Re: [Bug-AUCTeX] 11.83; Wrong intendation and vertical shifting in math mode display

[Frank Küster]

Ralf Angeli <address@hidden> wrote:

> * Frank Küster (2006-07-24) writes:
>
>> In the following equation code, some things go wrong:
>
> Here is the equation for reference:
>
>   P(r) = 2 \int_0^{\pi} \frac{1}{2\pi} d\phi \cdot 
>          2 \int_0^{\pi} \frac{1}{2\pi} 
>                         \frac{
>                           4 \pi \left( 
>                             \frac{
>                               r - r_0
>                             }{
>                               \cos\theta
>                             }\right)^2
>                         }{
>                           \left(
>                             \nicefrac{2}{3}\pi \langle r^2\rangle
>                           \right)^{\nicefrac{3}{2}}
>                         }
>                         e^{
>                           -\frac{
>                             3\left(
>                               \frac{r-r_0}{\cos\theta}
>                             \right)^2
>                           }{
>                             2\langle r^2\rangle
>                           }
>                         }
>
>> - the closing brace for the exponent that is started by `e^{' in the
>>   ninth-last line is wrongly indented,
>
> If I hit TAB with point in that line it is indented by two more spaces
> (CVS Emacs, CVS AUCTeX).  The resulting position (as shown above)
> looks correct to me.

Oops, I now see that the closing brace is indented one step more - but
not enough, as on your system.  But it is even more strange and might
have to do something with tab settings.  If I copy the lines from
e^... to the closing brace and yank them in the mail buffer, I get
correct indentation:

                        e^{
                          -\frac{
                            3\left(
                              \frac{r-r_0}{\cos\theta}
                            \right)^2
                          }{
                            2\langle r^2\rangle
                          }
                        }

However, in the TeX buffer it looks like this (now manually adjusted:

                        e^{
                        -\frac{
                          3\left(
                            \frac{r-r_0}{\cos\theta}
                          \right)^2
                        }{
                          2\langle r^2\rangle
                        }
                      }

The only difference to the screenshot is that the last line is indented
only 2 columns too less, not four.

>> and the `-\frac' in the line
>>   following `e^{' should be indented by two more places, too.
>
> Why?  There is one level of parenthetical grouping which means it is
> indented (compared to indentation of the line before) the amount
> specified in `TeX-brace-indent-level' which is 2 by default.  That
> means indentation in the example is correct.

No, in the screenshot as well as in the manually adjusted code above the
e and the minus are in the same column, not indented by 2.  

>> - The index zero two lines later looks weird - the underscore is in the
>>   middle of the `0' instead near its bottom where it should be
>
> The 0 is displayed as subscript because of the underscore in front of
> it.

Yes, but the underscore which makes it a subscript is higher than the
0's baseline, or the 0 is not shifted downwards enough.

>>> - in the two occurences of `^2' below, the digit is shifted below the
>>   baseline instead above.
>
> No, it's not shifted at all compared to the text it appears in.  It's
> just made a little bit smaller.

No, look at the screenshot.  Or look at
http://www.kuesterei.ch/auctex2.png where I magnified the screenshot,
scrolled it so that the lower border works as a ruler, and took a new
screenshot.  You can see that the 2 extends below the baseline of the
word "right" on the left of it.

> The last two peculiarities will not appear in the next version of
> AUCTeX because font locking of subscript and superscript will be
> limited to one level.

Hm, I think that makes sense.  However, I insist that there is a bug in
the placement of these sub- and superscripts, and if the code is not
completely removed, we'd rather find the culprit.

Regards, Frank
-- 
Frank Küster
Single Molecule Spectroscopy, Protein Folding @ Inst. f. Biochemie, Univ. Zürich
Debian Developer (teTeX/TeXLive)