[ticket #9290] Selection in nested insets is sometimes irreversible

[Alfredo Braunstein]

Continuing with a series of semi-infuriating navigation and selection
issues.

When starting a selection inside an inset (e.g. footnote) and extending to
the outside with shift-arrow keys, the selection correctly includes the
inset atomically. However, then it is impossible to re-enter the inset while
selecting, so exiting is irreversible.
This can be highly frustrating when editing e.g. complex formulae; a small
mistake and you exit the math inset and have to restart the selection from
scratch.
This works correctly *inside* mathed (you can reenter the inset where the
anchor is), so this behaviour is what the user would expect also in texted.
When selecting with the mouse, everything's fine, i.e. you can reenter the
inset.

A proposed fix is in attachment. Seems suspiciously simple... I'm worried of
course that it may introduce other issues (although from a quick test I
didn't see any). Could not test with tables because of #9291.

A/

[Scott Kostyshak]

Post by Alfredo Braunstein
Continuing with a series of semi-infuriating navigation and selection
issues.
When starting a selection inside an inset (e.g. footnote) and extending to
the outside with shift-arrow keys, the selection correctly includes the
inset atomically. However, then it is impossible to re-enter the inset while
selecting, so exiting is irreversible.
This can be highly frustrating when editing e.g. complex formulae; a small
mistake and you exit the math inset and have to restart the selection from
scratch.
This works correctly *inside* mathed (you can reenter the inset where the
anchor is), so this behaviour is what the user would expect also in texted.
When selecting with the mouse, everything's fine, i.e. you can reenter the
inset.
A proposed fix is in attachment. Seems suspiciously simple... I'm worried of
course that it may introduce other issues (although from a quick test I
didn't see any). Could not test with tables because of #9291.
A/

This is great! I have also been frustrated by this. Thank you for the
patch. It worked well for me in limited testing.

I think your bug report might be a duplicate of
http://www.lyx.org/trac/ticket/2346

Best,

Scott

[Alfredo Braunstein]

Post by Scott Kostyshak
This is great! I have also been frustrated by this. Thank you for the
patch. It worked well for me in limited testing.
I think your bug report might be a duplicate of
http://www.lyx.org/trac/ticket/2346

Yep, you are right, thanks. Humm, how do I mark it as duplicate of
2346? I see only an option of marking the resolution as "duplicate"
but no place to put the number of the original ticket.

A/

[Jean-Marc Lasgouttes]

Post by Alfredo Braunstein

Post by Scott Kostyshak
This is great! I have also been frustrated by this. Thank you for the
patch. It worked well for me in limited testing.
I think your bug report might be a duplicate of
http://www.lyx.org/trac/ticket/2346

Yep, you are right, thanks. Humm, how do I mark it as duplicate of
2346? I see only an option of marking the resolution as "duplicate"
but no place to put the number of the original ticket.

You just write it as a comment, and maybe do the same in #2346. Trac is
a bit minimalistic in some respects.

JMarc

[Jean-Marc Lasgouttes]

Post by Alfredo Braunstein
Continuing with a series of semi-infuriating navigation and selection
issues.
When starting a selection inside an inset (e.g. footnote) and extending to
the outside with shift-arrow keys, the selection correctly includes the
inset atomically. However, then it is impossible to re-enter the inset while
selecting, so exiting is irreversible.
This can be highly frustrating when editing e.g. complex formulae; a small
mistake and you exit the math inset and have to restart the selection from
scratch.
This works correctly *inside* mathed (you can reenter the inset where the
anchor is), so this behaviour is what the user would expect also in texted.
When selecting with the mouse, everything's fine, i.e. you can reenter the
inset.
A proposed fix is in attachment. Seems suspiciously simple... I'm worried of
course that it may introduce other issues (although from a quick test I
didn't see any). Could not test with tables because of #9291.

I am not sure of whether your patch is the right one, but we could apply
it and see what happens in master.

If you send a public ssh key to me, I'll give you git commit rights.

JMarc

[Alfredo Braunstein]

Post by Jean-Marc Lasgouttes
I am not sure of whether your patch is the right one, but we could apply
it and see what happens in master.

Of course the patch is correct! It is the rest of the code that may be wrong
;-)

Post by Jean-Marc Lasgouttes
If you send a public ssh key to me, I'll give you git commit rights.

I'll do.

A/

[Alfredo Braunstein]

Post by Jean-Marc Lasgouttes
I am not sure of whether your patch is the right one, but we could apply
it and see what happens in master.

Done. I've made the same change also to checkAndActivateInsetVisual, I
presume for visual navigation inside RTL, but I did not test it (I don't
know how).

A/