[Nano-devel] bug #1472421 on Red Hat

[Benno Schulenberg]

Hello Kamil,

To address https://bugzilla.redhat.com/show_bug.cgi?id=1472421 you seem to
have cherry-picked a patch from after nano-2.7.5 and applied it to 2.6.1.
(I'm not sure, because the commit you refer to fails to apply to 2.6.1.)

However, I suspect that that commit only makes sense together with the
other commits to that same area of code (but in time quite wide apart).

When I try the reproduction recipe given in the above bug with version
2.6.1, I see on my machine a high CPU usage not for just several seconds,
but for a *full minute*.  I haven't tried the versions in between, but
with 2.7.5 the time is down to seven or eight seconds.  With 2.8.1 there
is maybe one second of improvement.  So I doubt that the change you applied
to 2.6.1 is an actual correct fix.  It may avoid the problem of the original
bug poster, but it might have created/exposed other bugs.  Are you sure that
your patched 2.6.1 does not suffer from https://savannah.gnu.org/bugs/?50292
or https://savannah.gnu.org/bugs/?50293 ?

The main problem with pasting a big piece of text being so slow in nano, is
that it treats all input as if it's being typed in at the keyboard, so it
thinks it has to show all the changes to the text, and show them on screen.
In the end it seems that ncurses thinks: "fuck it, I'm not going to follow
all of these changes until you make up your mind and settle down".  To make
pasting work smoothly, we will have to implement 'bracketed paste mode', as
requested in https://savannah.gnu.org/bugs/?40060 .

Benno