bug#20245: 24.5; c++-mode font-lock broken by the last commit to emacs24 branch

bug#20245: 24.5; c++-mode font-lock broken by the last commit to emacs24 branch

[Oleh Krehel]

Hello,

The font-locking of at least C++ headers was broken in "d69e9f1".
Here's a screenshot: http://i.imgur.com/eT6X3hP.png. You can see three
artifacts there. They change number and location when I scroll.

It goes back to a good state if I revert the commit and re-compile.
I hope this gets fixed by 24.5 release, even by just reverting
"d69e9f1", since this one is a bigger bug than
http://debbugs.gnu.org/cgi/bugreport.cgi?bug=19669, which "d69e9f1" fixes.

Oleh

bug#20245: 24.5; c++-mode font-lock broken by the last commit to emacs24 branch

[Alan Mackenzie]

Hello, Oleh.

In article <mailman.3171.1427958137.31049.bug-gnu-emacs@gnu.org> you wrote:

> The font-locking of at least C++ headers was broken in "d69e9f1".
> Here's a screenshot: http://i.imgur.com/eT6X3hP.png. You can see three
> artifacts there. They change number and location when I scroll.

This was fixed very recently in the master branch.  Would you please try
the same file in master to confirm that the fix is a fix.

> It goes back to a good state if I revert the commit and re-compile.
> I hope this gets fixed by 24.5 release, even by just reverting
> "d69e9f1", since this one is a bigger bug than
> http://debbugs.gnu.org/cgi/bugreport.cgi?bug=19669, which "d69e9f1" fixes.

> Oleh

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#20245: 24.5; c++-mode font-lock broken by the last commit to emacs24 branch

[Oleh Krehel]

Hi, Alan.


On Thu, Apr 2, 2015 at 11:51 AM, Alan Mackenzie <acm@muc.de> wrote:
> Hello, Oleh.
>
> In article <mailman.3171.1427958137.31049.bug-gnu-emacs@gnu.org> you wrote:
>
>> The font-locking of at least C++ headers was broken in "d69e9f1".
>> Here's a screenshot: http://i.imgur.com/eT6X3hP.png. You can see three
>> artifacts there. They change number and location when I scroll.
>
> This was fixed very recently in the master branch.  Would you please try
> the same file in master to confirm that the fix is a fix.

I just tried "23bb07e", the latest commit on master, but this issue isn't fixed.
The include statements look the same as in the screenshot.

Oleh

bug#20245: 24.5; c++-mode font-lock broken by the last commit to emacs24 branch

[Alan Mackenzie]

Hello, Oleh.

In article <mailman.3178.1427970015.31049.bug-gnu-emacs@gnu.org> you wrote:
> Hi, Alan.

> On Thu, Apr 2, 2015 at 11:51 AM, Alan Mackenzie <acm@muc.de> wrote:
>> In article <mailman.3171.1427958137.31049.bug-gnu-emacs@gnu.org> you wrote:

>>> The font-locking of at least C++ headers was broken in "d69e9f1".
>>> Here's a screenshot: http://i.imgur.com/eT6X3hP.png. You can see three
>>> artifacts there. They change number and location when I scroll.

>> This was fixed very recently in the master branch.  Would you please try
>> the same file in master to confirm that the fix is a fix.

> I just tried "23bb07e", the latest commit on master, but this issue isn't fixed.
> The include statements look the same as in the screenshot.

That's a pity.  It looks like there's a separate problem with the "fix"
from February.

Is there any chance you could send me a copy of the file which triggers
the bug, a description of exactly what you do to see it, and (unless you
start from emacs -Q) a copy of your CC Mode configuration (generated by
C-c C-b)?

Thanks!

> Oleh

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#20245: 24.5; c++-mode font-lock broken by the last commit to emacs24 branch

[Oleh Krehel]

Hi Alan,

On Thu, Apr 2, 2015 at 2:48 PM, Alan Mackenzie <acm@muc.de> wrote:

>> I just tried "23bb07e", the latest commit on master, but this issue isn't fixed.
>> The include statements look the same as in the screenshot.
>
> That's a pity.  It looks like there's a separate problem with the "fix"
> from February.
>
> Is there any chance you could send me a copy of the file which triggers
> the bug, a description of exactly what you do to see it, and (unless you
> start from emacs -Q) a copy of your CC Mode configuration (generated by
> C-c C-b)?

