Actual task: I want to print the matrix (of my own implementation) in humanly readable format. As a pre-requisite, I figured I need to be able to specify "fit the number representation into X characters". I found #printShowingDecimalPlaces: and #printPaddedWith:to: in Float and Integer classes (the first method is in more general Number class). Individually, they work, but the former works on fractional part only and the the latter on part before fractional, e.g.:
10.3 printPaddedWith: Character space to: 5.
"' 10.3'"
-10.3 printPaddedWith: Character space to: 5.
"' -10.3'"
10.3 printShowingDecimalPlaces: 3.
"'10.300'"
Also, their action on very large (or equally small numbers) in scientific form is not ideal:
12.3e9 printShowingDecimalPlaces: 3.
"'12300000000.000'"
12.3e9 printPaddedWith: Character space to: 5.
"' 1.23e10'"
So, I would like something like Common Lisp's (FORMAT T "~10g" 12.3d9) or C's printf("%10g", 12.3e9), that (a) restricts the whole width to 10 characters and (b) chooses the most suitable format depending on the size of the number. Is there something like this in Pharo?
For versatile printing options, I suggest loading NumberPrinter package from http://ss3.gemstone.com/ss/NumberPrinter/
(FloatPrinter fixed) digitCount: 2; print: 10.3.
-> '10.30'
I did not try it in recent Pharo versions though.
EDIT: Ah, but I see no format for handling exponents multiple of 3, maybe you would have to create a subclass for such format.
EDIT:
Or I missunderstood: you don't want it to print as '12.3e9' but rather '1.23e10'? note that apart significand digitCount, you need extra size for at worst 1 for sign + 1 for fraction separator + 1 for exponent letter + 1 for exponent sign + 3 for exponent (worst case for double precision floating point).
The more or less equivalent to g format would be something like this:
(FloatPrinter freeFormat)
totalWidth: 13; "size of the generated string"
digitCount: 6; "number of significant figures"
print: -12.3e-205.
->' -1.23e-204'