Neil Hodgson wrote:
Robert Roessler:
I followed up on John's look at this with an atom set to match
DROPFILES_DND... and things work fine with one of the latest Win32
versions of GTK (2.8.18) and glib (2.10.3).
So it looks to me like one can just take the released version of
ScintillaGTK.cxx and use the initial if clause with an
"|| selection_data->type == atomDROPFILES_DND"
added (with the appropriate declarations and intializations for
atomDROPFILES_DND, of course) - that is what I have done for testing
purposes.
I looked up the code and can't see why you wouldn't get text/uri-list
http://www.koders.com/c/fid1A78BB9F26B1B6824DAF0D8DE61CE2CDFBB5484E.aspx
Are you calling the WIn32 DragAcceptFiles function on the window?
That's the normal way of declaring you want to receive drop file
events. If this isn't called you may be seeing lower level drop
events.
How about the DROPFILES_DND is what you get when it really came from a
Windows WM_DROPFILES, while the other is a more general case dealing
with "synthesized" drop data types (just conjecture)?
I don't think either my OCaml app code or my Scintilla wrapper code
would ever be calling DragAcceptFiles or any other Win32 API... I have
an OCaml binding of a GTK Scintilla widget, both of which work hard to
shield client code from specific OS issues and provide an "abstract"
interface for receiving dropped "URIs" - realized in the form of the
Scintilla SCN_URIDROPPED notification. :)
Now it is certainly true that either (or both) of the Scintilla GTK
platform layer or the GTK-on-Windows platform layers can and probably
should issue DragAcceptFiles... after all, *someone* has to. And
clearly someone *does*, since I do get the drop - but only if the
Scintilla GTK platform layer also checks for the GTK DROPFILES_DND
data type.
Finally, how about uncommenting the "unrecognized" case error print -
that seems to be the GTK/glib way anyway... this would have been a
little easier to diagnose if that message had shown.
I'm not real happy about this although a few printfs have stayed
enabled. Calling applications should receive silence and so not mess
up transcripts that include SciTE actions except for real errors and
this looks like an informational message.
OK - I agree that I usually (only?) see these GTK/glib stderr messages
on "critical" problems...
The version which just adds the atomDROPFILES_DND to the initial
accepted type is available at
http://www.rftp.com/ScintillaGTK.cxx
Robert Roessler
[EMAIL PROTECTED]
http://www.rftp.com
_______________________________________________
Scintilla-interest mailing list
[email protected]
http://mailman.lyra.org/mailman/listinfo/scintilla-interest
