@ multitable_body
 row
 balanced_braces
 bracketed_arg
-bracketed_linemacro_arg
+bracketed_linemacro_arg                   text
 
 # container of definition line argument content
 def_line_arg
@@ -104,10 +104,10 @@ def_typearg
 def_arg
 
 # text for def commands, definition line direct arguments
-delimiter
-spaces
+delimiter                                 text
+spaces                                    text
 # type of text element inserted in untranslated_def_line_arg
-untranslated
+untranslated                              text
 
 # for unexpanded parts of conditionals
 elided_rawpreformatted
@@ -124,7 +124,7 @@ text
 
 # for HTML converter
 _code
-_converted
+_converted                                text
 _string
 
 # not in parser, for virtual element associated to special output units
diff --git a/tp/Texinfo/XS/main/manipulate_tree.c 
b/tp/Texinfo/XS/main/manipulate_tree.c
index dcade5d597..ce4966d7f6 100644
--- a/tp/Texinfo/XS/main/manipulate_tree.c
+++ b/tp/Texinfo/XS/main/manipulate_tree.c
@@ -176,6 +176,7 @@ copy_tree_internal (ELEMENT* current, ELEMENT *parent)
   if (current->text.space > 0)
     {
       text_append (&new->text, current->text.text);
+      copy_associated_info (&current->extra_info, &new->extra_info);
       return new;
     }
 
diff --git a/tp/Texinfo/XS/parsetexi/parser.c b/tp/Texinfo/XS/parsetexi/parser.c
index e9755e76e6..cfbd12b0da 100644
--- a/tp/Texinfo/XS/parsetexi/parser.c
+++ b/tp/Texinfo/XS/parsetexi/parser.c
@@ -23,6 +23,7 @@
 #include "element_types.h"
 #include "tree_types.h"
 #include "global_commands_types.h"
+#include "types_data.h"
 #include "tree.h"
 #include "extra.h"
 #include "builtin_commands.h"
@@ -173,7 +174,7 @@ check_space_element (ELEMENT *e)
         || e->cmd == CM_c
         || e->cmd == CM_comment
         || e->cmd == CM_COLON
-        || (e->text.space > 0
+        || (type_data[e->type].flags & TF_text
             && (e->text.end == 0
                 || !*(e->text.text + strspn (e->text.text, whitespace_chars))))
      ))
@@ -201,8 +202,8 @@ text_contents_to_plain_text (ELEMENT *e, int 
*superfluous_arg)
   text_init (&result);
   for (i = 0; i < e->contents.number; i++)
     {
-      ELEMENT *e1 = contents_child_by_index (e, i);
-      if (e1->text.end > 0)
+      const ELEMENT *e1 = contents_child_by_index (e, i);
+      if (type_data[e1->type].flags & TF_text && e1->text.end > 0)
         ADD(e1->text.text);
       else if (e1->cmd == CM_AT_SIGN
                || e1->cmd == CM_atchar)
@@ -732,23 +733,10 @@ merge_text (ELEMENT *current, const char *text, size_t 
len_text,
   last_element = last_contents_child (current);
   if (!no_merge_with_following_text
       && last_element
-      /* There is a difference between the text being defined and empty,
-         and not defined at all.  The latter is true for 'brace_command_arg'
-         elements.  We need either to make sure that we initialize all elements
-         with text_append (&e->text, "") where we want merging with following
-         text, or treat as a special case here. */
-      && (last_element->text.space > 0
-            && !strchr (last_element->text.text, '\n')))
-    {
-      /*
-      if (last_element->type != ET_normal_text && last_element->type != 
ET_empty_line
-          && last_element->type != ET_ignorable_spaces_after_command
-          && last_element->type != ET_internal_spaces_after_command
-          && last_element->type != ET_internal_spaces_before_argument
-          && last_element->type != ET_spaces_after_close_brace)
-        fprintf (stderr, "EEE %s '%s'\n", 
element_type_names[last_element->type],
-                         last_element->text.text);
-       */
+      /* can actually be normal_text, and some space elements */
+      && type_data[last_element->type].flags & TF_text
+      && !strchr (last_element->text.text, '\n'))
+    {
       /* Transfer source marks */
       if (transfer_marks_element
           && transfer_marks_element->source_mark_list.number > 0)
@@ -872,7 +860,7 @@ abort_empty_line (ELEMENT **current_inout)
   return retval;
 }
 
