BlackBox Framework Center

Robert wrote:Docu: Stores
Procedure: WriteXString

Slightly off-topic: The entire Module SMath (SHORTREAL versions of Cos, Exp, etc) went missing from the Oms distribution around about the time of BlackBox 1.3. I use it (not for speed, but to save space doing big sums). Should we consider including it?

Leave it as it is. Both.

- Josef

Robert wrote:Docu: From Pascal to Component Pascal
Line # 9
Was: "Arrays are always defined over the integer range 0..max-1"
Should be: "Arrays are always defined over the integer range 0..max" or "Arrays are always defined over the integer range 0..count-1"

should be: 0..length-1

because the Language Report also uses the term 'length'.

- Josef

I agree, that "X procedures" can be kept for 1.7 with the existed "historical" documentation.

"Count" was changed to "length".

http://blackboxframework.org/unstable/i ... a1.412.zip

Robert wrote:Slightly off-topic: The entire Module SMath (SHORTREAL versions of Cos, Exp, etc) went missing from the Oms distribution around about the time of BlackBox 1.3. I use it (not for speed, but to save space doing big sums). Should we consider including it?

I do not think, that we should include it to basic distribution. It is better to make in available in some component like Lib.

As far as I see, SMath is and was always included in the distribution.
Only the docu file is very short. It says that it is the same as Math
but with SHORTREAL instead of REAL.
The only possible change would be to provide a long form of the docu.

- Josef

Josef Templ wrote:As far as I see, SMath is and was always included in the distribution.
Only the docu file is very short. It says that it is the same as Math
but with SHORTREAL instead of REAL.
The only possible change would be to provide a long form of the docu.

Apologies, it is there. And I was not thinking of padding out the Docu.


Maybe it was removed about 1.3, and reinstated later, maybe my memory is just playing me tricks.

I found a bug in the docu of TextMappers related to the syntax of Sets.

CONST interpretSets
Option element
Possible element of Scanner.opts. If present, the scanner will recognize set values: sets of integers in the range MIN(SET) .. MAX(SET) as defined by the language Component Pascal; otherwise "{", ".", and enclosed integers will be returned individually. (The syntax of set values is {" integer [".." integer "]" { "," integer [".." integer] "}.)

This syntax does not allow for empty sets. Also some double quotes are wrong.
The syntax must be be:
"{" [integer [".." integer] { "," integer [".." integer]] "}"

- Josef

CONST interpretSets
... otherwise "{", ".", and enclosed integers will be returned individually.

There is an interesting ambiguity here, which I have deliberately avoided mentioning as I thought the discussion on scanning SETs was intense enough!.

What does the "otherwise" refer to: the presence of the option "interpretSets", or the presence of a "set as defined by the language".

With the second interpretation even the new scanning procedure Set is not correct, and one could make an argument that it should be.
Rather than just setting the the type to invalid whenever the character "{" is found (and not part of a valid set) it should set the type to char and rewind to that point.

I miss in "Help" -> "Contents" list the link to the Map of the Ctl subsystem
and the link to document of the Direct-To-COM Compiler (Docu/DTC-Env.odc).
Map to the Std subsystem is also missing.
- Helmut

Robert wrote:

CONST interpretSets
... otherwise "{", ".", and enclosed integers will be returned individually.

There is an interesting ambiguity here, which I have deliberately avoided mentioning as I thought the discussion on scanning SETs was intense enough!.

What does the "otherwise" refer to: the presence of the option "interpretSets", or the presence of a "set as defined by the language".

With the second interpretation even the new scanning procedure Set is not correct, and one could make an argument that it should be.
Rather than just setting the the type to invalid whenever the character "{" is found (and not part of a valid set) it should set the type to char and rewind to that point.

"otherwise" clearly refers to the situation where "interpretSets" is not in the options.
The individual symbols of a set are only returned as individual symbols if the option is not set.
This also corresponds to the original implementation.
May be we need a different wording for avoiding any doubts.

@Ivan: the grammar in the docu needs to be fixed. It is simply wrong.

- Josef