My todo list goes public...

[lyx.git] / src / mathed / BUGS

Items marked with

  //   - are probably fixed, the reporter is asked to verify this
         and report success or failure

  ??   - are not reproducable, the reporter is asked to verify this
         and report success or failure

  ::   - are questions or comments to the reporter, containing question
         on how to reproduce the bug exactly or things like that

  !!   - mark "not a bug, a feature" replies, usually with a request for
         further discussion

Unmarked items are known unfixed but probably unverified bugs.

----------------------------------------------------------------------

General hints for bug reports: 

  - keep the items small

  - do not provide excessive information on how to reproduce the bug
    if this is obvious

  - if a bug is partly fixed, report it as "fixed" and a new bug item
    that contains the unfixed part only

  - check this list regularly, comment on the marked items.

  - plain ASCII text please, not much more than 70 chars per column 

----------------------------------------------------------------------



Dekel:

Array/eqnarray

// - Ctrl+enter doesn't work

// - When creating a matrix/align/eqnarray all columns are centered.


Macros:

?? - When changing the macro definition, all instances are not updated.

:: I cannot reproduce this.

?? - If I have macros \newcommand{\foo}[1]{[#1]} and
??   \newcommand{\foox}{\foo{x}} then when an instance of \foox is drawn on
??   screen, there are two black frames.  While this is "correct", it is
??   annoying.

:: This is partially fixed?

- LyX crashes when you define a recursive macro

Misc:

- If I have x_{1}^{2}, I put the cursor before the 2, and press backspace,
  the result is 'x2'. The correct result should be 'x_{1}2'

- If I mark some text and then press ^, then the marked text is deleted.
  The correct behavior is to put the marked text in the superscript.

- Pressing <down> when cursor is in subscript will move the cursor up
  (to the base of the subscript) instead of exiting the inset.
  The same problem happen in fractions

!! If you type something there you will see that the corsor has left the
!! inset. That's mor or less a "feature". If you don't like it, propose a
!! method how a subscript inset can figure out what is "optically" lower
!! on screen.

// - When you press the mouse just to the left of the middle point of
//   some char, the cursor will be positioned to the right of the char.

- It is possible to put two or more consecutive spaces in math text mode

// - Text in superscript is not smaller than normal text.

?? - After insertion of \sum, the cursor is in invalid location
??    There is still a problem when doing c-m \sum S-C-M

?? - The height of the frame is often wrong (e.g. x^2)
??    For example $\sum x$ in an inline formula

- InsetFormula::validate is broken

// - Changing math space length by pressing the space-bar key doesn't work
!!  it's M-m space now. 

- The drawing of decorations on screen is very different than latex.
   (a low priority bug).

- If I change '\alpha' to bold, the result is \mathbf{a}

- Just moving the cursor in a math inset causes changing the buffer status
  to (changed).


Eran Tromer:

- When exiting the formula during selection, the anchor moves to the
  left of the formula and there's no way to go back.

// - Scripts are too large. $\log_2$ exceeds inset frame (non-displayed).

- You need two <right> to get past \vec{v}, and all sort of other
  insetish behavior.

- \vec{several-characters} doesn't work. \overrightarrow does.

- When the cursor is in a subscript (and either isn't at the end of 
  the subscript or there's no subsubscript), pressing <down> should
  try to *really* go down (to the next line, or denumerator, or 
  whatever). Currently it goes *up*. Ditto for superscripts.

:: Please see comment to Dekel's report on the same problem.

// - When choosing to insert "\left\Vert \right\Vert" parenthesis using
//   the maths [sic] panel, a "\left( \right." is inserted instead.

 - Formula inside tabular cell: red frame of cell inset and purple 
   frame of formula can overlap, causing leftovers when leaving 
   formula. To get an example, insert a 1x1 tabular into a new document,
   enter the cell its cell and press 
   M-m ( M-f 1 <right> <right> <right>       (zoom=100, screenDPI=100)

- When selecting multiple cells in a array using the keyboard, <left>
  etc. should can move whole cell at a time -- no need to navigate 
  within cells.

- When selecting, maybe give a visual indication of the "original" 
  anchor, when it differs from the "actual" one.

// - C-v always pastes into the end of the current math inset, instead of
//   the cursor location.

