@b4n requested changes on this pull request.

Is GEANY_WORDCHARS actually useful to anyone? I think it's pretty much useless in most situations, and wrong in many other (e.g. it's not a good universal way to recognize a "word", not even for programming languages).
And it's fairly easy to copy over, or even generate using I don't know, '_' + string.lowercase + string.uppercase + string.digits in Python, and probably something similar in another language.

Also, most if not all the other macros are irrelevant to non-C, like foreach_document, documents and alike. The only potentially interesting things would be TAG_*, but I'm not even sure we really think it's part of the API (OK, it is, but I'm not sure we did know), and no plugin is using it ATM AFAICT.

GEANY_API_VERSION is also probably uninteresting as I guess the proxy is more relevant here.

Anyway, the code here seem mostly good to me, but I'm not very sure it's a good or useful idea. And if we do want macros, we should probably mark a lot of the existing ones @girskip in any case.

Note also: Due to a bug in g-ir-scanner, GEANY_ABI_VERSION is parsed
incorrectly (simply without the GTK version shift) but the ABI is
of little relevance for GIR consumers.

FWIW GEANY_ABI_SHIFT is not part of the API, and thus is not in the gtkdoc header, no surprise g-ir-scanner has a hard time :)

In scripts/gen-api-gtkdoc.py:

> +class DoxyDefine(DoxyElement):
+    @staticmethod
+    def from_memberdef(xml):
+        name = xml.find("name").text
+        v = xml.find("initializer")
+        if (v is not None):
+            proc = DoxygenProcess()
+            text = proc.process_element(v)
+            params = ",".join(p.text for p in xml.xpath(".//param/defname"))
+            if (params != ""):
+                params = "(%s)" % params
+            d = "#define %s%s %s" % (name, params, text)
+        else:
+            d = "#define %s" % name
+
+        e = DoxyEnum(name, d)

Shouldn't this be DocyDefine(name, d)? (not that it actually matters much, but still)

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.