issue-#95 TextMapper Scanner & SETS

[Robert]

It is not a bug - it is a feature.

(In this quote I have changed a word to "feature" as I think this is what you meant!)

I guess we will never know if it is a deliberate feature.

I think the current situation is unsatisfactory. If it is a feature we should update the documentation to describe it. If it is a bug we should remove it.

Personally I much prefer to regard it as a bug.

[Ivan Denisov]

I think, that we should vote about this. Because we definitely have two opinions.
1. {1, 2 3} should abort compilation with error mark
2. {1, 2 3} should not abort compilation

[Ivan Denisov]

Robert, does the ready solution exist already?

[Josef Templ]

In my opinion, this topic is not a real problem and does not need a change and therefore does not need a vote.
We have enough to do to catch up with the backlog of changes in the CPC 1.7 change list.

Please keep focused on real bugs and improvements.

- Josef

[Robert]

I think, that we should vote about this. Because we definitely have two opinions.
1. {1, 2 3} should abort compilation with error mark
2. {1, 2 3} should not abort compilation

It is not a problem with compilation; the compiler does reject invalid sets. It is when you use a Scanner directly that this issue occurs.

Robert, does the ready solution exist already?

My suggestion is in my post dated 4-Oct. You need to change the name from Sett to Set.

[Josef Templ]

I know that it is not in the compiler but in the text scanner.
See my previous posting http://forum.blackboxframework.org/view ... =295#p2534.
There is no need for a change. Just leave it as it is unless there is a problem.
Having a more tolerant syntax here is not a problem but could actually be considered a feature.

- Josef

[Ivan Denisov]

Josef, I the solution exist, we do not need to suspend the issue. So I will prepare the branch for voting on this issue.
If you think, that it is not important issue, you can vote No.

[Josef Templ]

If we create an issue for every little question or personal preference a center member has we waste so much time and energy that we will never come to a release. This should be understood by all center members.

[Robert]

If we create an issue for every little question or personal preference a center member has we waste so much time and ...

This wasted my time in the past as I thought that I could use the BlackBox code as a model for a private routine without realising that the Scanner code did not do what I expected - ie recognise correctly formatted SETs. I now have the routine I need, so it is no longer a problem for me. The reason for considering correcting the existing code is to avoid wasting other people's time.
The correction could consist of modifying the code to comply with the Docu, or modifying the Docu to explain the syntax that interpretSets does recognize.

I prefer the first option.

[Ivan Denisov]

Please, take a look at the changes.
http://redmine.blackboxframework.org/pr ... 22ba0cbbbc

I changed the LOOP statement to WHILE. Because LOOP is rarely used in BlackBox and also it is excluded from new revision of Oberon. It is better not use it in all Center fixes if it is possible.

The test version:
http://blackboxframework.org/unstable/i ... a1.374.zip