[EMAIL PROTECTED] wrote:
Author: spitz
Date: Mon Mar 3 13:22:17 2008
New Revision: 23410
URL: http://www.lyx.org/trac/changeset/23410
Log:
* src/insets/InsetLabel.cpp (doDispatch):
- restore call to changeRefsIfUnique.
The references were not updated if the label name was changed.
Abdelrazak Younes wrote:
The references were not updated if the label name was changed.
Then this is a bug of in InsetLabel::update() because the code for this
is present in there.
looks like it.
Jürgen
Abdelrazak Younes wrote:
But it used to work fine, including for references or labels in child
documents.
I cannot imagine how I could have broken it. But I'll recheck rev. 23405.
Jürgen
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
But it used to work fine, including for references or labels in child
documents.
I cannot imagine how I could have broken it. But I'll recheck rev. 23405.
I didn't accuse you :-), maybe I broke it before committing this stuff
I am simply
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
The references were not updated if the label name was changed.
Then this is a bug of in InsetLabel::update() because the code for this
is present in there.
looks like it.
But it used to work fine, including for references or labels in
Abdelrazak Younes wrote:
I cannot imagine how I could have broken it. But I'll recheck rev. 23405.
I didn't accuse you :-), maybe I broke it before committing this stuff
Looks like it (I just checked)
I am simply saying that fixing update() might be a better idea.
Sure, but I tried and
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
I cannot imagine how I could have broken it. But I'll recheck rev. 23405.
I didn't accuse you :-), maybe I broke it before committing this stuff
Looks like it (I just checked)
I am simply saying that fixing update() might be a better
Abdelrazak Younes wrote:
I think the reason is the const_iterator used in InsetLabel::update(). I
guess gcc silently fail to update the InsetRef pointers because of that.
So I'll switch a non const iterator and revert your fix. Please tell me
if that's OK.
No, does not work either.
Jürgen
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
I think the reason is the const_iterator used in InsetLabel::update(). I
guess gcc silently fail to update the InsetRef pointers because of that.
So I'll switch a non const iterator and revert your fix. Please tell me
if that's OK.
No, does
Abdelrazak Younes wrote:
OK, I'll put back your change.
Wait ...
Jürgen
Jürgen Spitzmüller wrote:
OK, I'll put back your change.
Wait ...
It works if you call update from outside. E.g., if you change in the cut and
paste check I introduced:
- lab.update(oldname, false);
+ lab.update(oldname, true);
the refs are updated on paste (which is unwanted in this case
Jürgen Spitzmüller wrote:
Jürgen Spitzmüller wrote:
OK, I'll put back your change.
Wait ...
It works if you call update from outside. E.g., if you change in the cut and
paste check I introduced:
- lab.update(oldname, false);
+ lab.update(oldname, true);
the refs are updated on paste
Abdelrazak Younes wrote:
Could you please remove it and always pass
true?
- void update(docstring const new_label, bool updaterefs = true);
+ void update(docstring const new_label, bool updaterefs);
It does not even work if I remove the bool altogether.
I have a similar problem with my
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
Could you please remove it and always pass
true?
- void update(docstring const new_label, bool updaterefs = true);
+ void update(docstring const new_label, bool updaterefs);
It does not even work if I remove the bool altogether.
I have a
Abdelrazak Younes wrote:
This strikes me like the identical bug.
Hum, that might just be a redraw problem...
But the name is not changed indeed (as view source reveals).
Jürgen
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
This strikes me like the identical bug.
Hum, that might just be a redraw problem...
But the name is not changed indeed (as view source reveals).
I am out of idea then...
Abdel.
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
This strikes me like the identical bug.
Hum, that might just be a redraw problem...
But the name is not changed indeed (as view source reveals).
Another idea:
void InsetCommand::metrics(MetricsInfo mi, Dimension dim) const
{
if
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
Could you please remove it and always pass
true?
- void update(docstring const new_label, bool updaterefs = true);
+ void update(docstring const new_label, bool updaterefs);
It does not even work if I remove the bool altogether.
And if
Abdelrazak Younes wrote:
It does not even work if I remove the bool altogether.
And if you rename the method to say, updateCommand()?
No (but I prefer that name).
Jürgen
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
It does not even work if I remove the bool altogether.
And if you rename the method to say, updateCommand()?
No (but I prefer that name).
Another idea... This might be related to the way we update the dialogs.
For example, I was debugging
Abdelrazak Younes wrote:
Another idea... This might be related to the way we update the dialogs.
Good news: it works now.
