Matthew Leeds authored Jun 25, 2016 and Sébastien Wilmet committed Jun 30, 2016

gtk-doc treats backslash as the escape character, so devhelp displays
"\." as "." and "\\" as "\". In the documentation for
gtk_source_utils_escape_search_text() and
gtk_source_utils_unescape_search_text(), we want to display "\" and "\\"
literally so they have to be surrounded by backticks.

https://bugzilla.gnome.org/show_bug.cgi?id=768040

Matthew Leeds authored Jun 23, 2016 and Sébastien Wilmet committed Jun 24, 2016

https://bugzilla.gnome.org/show_bug.cgi?id=767978

Matthew Leeds authored Jun 23, 2016 and Sébastien Wilmet committed Jun 24, 2016

When at_word_boundaries is set in the SearchSettings and an insertion
occurs on a word boundary, the affected occurrence needs to be
re-scanned. This commit changes remove_occurrences_in_range to
properly adjust the given boundaries to accomplish this.

https://bugzilla.gnome.org/show_bug.cgi?id=767978

libxml2 headers could also be in $(prefix)/include/libxml2, which is what
happens on *nix, so look for them there.

Note that the libxml2 NMake Makefiles however does copy all its headers
under $(prefix)/include/libxml rather than
$(prefix)/include/libxml2/libxml, but let's do this for consistency's
sake, especially as this is already done for libcroco and librsvg.

The warning was:
warning C4090: 'function' : different 'const' qualifiers

But it's normally valid C code, so it's probably a false negative of the
Visual Studio compiler.

_gtk_source_style_scheme_apply() doesn't accept a NULL StyleScheme,
which is normal.

Factor out the update_style() function, to avoid code duplication, since
the code was correct at one place and incorrect at another place.

https://bugzilla.gnome.org/show_bug.cgi?id=760984

I'm surprised this didn't trigger any compiler warnings or errors.

There was a good reason behind
commit 949cd128: the code was not clear.
It was not evident to know why that code was useful.

If the title of this commit doesn't make any sense, it's because the
::highlight-updated signal has a bad name. When the ContextEngine sends
it, the _highlight_ is actually not yet updated.

Now with the new implementation and the comment in the code, I hope
it'll be a bit clearer. But it would be nice to do something about the
signal name.

https://bugzilla.gnome.org/show_bug.cgi?id=767565

Cairo headers could be in $(prefix)/include/cairo, so look for them there
as well.

This should fix a warning on ARM:
https://bugzilla.gnome.org/show_bug.cgi?id=767279

Ouch. I'm surprised that file loading worked fine during all this time.

It's strange that on my machine (Fedora 24) this didn't trigger any
warnings or errors. -Wsign-compare is present on the command line.

g_utf8_strlen() returns a glong. minimum_word_size is a guint. So cast
the latter to glong. This should make the error to disappear, but there
are maybe other places in GtkSourceView where a glong is compared to a
guint. So finding a way to reproduce the warning/error would be better.

https://bugzilla.gnome.org/show_bug.cgi?id=767254

To fix this critical message:
Gtk-CRITICAL **: gtk_widget_get_preferred_width_for_height: assertion
'height >= 0' failed

Thanks to slaf for the suggestion on IRC.

The scrollbars size are no longer taken into account, since overlay
scrollbars are used. The comment was not updated. But I still think the
comment is useful.

Move some keywords from errors to valid keywords

https://bugzilla.gnome.org/show_bug.cgi?id=765639

https://bugzilla.gnome.org/show_bug.cgi?id=765495

https://bugzilla.gnome.org/show_bug.cgi?id=765495

My apologies for breaking this up into two.

When cleaning up matching brackets, gtk_text_buffer_remove_tag() is called
for the entire buffer. This causes a damage in the view causing the entire
pixelcache to be dropped.

Further more, on every ::insert-text we perform the bracket updating
meaning we damage the views on every keystroke.

This walks the tag toggles manually to to reduce the ranges we damage
to just the locations containing the tag.

https://bugzilla.gnome.org/show_bug.cgi?id=765363

It is often the case that a GtkSourceMap will be packed into the widget
hierarchy, but not visible unless the user has enabled the feature.

This means that we receive a lot of signal notifications when we can't
do anything about it. So instead, toggle the signal activity based on
whether or not we are visible.

https://bugzilla.gnome.org/show_bug.cgi?id=765358

We can remove the calls to check if the cursor is visible for every visible
line by keeping track of when the property changes. This removes a
gtk_text_view_get_cursor_visible() and gtk_text_gutter_render_get_view()
for every line in the rendering process.

https://bugzilla.gnome.org/show_bug.cgi?id=765356

This makes the cost of ::query_data() for the line gutter renderer
equivalent to the cost of the g_strdup_printf() previously.

https://bugzilla.gnome.org/show_bug.cgi?id=765356

Avoid looping division to count digits. Not huge gains or anything, but
common to avoid loop+division in general. ~25% faster for numbers into
1000s.

https://bugzilla.gnome.org/show_bug.cgi?id=765356