- Placement of cursor using mouse is seriously broken.
  Example 1: large fractions.
  Example 2: type into a new document: 
  "M-d xxxxxxxxx <right> M-x footnote-insert"
  Can't reach most places in the formula using mouse.



Rainer Dorsch:

- Entering \mathbf{c} in math mode is displayed as written (without
  backslash)


Marcus (Suran@gmx.net) 

// - In math-mode I can switch back to text-mode in a formula but then I am
//   not able to type Umlauts.

?? - Having selected some part of a formula and creating a fraction with that
??   part in the top my selection-buffer is replaced with the part.

- If the math-panel has the focus I can type text but not switch into
  math-mode or use some of the other keyboard-shortcuts.

- In the default-configuration there are shortcuts for al greek characters
  _except_ varphi wich is used extensively as a default-variable-name.


From: Álvaro Tejero Cantero <alvaro@antalia.com>
 
- the most annoying bug by far: bold is "inherited" to super and subscripts.
 
- when you get into equation array mode, the previous equation(s) remain(s) 
 unaligned. I would be soo much faster to have LyX pick the last equal sign 
 from the previous line as a reference for the alignment (and a good heuristic 
 for the 80% of cases, with the advantage that we don't lose anything 
 enforcing it). This is another case that slows down a lot when typing 
 because, as the previous one, requires going back and correcting.
 
- the movement is sometimes clumsy inside this arrayed equations: ie. C-a (or 
 "home" are understood as "beginning of the formula" instead of "beginning of 
 the equation". Again, a lot of movement has to be made. I suggest mapping 
 "beginning of doc" with "beginning of formula" and so on (very intuitive), in 
 case you don't deem appropriate the "namespace idea" below.
 
//- movement between macro boxes is also weird, but in this one it's perhaps 
// just me.  
 
- I suggest creating a different "kewybinding namespace" for the formulas, 
 since you could put to good use all those keybindings from the menu (M-?, 
 C-?) thath currently do their job PLUS getting you out of the formula. 
 Seriously, it'd be great to have more keys free, so M-d t would be time 
 derivative and M-d ? derivative with respect to the variable ?. And so on.
 
- I'm no experienced C++ programmer, but if you consider it appropriate, I 
 could write a scritp in python for this one (I'm also very optimistic). Tell 
 me what you think:
  
   Flattening macros. Sometimes it's annoying the fact that once you have 
 written a macro, you can't touch at it's "constant parts". I call flattening 
 to the process of substituting all macros with LaTeX code.
 
   Task: designing a macro substitution system that reads from a file 
 (possibly the same file as the document's) the macros and parses the document 
 doing the appropriate replacements
 
  This is very useful, because sometimes you have a big expression in a macro 
 and you want to change an index only. What do you do then?. You retype 
 everything (perhaps several times in the document) or you create extremely 
 generic and parametrizable macros that aren't very fast to fill in the 
 majority of cases.
 
 
- undo-redo seems too coarse in math-mode. You end up losing substantial 
 fragments of your formula. I hate that when it affects tensors,;).
 
- cut&paste inside math-mode doesn't work the X fashion (middle button doesn't 
 paste anything).
 
- I remember having heard that a search-replace function was planned, so I 
 won't repeat that. Only that the flattening option would be then easier to 
 implement on top of that.
 
- There's something I keep wondering about.. how do new math symbols get into 
 LyX. Do they get ever?. I don't know if it can be done, but if you tell me 
 where to "draw" those symbols or how, I would very glad start doing it. LyX 
 deserves to have support (visual support) for all the AMSTeX constructs. But 
 there must be something structural going on there, because otherwise more 
 math symbols would be ready by now.. just a thought.
 
- Some math symbols aren't very well supported (to my knowledge). I'm 
  thinking of underbraces with extra data in them,or [] options.


Herbert Voss:

- it's not possible to enter superscript when the
  ^-char works as a dead key. with the second ^-
  or the space the cursor jumps outside the mathbox.


Jules Bean:


- I've been using Lyx for a year or two now, and I have some thoughts on
  improving the interface. Not the GUI, as such, but the basic Human-Machine
  interface --- the efficiency with which I can enter information into it. 

