> The attached patch adds a new 'command' called
>
> 'command-sequence'
>
> This allows a key sequence to be bound to a seqeunce of other lyx-commands
>
> Example:
> \bind "F12" "command-sequence math-insert alpha; math-insert beta;"
>
> Every single command in the sequence has to be terminated by a semicolon.
> (Ok, this probably makes it difficult if you have a semicolon in your
> argument...)
I think this patch should go in! This is something that I wanted to
do a long time ago, but never had the time for. Also, hopefully
that will keep André morale up to send in lots of other pathces ;-)
And now for something completely different:
ObFridayInsult:
Regarding the forever ongoing SpaceLess discussion. I was
the one to introduce the SpaceLess function a long time ago, because
I was annoyed that it was not possible to have spaces in filenames,
simply because LaTeX is braindead. So I hacked up the SpaceLess
function to fix this, and people saw that it was good. Then, over
time, it turned out that LaTeX had a problem with other kinds of
characters, and slowly SpaceLess involved to a general mechanism
that converted problematic LaTeX characters to something else that
was tolerable. Then, it turned out that xdvi, shells, OS/2 and
other stuff had problems with other characters, and the SpaceLess
function was extended to convert all kinds of characters to underscores
or other substitutes.
The result of this process was that - was converted to _, and people
became rightfully angry. Then, a huge discussion emerged on this
list, and people argued that we should use & 0x7f, isgraph, isalpha,
this and that, should we use X or _ as substition, what about different
encodings, and so on. The discussion concerned the tiny small
details instead of the big picture.
On top of all this, Lgb sweeps away the discussion by insisting that
the SpaceLess function should only be used for temporary files.
Also, Juergen insists that this is not a problem for LyX, because
if people use a strange name, it's their own problem when they
export the file. Come on Juergen, you can do better ;-)
So, now we are in a position where the discussion is still going,
and the result might be that we convert all filenames to a string
of Xs or even _s, or not at all, and nobodyu will be happy.
Now, let's be a little productive instead, huh?
These are the issues:
1) The conversion is needed, because I want to be able to use
spaces and other stuff in *my* filenames. There is *no*
difference between temporary and exported filenames in this
regard.
2) We want to keep the conversion to a minimum. A filename like ___s____.tex
is no fun.
3) LaTeX, xdvi, ghostview, DocBook, OS/2, SGMLtools, the shell all have
different sets of problematic characters, and escaping rules.
4) The current code all uses the same SpaceLess function.
5) The export code all just spit out a file without asking for path
or filenames.
Now, where do we go from here?
Can we handle this very extremely hyper-super-difficult problem that
is completely and utterly impossible to solve?
Of course we can not, so let's just keep arguing that ; should be
converted to , or maybe it shold be :, or maybe we should use an
STL map to do this, or maybe a doubly linked hashed crossindexed
string of ropes with extra flags for encodings and strange influations
from other outer space. That should keep us occupied for a while.
Greets,
Asger