Jürgen
[EMAIL PROTECTED] wrote:
Author: spitz
Date: Mon Mar 3 13:22:17 2008
New Revision: 23410
URL: http://www.lyx.org/trac/changeset/23410
Log:
* src/insets/InsetLabel.cpp (doDispatch):
- restore call to changeRefsIfUnique.
The references were not updated if the label name was changed.
Abdelrazak Younes wrote:
> > The references were not updated if the label name was changed.
>
> Then this is a bug of in InsetLabel::update() because the code for this
> is present in there.
looks like it.
Jürgen
Abdelrazak Younes wrote:
> But it used to work fine, including for references or labels in child
> documents.
I cannot imagine how I could have broken it. But I'll recheck rev. 23405.
Jürgen
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
But it used to work fine, including for references or labels in child
documents.
I cannot imagine how I could have broken it. But I'll recheck rev. 23405.
I didn't accuse you :-), maybe I broke it before committing this stuff
I am simply
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
The references were not updated if the label name was changed.
Then this is a bug of in InsetLabel::update() because the code for this
is present in there.
looks like it.
But it used to work fine, including for references or labels in
Abdelrazak Younes wrote:
> > I cannot imagine how I could have broken it. But I'll recheck rev. 23405.
>
> I didn't accuse you :-), maybe I broke it before committing this stuff
Looks like it (I just checked)
> I am simply saying that fixing update() might be a better idea.
Sure, but I tried
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
I cannot imagine how I could have broken it. But I'll recheck rev. 23405.
I didn't accuse you :-), maybe I broke it before committing this stuff
Looks like it (I just checked)
I am simply saying that fixing update() might be a better
Abdelrazak Younes wrote:
> I think the reason is the const_iterator used in InsetLabel::update(). I
> guess gcc silently fail to update the InsetRef pointers because of that.
> So I'll switch a non const iterator and revert your fix. Please tell me
> if that's OK.
No, does not work either.
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
I think the reason is the const_iterator used in InsetLabel::update(). I
guess gcc silently fail to update the InsetRef pointers because of that.
So I'll switch a non const iterator and revert your fix. Please tell me
if that's OK.
No, does
Abdelrazak Younes wrote:
> OK, I'll put back your change.
Wait ...
Jürgen
Jürgen Spitzmüller wrote:
> > OK, I'll put back your change.
>
> Wait ...
It works if you call update from outside. E.g., if you change in the cut and
paste check I introduced:
- lab.update(oldname, false);
+ lab.update(oldname, true);
the refs are updated on paste (which is unwanted in this
Jürgen Spitzmüller wrote:
Jürgen Spitzmüller wrote:
OK, I'll put back your change.
Wait ...
It works if you call update from outside. E.g., if you change in the cut and
paste check I introduced:
- lab.update(oldname, false);
+ lab.update(oldname, true);
the refs are updated on paste
Abdelrazak Younes wrote:
> Could you please remove it and always pass
> true?
>
> - void update(docstring const & new_label, bool updaterefs = true);
> + void update(docstring const & new_label, bool updaterefs);
It does not even work if I remove the bool altogether.
I have a similar problem
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
Could you please remove it and always pass
true?
- void update(docstring const & new_label, bool updaterefs = true);
+ void update(docstring const & new_label, bool updaterefs);
It does not even work if I remove the bool altogether.
I have
Abdelrazak Younes wrote:
> > This strikes me like the identical bug.
>
> Hum, that might just be a redraw problem...
But the name is not changed indeed (as view source reveals).
Jürgen
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
This strikes me like the identical bug.
Hum, that might just be a redraw problem...
But the name is not changed indeed (as view source reveals).
I am out of idea then...
Abdel.
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
This strikes me like the identical bug.
Hum, that might just be a redraw problem...
But the name is not changed indeed (as view source reveals).
Another idea:
void InsetCommand::metrics(MetricsInfo & mi, Dimension & dim) const
{
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
Could you please remove it and always pass
true?
- void update(docstring const & new_label, bool updaterefs = true);
+ void update(docstring const & new_label, bool updaterefs);
It does not even work if I remove the bool altogether.
And if
Abdelrazak Younes wrote:
> > It does not even work if I remove the bool altogether.
>
> And if you rename the method to say, updateCommand()?
No (but I prefer that name).
Jürgen
Jürgen Spitzmüller wrote:
Abdelrazak Younes wrote:
It does not even work if I remove the bool altogether.
And if you rename the method to say, updateCommand()?
No (but I prefer that name).
Another idea... This might be related to the way we update the dialogs.
For example, I was debugging
Abdelrazak Younes wrote:
> Another idea... This might be related to the way we update the dialogs.
Good news: it works now.
Jürgen
42 matches
Mail list logo