Before, *_html_parser_state() operated with struct html_element *. Now, it is
transparent for the renderer (just void *), so that DOM won't have to provide
this struct but will be able to use something internal.
Normally, you can add some objects or subdirs to the build if some
config variable is set to yes, by doing:
SUBDIRS-$(CONFIG_FOO)
Now you can do it the other way too, including stuff if some
variable is set to no:
SUBDIRS-unless$(CONFIG_FOO)
This will be used for switching between two HTML parser implementations.
...as struct text_style. This way it might be possible later to
add other default formatting attributes by CSS and it allows
quite a code simplification in the DOM renderer.
Currently, all DOM, HTML and plain renderers had their own routine for
conversion from text style to screen attribute. This moves text_style and
text_style_format from html/parser.h to renderer.h and introduces new generic
routine get_screen_chracter_template() that is used by all the specific
rendering engines.
The parsing in parse_set and parse_unset saves, overwrites with a NUL, and restores a character in the string that is being parsed. If there is a malloc failure between overwriting and restoring, the restore is not done. This commit changes that behaviour to restore before returning.
Now, CSS is initialized separately for each of the renderers, so that
also RSS doesn't just choose styles of random DOM node types.
init_template_by_style() is introduced as the common backend for
loading CSS properties.
Introduce get_option_shadow. This routine takes an option, the tree under which that option resides, and another tree. It returns a corresponding option with parallel ancestry in the second tree, creating that option and ancestry if it does not already exist.
Add enum copy_option_flags.
Add the CO_NO_LISTBOX_ITEM flag, which is used to suppress listbox creation for shadowed trees.
Add the CO_SHALLOW flag, which is used to suppress the duplication of unwanted children for shadowed trees.
Add a flags parameter to copy_option and tree_dup (and out of necessity, struct option_type_info and str_dup).
This fixes ELinks crashing on this with terminal width e.g. 103:
<p align="justify">
xxxx xxxx xxxx xxxxx xxxxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx xxxxx
xxxxx xxxx xxxxxxx xxx xxxxxxx xxx xxxx xxxx xxxx xx xxxx x xxx xxxx
xxxx xx xxxx xxxx xxxx xxx—xxx xxxx xx—xxx xxxx<em> </em>x
xxxx </p>
This test was removed for an unknown reason in commit
b1cc717789.
Discovered together with Miciah.
Update a comment in encode_multipart, which refers to html_form_control, which has since been renamed to html_special_form_control.
The comment was added with this commit:
commit b4dee890a61a6c8a27a8e4cd1dc3b3b93f1cdb08
Author: Petr Baudis <pasky@ucw.cz>
Date: Fri May 10 13:26:55 2002 +0000
Don't decode and back encode hidden form items (by mikulas, from 0.97).
The function was renamed with this commit:
commit c9d72739c715b3b0c7c6fec582780c1e8f444fc4
Author: Petr Baudis <pasky@ucw.cz>
Date: Sat Dec 18 02:22:28 2004 +0000
html_(tag|form*) -> html_special_\1, to naming prevent conflicts with HTML element handlers. As suggested by Jonas.
This change:
- Adds a check for the doxygen program to configure.
- Moves the Doxyfile from src/Doxyfile to doc/Doxyfile.in.
- Generates a doc/Doxyfile from doc/Doxyfile.in inserting
an absolute path to the source directory, so that it
also works when builddir != srcdir.
- Adds `make api` rule for running doxygen; it depends on
api/doxygen file which is never created to force the rule
to always run.