-/* The caller verifies that the element is a text element */
+/* The caller verifies that last_elt is a text element */
 static void
 isolate_last_space_internal (ELEMENT *current, ELEMENT *last_elt)
 {
@@ -921,18 +909,13 @@ isolate_last_space_internal (ELEMENT *current, ELEMENT 
*last_elt)
     }
 }
 
-/* The callers verify that the last_elt is a text element */
+/* The caller verifies that last_elt is a text element */
 static void
-isolate_trailing_space (ELEMENT *current, enum element_type spaces_type)
+isolate_trailing_space (ELEMENT *current, ELEMENT *last_elt,
+                        enum element_type spaces_type)
 {
-  ELEMENT *last_elt;
-  char *text;
-  int text_len;
+  char *text = last_elt->text.text;
 
-  last_elt = last_contents_child (current);
-  text = element_text (last_elt);
-
-  text_len = last_elt->text.end;
 
   /* If text all whitespace */
   if (text[strspn (text, whitespace_chars)] == '\0')
@@ -943,6 +926,7 @@ isolate_trailing_space (ELEMENT *current, enum element_type 
spaces_type)
     {
       ELEMENT *new_spaces;
       int i, trailing_spaces;
+      int text_len = last_elt->text.end;
 
       trailing_spaces = 0;
       for (i = text_len - 1;
@@ -988,6 +972,9 @@ isolate_last_space (ELEMENT *current)
 
   last_elt = last_contents_child (current);
 
+  if (!(type_data[last_elt->type].flags & TF_text))
+    goto no_isolate_space;
+
   text_len = last_elt->text.end;
   if (text_len <= 0)
     goto no_isolate_space;
@@ -1002,7 +989,7 @@ isolate_last_space (ELEMENT *current)
   debug_parser_print_element (last_elt, 0); debug ("");
 
   if (current->type == ET_menu_entry_node)
-    isolate_trailing_space (current, ET_space_at_end_menu_node);
+    isolate_trailing_space (current, last_elt, ET_space_at_end_menu_node);
   else
     isolate_last_space_internal (current, last_elt);
 
@@ -2309,25 +2296,30 @@ process_remaining_on_line (ELEMENT **current_inout, 
const char **line_inout)
            || cmd == CM_seealso
            || cmd == CM_subentry)
           && current->contents.number > 0
-          && last_contents_child (current)->text.end > 0
        /* it is important to check if in an index command, as otherwise
           the internal space type is not processed and remains as is in
           the final tree. */
           && (command_flags(current->parent) & CF_index_entry_command
                || current->parent->cmd == CM_subentry))
         {
-          if (cmd == CM_subentry)
-            {
-              isolate_trailing_space (current, ET_spaces_at_end);
-            }
-          else
-           /* an internal and temporary space type that is converted to
-              a normal space without type if followed by text or a
-              "spaces_at_end" if followed by spaces only when the
-              index or subentry command is done. */
+          ELEMENT *last_elt = last_contents_child (current);
+
+          if (last_elt && type_data[last_elt->type].flags & TF_text
+              && last_elt->text.end > 0)
             {
-              isolate_trailing_space (current,
-                                      
ET_internal_spaces_before_brace_in_index);
+              if (cmd == CM_subentry)
+                {
+                  isolate_trailing_space (current, last_elt, ET_spaces_at_end);
+                }
+              else
+               /* an internal and temporary space type that is converted to
+                  a normal space without type if followed by text or a
+                  "spaces_at_end" if followed by spaces only when the
+                  index or subentry command is done. */
+                {
+                  isolate_trailing_space (current, last_elt,
+                               ET_internal_spaces_before_brace_in_index);
+                }
             }
         }
 
diff --git a/tp/Texinfo/XS/parsetexi/source_marks.c 
b/tp/Texinfo/XS/parsetexi/source_marks.c
index 34c7d76342..ff0fef86d2 100644
--- a/tp/Texinfo/XS/parsetexi/source_marks.c
+++ b/tp/Texinfo/XS/parsetexi/source_marks.c
@@ -18,6 +18,7 @@
 #include <stdlib.h>
 
 #include "tree_types.h"
+#include "types_data.h"
 #include "tree.h"
 /* for count_multibyte and other */
 #include "utils.h"
@@ -63,7 +64,8 @@ place_source_mark (ELEMENT *e, SOURCE_MARK *source_mark)
     {
       ELEMENT *last_child = last_contents_child (e);
       mark_element = last_child;
-      if (last_child->text.end > 0)
+      if (type_data[last_child->type].flags & TF_text
+          && last_child->text.end > 0)
         source_mark->position = count_multibyte (last_child->text.text);
     }
   else