Re: Image size error reporting

I think a common function to extract line number from the DOM is a good idea.

There isn't an image-specific position type; there's a "general" type.
I would simply omit <position> for errors that do not usefully pertain
to a position in the source document -- like in an image.

Sean

On 7/9/07, Laura Holmes <holmes@google.com> wrote:
> Noted.
>
> In regard to the error position, I simply meant that as of right now, the
> test results do not record position information, such as line number. I
> think that that the code for position reporting should be in the
> functions.xsl doc? We can make a Java call from within each individual XSL
> test in order to give line number to the result reporting function, and then
> we can add any other data we might want.
>
> I don't think I was aware of an image specific error position type. I think
> there was a link floating around that had some basic sample position
> results? If someone can re-point me to that, I can take a stab at adding
> this to the result function in the functions.xsl file. If there's any more
> information I should know about implementing general position error
> reporting, please send it my way as well.
>
> Laura
>
>
> On 7/7/07, Sean Owen <srowen@google.com> wrote:
> > Do whatever is easiest. If there are errors, the only real requirement
> > is to report the first error encountered accurately. Beyond that, all
> > else is a bonus. I am more concerned with getting everything basically
> > working, and not adding any more complexity.
> >
> > What's the issue with error position? we have a 'general' position
> > type that should always be used for image errors. There no such thing
> > as position for images -- I suppose you could refer to a byte that
> > didn't make sense but that doesn't seem useful or easy to obtain from
> > APIs.
> >
> > Sean
> >
> > On 7/6/07, Laura Holmes <holmes@google.com> wrote:
> > > Hi all,
> > > I'm working on the image resizing and specify test right now and I
> wanted to
> > > ask for opinions. Here are options for how it is implemented:
> > >
> > > 1) As is, we have a series of if statements testing both height and
> width
> > > (after we've established that indeed these tags are there). This means:
> > > - if we have a improperly formatted value, it records a test result
> value
> > > that reflects the improper formatting and a test result that either
> warns or
> > > fails depending on how the improper value is evaluated numerically
> > > - both width and height at evaluated simultaneously, so we can record
> the
> > > line number in which the error occurred, but not which value
> > >
> > > 2) We can choose not to proceed with the tests after it's been
> established
> > > that there's an improper value. However, if we stop here, if there's
> another
> > > issue with the properly formatted value, we never report that
> > >
> > > 3) We can test height and width values separately. If it is determined
> that
> > > the height or width is improperly formatted, we report that error, but
> > > continue testing the other value.
> > >
> > > Also, I'm still not quite sure what we determined about the position
> > > information format for error reporting. Roland, are you working on that?
> > >
> > > Thanks!
> > >
> > > Cheers,
> > > Laura
> > >
> >
>
>

Received on Monday, 9 July 2007 15:16:18 UTC