Re: [PATCH] Avoid a divide by zero crash.

[Dmitry Timoshkov]

Max Woodbury  wrote:
> While tmHeight is not _supposed_ to ever be zero, it is an external
>
> input and has been found to be zero in a few rare fonts.  Detecting
>
> and fixing THAT problem should be handled elsewhere.  This change just
>
> keeps the application from being blown away by a divide by zero
>
> exception.

This change needs at least more details and preferrably a test case.

[PATCH] Avoid a divide by zero crash.

[Max Woodbury]

On 03/01/2014 04:24 AM, Dmitry Timoshkov wrote:
> Max Woodbury  wrote:
> > While tmHeight is not _supposed_ to ever be zero, it is an external
> >
> > input and has been found to be zero in a few rare fonts.  Detecting
> >
> > and fixing THAT problem should be handled elsewhere.  This change just
> >
> > keeps the application from being blown away by a divide by zero
> >
> > exception.
>
> This change needs at least more details and preferrably a test case.

The test case was submitted about a year ago.

What more details do you think are needed?

tmHeight is an external input, so it can have any value including
zero.  As such, it should NOT be used as a divisor.  This change uses
it as a multiplier on the other side of the comparison so it should not
cause THAT particular problem.  A little additional algebra simplifies
the code.

max

[PATCH] Avoid a divide by zero crash.

[Dmitry Timoshkov]

Max Woodbury  wrote:
> > > While tmHeight is not _supposed_ to ever be zero, it is an external
> > >
> > > input and has been found to be zero in a few rare fonts.  Detecting
> > >
> > > and fixing THAT problem should be handled elsewhere.  This change just
> > >
> > > keeps the application from being blown away by a divide by zero
> > >
> > > exception.
> >
> > This change needs at least more details and preferrably a test case.
>
> The test case was submitted about a year ago.
>
> What more details do you think are needed?
>
> tmHeight is an external input, so it can have any value including
> zero.  As such, it should NOT be used as a divisor.  This change uses
> it as a multiplier on the other side of the comparison so it should not
> cause THAT particular problem.  A little additional algebra simplifies
> the code.

Sounds like a deja-vu.

[PATCH] Avoid a divide by zero crash.

[Max Woodbury]

On 03/01/2014 05:07 AM, Dmitry Timoshkov wrote:
> Max Woodbury  wrote:
> > > > While tmHeight is not _supposed_ to ever be zero, it is an external
> > > >
> > > > input and has been found to be zero in a few rare fonts.  Detecting
> > > >
> > > > and fixing THAT problem should be handled elsewhere.  This change just
> > > >
> > > > keeps the application from being blown away by a divide by zero
> > > >
> > > > exception.
> > >
> > > This change needs at least more details and preferrably a test case.
> >
> > The test case was submitted about a year ago.
> >
> > What more details do you think are needed?
> >
> > tmHeight is an external input, so it can have any value including
> > zero.  As such, it should NOT be used as a divisor.  This change uses
> > it as a multiplier on the other side of the comparison so it should not
> > cause THAT particular problem.  A little additional algebra simplifies
> > the code.
>
> Sounds like a deja-vu.

8-) and your question was answered!?!

[PATCH] Avoid a divide by zero crash.

[Dmitry Timoshkov]

Max Woodbury  wrote:
> > > > > While tmHeight is not _supposed_ to ever be zero, it is an external
> > > > >
> > > > > input and has been found to be zero in a few rare fonts.  Detecting
> > > > >
> > > > > and fixing THAT problem should be handled elsewhere.  This change just
> > > > >
> > > > > keeps the application from being blown away by a divide by zero
> > > > >
> > > > > exception.
> > > >
> > > > This change needs at least more details and preferrably a test case.
> > >
> > > The test case was submitted about a year ago.
> > >
> > > What more details do you think are needed?
> > >
> > > tmHeight is an external input, so it can have any value including
> > > zero.  As such, it should NOT be used as a divisor.  This change uses
> > > it as a multiplier on the other side of the comparison so it should not
> > > cause THAT particular problem.  A little additional algebra simplifies
> > > the code.
> >
> > Sounds like a deja-vu.
>
> 8-) and your question was answered!?!

I'm afraid that nothing has changed since last time.

[PATCH] Avoid a divide by zero crash.

[Max Woodbury]

On 03/01/2014 05:16 AM, Dmitry Timoshkov wrote:
> Max Woodbury  wrote:
> > > > > > While tmHeight is not _supposed_ to ever be zero, it is an external
> > > > > >
> > > > > > input and has been found to be zero in a few rare fonts.  Detecting
> > > > > >
> > > > > > and fixing THAT problem should be handled elsewhere.  This change just
> > > > > >
> > > > > > keeps the application from being blown away by a divide by zero
> > > > > >
> > > > > > exception.
> > > > >
> > > > > This change needs at least more details and preferrably a test case.
> > > >
> > > > The test case was submitted about a year ago.
> > > >
> > > > What more details do you think are needed?
> > > >
> > > > tmHeight is an external input, so it can have any value including
> > > > zero.  As such, it should NOT be used as a divisor.  This change uses
> > > > it as a multiplier on the other side of the comparison so it should not
> > > > cause THAT particular problem.  A little additional algebra simplifies
> > > > the code.
> > >
> > > Sounds like a deja-vu.
> >
> > 8-) and your question was answered!?!
>
> I'm afraid that nothing has changed since last time.

:-(

Last time you said you would fix the problem elsewhere.  You have NOT 
done that.  The problem persists for those with the unusual fonts 
installed.  It still needs a fix.  You can still make the other changes 
when you get around to it.  That is NOT a reason for rejecting this patch.