The bug is reproducible with emacs -Q. You can get an example file at:
https://github.com/dune-project/dune-pdelab-howto/blob/master/src/navier-stokes/cgstokes_instat.cc

To trigger the bug, just M->, and then scroll up with the mouse to the
start of the file where the headers are.  Random headers will be
partially highlighted with 'default instead of 'font-lock-string-face.

Oleh

bug#20245: 24.5; c++-mode font-lock broken by the last commit to emacs24 branch

[Alan Mackenzie]

Hello, Oleh.

On Thu, Apr 02, 2015 at 03:10:57PM +0200, Oleh Krehel wrote:
> The bug is reproducible with emacs -Q. You can get an example file at:
> https://github.com/dune-project/dune-pdelab-howto/blob/master/src/navier-stokes/cgstokes_instat.cc

> To trigger the bug, just M->, and then scroll up with the mouse to the
> start of the file where the headers are.  Random headers will be
> partially highlighted with 'default instead of 'font-lock-string-face.

Thanks.  I've got that file, and I do indeed see the bug in it.

> Oleh

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#20245: 24.5; c++-mode font-lock broken by the last commit to emacs24 branch

[Alan Mackenzie]

Hello, Oleh.

On Thu, Apr 02, 2015 at 12:19:36PM +0200, Oleh Krehel wrote:
> On Thu, Apr 2, 2015 at 11:51 AM, Alan Mackenzie <acm@muc.de> wrote:
> > In article <mailman.3171.1427958137.31049.bug-gnu-emacs@gnu.org> you wrote:

> >> The font-locking of at least C++ headers was broken in "d69e9f1".
> >> Here's a screenshot: http://i.imgur.com/eT6X3hP.png. You can see three
> >> artifacts there. They change number and location when I scroll.

> > This was fixed very recently in the master branch.  Would you please try
> > the same file in master to confirm that the fix is a fix.

> I just tried "23bb07e", the latest commit on master, but this issue isn't fixed.
> The include statements look the same as in the screenshot.

I've looked at this problem and now understand its cause.  It's too
involved to fix for Emacs 24.5, so we've decided to revert that commit,
as you've suggested.  A proper fix will surely be included in Emacs 25.1
(or 24.6, whichever comes first).

Thanks for taking the trouble to report this bug.  I'm closing it now.

> Oleh

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#20245: 24.5; c++-mode font-lock broken by the last commit to emacs24 branch

[Alan Mackenzie]

Hello, Oleh.

On Fri, Apr 03, 2015 at 12:46:53PM +0000, Alan Mackenzie wrote:

> On Thu, Apr 02, 2015 at 12:19:36PM +0200, Oleh Krehel wrote:
> > On Thu, Apr 2, 2015 at 11:51 AM, Alan Mackenzie <acm@muc.de> wrote:
> > > In article <mailman.3171.1427958137.31049.bug-gnu-emacs@gnu.org> you wrote:

> > I just tried "23bb07e", the latest commit on master, but this issue isn't fixed.
> > The include statements look the same as in the screenshot.

> I've looked at this problem and now understand its cause.  It's too
> involved to fix for Emacs 24.5, so we've decided to revert that commit,
> as you've suggested.  A proper fix will surely be included in Emacs 25.1
> (or 24.6, whichever comes first).

I've now committed a (hopefully) proper fix for this into the master
branch at savannah.  Feel free to download this and check that it does
indeed work properly.  If not ...

> > Oleh

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#20245: 24.5; c++-mode font-lock broken by the last commit to emacs24 branch

[Oleh Krehel]

Alan Mackenzie <acm@muc.de> writes:

Hi Alan,

> On Fri, Apr 03, 2015 at 12:46:53PM +0000, Alan Mackenzie wrote:
>
> I've now committed a (hopefully) proper fix for this into the master
> branch at savannah.  Feel free to download this and check that it does
> indeed work properly.  If not ...

It works great. Thanks a lot.

regards,
Oleh