From cf5cf5491d877a3ac92a9995df69be6f767e1b3a Mon Sep 17 00:00:00 2001
From: Ivica Ico Bukvic <ico@vt.edu>
Date: Sun, 6 Sep 2020 18:38:00 -0400
Subject: [PATCH] Fixed bug that prevented proper reinstantiation after
deselect
* Previously reinstantiation only worked every other time the object was deselected. Now we force it to properly reinstantiate after each deselect.
---
externals/pddp/pddplink.c | 72 +++++++++++++++++++++------------------
1 file changed, 39 insertions(+), 33 deletions(-)
diff --git a/externals/pddp/pddplink.c b/externals/pddp/pddplink.c
index 847cf1f6a..7120241ca 100644
--- a/externals/pddp/pddplink.c
+++ b/externals/pddp/pddplink.c
@@ -110,39 +110,6 @@ static void pddplink_select(t_gobj *z, t_glist *glist, int state)
}
}
-static void pddplink_activate(t_gobj *z, t_glist *glist, int state)
-{
- t_pddplink *x = (t_pddplink *)z;
- t_rtext *y = glist_findrtext(glist, (t_text *)x);
- rtext_activate(y, state);
- x->x_rtextactive = state;
- if (!state) {
- /* Big workaround for pddplink without the -box option...
-
- After this call, Pd calls text_setto to see if it needs to
- instantiate the object. For nearly all t_text objects, the
- object doesn't get rebuilt if the text inside the box remains
- the same. Instead, it just calls rtext_senditup which (eventually)
- sends a message to the GUI to simply supply new text for the
- current object. (It shouldn't do this if the text hasn't actually
- changed, but that's another story...).
-
- Anyway, since pddplink has this weird widget behavior, the text
- displayed is different than the text we typed in the box-- hence
- the x->x_vistext member here. So if we happened to edit the box
- without changing the text, rtext_senditup would update the link
- to be "pddplink foo" instead of just "foo".
-
- To prevent this, we just zero out the object's te_binbuf to ensure
- that text_setto re-instantiates. That ensures the correct text is
- printed for the link */
- t_binbuf *b = binbuf_new();
- t_binbuf *old = x->x_ob.te_binbuf;
- x->x_ob.te_binbuf = b;
- binbuf_free(old);
- }
-}
-
static void pddplink_vis(t_gobj *z, t_glist *glist, int vis)
{
t_pddplink *x = (t_pddplink *)z;
@@ -188,6 +155,45 @@ static void pddplink_vis(t_gobj *z, t_glist *glist, int vis)
}
}
+static void pddplink_activate(t_gobj *z, t_glist *glist, int state)
+{
+ t_pddplink *x = (t_pddplink *)z;
+ t_rtext *y = glist_findrtext(glist, (t_text *)x);
+ rtext_activate(y, state);
+ x->x_rtextactive = state;
+ if (!state) {
+ /* Big workaround for pddplink without the -box option...
+
+ After this call, Pd calls text_setto to see if it needs to
+ instantiate the object. For nearly all t_text objects, the
+ object doesn't get rebuilt if the text inside the box remains
+ the same. Instead, it just calls rtext_senditup which (eventually)
+ sends a message to the GUI to simply supply new text for the
+ current object. (It shouldn't do this if the text hasn't actually
+ changed, but that's another story...).
+
+ Anyway, since pddplink has this weird widget behavior, the text
+ displayed is different than the text we typed in the box-- hence
+ the x->x_vistext member here. So if we happened to edit the box
+ without changing the text, rtext_senditup would update the link
+ to be "pddplink foo" instead of just "foo".
+
+ To prevent this, we just zero out the object's te_binbuf to ensure
+ that text_setto re-instantiates. That ensures the correct text is
+ printed for the link */
+ t_binbuf *b = binbuf_new();
+ t_binbuf *old = x->x_ob.te_binbuf;
+ x->x_ob.te_binbuf = b;
+ binbuf_free(old);
+ /* ico@vt.edu 20200906: this is ugly but necessary because otherwise
+ the object only activates correctly (reverts to its active link)
+ every other time when being deselected
+ */
+ pddplink_vis(z, glist, 0);
+ pddplink_vis(z, glist, 1);
+ }
+}
+
static int pddplink_wbclick(t_gobj *z, t_glist *glist, int xpix, int ypix,
int shift, int alt, int dbl, int doit);
--
GitLab