There are a couple of general principles I'm try to uphold here: 

a) If something's easy & quick in LaTeX then it should be easy & quick in
LyX (unless it really isn't very common).  b) Actions which are used
frequently should be a single key-press, even if that's not very easy to
remember. You learn it.  c) Actions which are used rarely should be
mnemonic -- easy to remember -- even if they are multiple keypresses.
Things you do rarely you care less about the time it takes to perform. 

2) Math-mode entry  Under the xemacs keybindings, which I use, (damn it,
I don't want to confuse my poor muscle memory more than necessary by
learning too many different bindings for the same actions), the key
sequence to enter math-mode is 'M-c m'. I'm not sure exactly what to
propose instead, but I really think this needs to be only one keypress. 

As an aside, you may think that I'm whining over nothing. However, when
you enter math mode as often as I do (often more than once a sentence) it
gets very annoying, especially as compared to simply typing '$' in plain
emacs. Also, you may say 'why don't you just change your bindings file?'. I
will ;) but I wanted to start some discussion on this since it can benefit
everyone! 

More serious, though, than the number of characters which need to be typed
is the confusing nature of the command. 'M-c m', typed once, puts you into
math-mode. However, typing 'M-c m' again doesn't put you out of math-mode
--- it puts you into math-text mode. Then hitting it again puts you back
into normal math-mode.  IMO, 'modal' keys should either be idempotent (so
hitting it the second time does nothing) or self-inverting. In fact, the
inverse to 'M-c m' is either 'ESC' or simply a space typed at the end of
the block --- which is confusing, since they're not of the same 'shape' as
the command that got you in there. 

Now, I'm not saying that 'space' shouldn't be allowed as a short-cut to get
you out of math-mode; it's a most useful and natural one, I like it a lot.
However, on balance I think M-c m should also have that effect. 

3) Math-mode inconsistencies 

Sometimes 'the same action' has the same keystroke both within and
without math-mode. This is very sensible. However, it is very annoying when
they don't behave the way you're expecting them to. 

For example, 'M-c e' puts you into 'emphasise' mode.  Ignoring the fact
that in text mode this is italics, and in math-mode it stands for the
calligraphic character set, I think of these as the same action, so I like
the fact that they have the same keys. However, in math-mode, 'M-c e' is
idempotent, (and you need 'M-c space' to get back into normal) whereas in
text-mode 'M-c e' is self-inverse. These are the two possibilities I listed
as acceptable before, but consistency would be nice ;-) IMO, self-inverse
would be best for both. 

4) ERT in math-mode 

Math-mode doesn't handle ERT very well at all. It flips back out of
TeX-mode as soon as you type any special character.  Very annoying if I
wish to typeset an \xymatrix inside math-mode (see also my comment above
about table-like environments), since I can't type the &. In fact, my
xymatrices end up as $$\xymatrix{..}$$ in text-mode ERT. Also rather
annoying when you want to use a \macro with more than one argument (but not
fatally so). 

5) Proposal : a 'ligatures' or 'autocorrect' system 

One of the very minor, but useful, features of TeX is the way it lets you
type the nearest approximation to what you want using a 'typewriter
keyboard', and substitutes the typographically neat equivalent. In
particular, 'fancy' quotes (") and en and em dashes (---). I propose that
this UI element could be taken up a level into LyX, with a system that does
the following (for example): 

-> becomes \rightarrow 
<- becomes \leftarrow 
=> becomes \Rightarrow (etc..) 
==> becomes \Longrightarrow (etc..) 

This may only be appropriate in math mode, of course.  This family bug me
in particular because they take ages to type using a \-escape. Undoubtedly
sharp minds will think of others, and also we need some way of actually
typing those sequences as literals when we want them. 

6) Scope macros: 

The current macro system is clever, but could be neater. One improvement
I'd like is to let LyX know about TeX's scoping rules... 


Yves Bastide:

- use AMS's \text instead of \mbox.  It supports accented characters,
  among others...  (selected via validate()?)


Jean-Marc:

> \def\neq{\not=} \let\ne=\neq
>    \let\le=\leq
>    \let\ge=\geq
>     \let\owns=\ni
>    \let\gets=\leftarrow
>    \let\to=\rightarrow
> % \let\@@sqrt\sqrtsign
> \let\|=\Vert

Angus:

- make math lables editable