Skip to content

glib: fix ABI issues with macros/constants #135

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 1 commit into from
Closed

glib: fix ABI issues with macros/constants #135

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
31 changes: 1 addition & 30 deletions GLib-2.0.gir
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12890,32 +12890,6 @@ module.</doc>
<doc xml:space="preserve">Multiplying the base 2 exponent by this number yields the base 10 exponent.</doc>
<type name="gdouble" c:type="gdouble"/>
</constant>
<constant name="LOG_DOMAIN" value="0" c:type="G_LOG_DOMAIN">
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Instead this can be ignored in the Gir.toml file which will drop it from the ffi tests

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I tried that but couldn't get it to work, let me try again

Copy link
Member

@bilelmoussaoui bilelmoussaoui Nov 30, 2022 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

we do that in gtk for the version constants for example

<doc xml:space="preserve">Defines the log domain. See [Log Domains](#log-domains).

Libraries should define this so that any messages
which they log can be differentiated from messages from other
libraries and application code. But be careful not to define
it in any public header files.

Log domains must be unique, and it is recommended that they are the
application or library name, optionally followed by a hyphen and a sub-domain
name. For example, `bloatpad` or `bloatpad-io`.

If undefined, it defaults to the default %NULL (or `""`) log domain; this is
not advisable, as it cannot be filtered against using the `G_MESSAGES_DEBUG`
environment variable.

For example, GTK+ uses this in its `Makefile.am`:
|[
AM_CPPFLAGS = -DG_LOG_DOMAIN=\"Gtk\"
]|

Applications can choose to leave it as the default %NULL (or `""`)
domain. However, defining the domain offers the same advantages as
above.</doc>
<type name="gchar" c:type="gchar"/>
</constant>
<constant name="LOG_FATAL_MASK" value="5" c:type="G_LOG_FATAL_MASK">
<doc xml:space="preserve">GLib log levels that are considered fatal by default.

Expand Down Expand Up @@ -13884,7 +13858,7 @@ levels using g_log_set_handler() and g_log_set_fatal_mask().</doc>
<member name="level_debug" value="128" c:identifier="G_LOG_LEVEL_DEBUG">
<doc xml:space="preserve">log level for debug messages, see g_debug()</doc>
</member>
<member name="level_mask" value="-4" c:identifier="G_LOG_LEVEL_MASK">
<member name="level_mask" value="4294967292" c:identifier="G_LOG_LEVEL_MASK">
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

shouldn't this be fixed upstream?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You can add (type uint) annotation for this IIRC

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

And (value 4294967292). That's done in GStreamer in many places. It also confuses the Python bindings if you don't do that, but I guess nobody is using G_LOG_LEVEL_MASK from Python.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we have this problem with any flags type, we consider them u32 but in C anything in an enum has to be int. Maybe I will just have gir cast them to unsigned

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah it looks like those annotations won't work on an enum constant, they only work on a #define

<doc xml:space="preserve">a mask including all log levels</doc>
</member>
</bitfield>
Expand Down Expand Up @@ -45766,9 +45740,6 @@ application domain</doc>
</parameter>
</parameters>
</function-macro>
<constant name="macro__has_attribute___noreturn__" value="0" c:type="g_macro__has_attribute___noreturn__">
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Similar to the other constant

<type name="gint" c:type="gint"/>
</constant>
<function-macro name="macro__has_builtin" c:identifier="g_macro__has_builtin" introspectable="0">
<parameters>
<parameter name="x">
Expand Down
7 changes: 7 additions & 0 deletions fix.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,6 +12,13 @@ xmlstarlet ed -L \
-u '//*[@glib:error-domain="g-option-context-error-quark"]/@glib:error-domain' -v g-option-error-quark \
GLib-2.0.gir

# fix layout checks
xmlstarlet ed -L \
-d '//_:constant[@name="LOG_DOMAIN"]' \
-d '//_:constant[@name="macro__has_attribute___noreturn__"]' \
-u '//_:member[@c:identifier="G_LOG_LEVEL_MASK"]/@value' -v "4294967292" \
GLib-2.0.gir

# GtkEntry icon signals incorrect assume GdkEventButton when other variants may be passed
xmlstarlet ed -L \
-u '//_:class[@name="Entry"]/glib:signal[@name="icon-press"]//_:parameter[@name="event"]/_:type[@name="Gdk.EventButton"]/@name' -v "Gdk.Event" \
Expand Down