Scott Wallace wrote:

Hi, Karl,

Nice idea!

Four quick observations:

(1) The little green up/down carets should not occur on the comma tile, since there are no alternative operators to cycle among in this case.

Right.

(2) The balloon-help message that one gets for the comma tile is not right.

Yup.

(3) The extension arrow should occur not only on string-valued *constant* tiles, but also on string-valued *variable* tiles -- cf. how this is handled in the case of numeric-valued variable tiles.

I have also noticed that the string variables can't use the text category tiles so you can't get the first letter of a variable etc. I'll see if I can fix that.

(4) The extension arrow for a string-valued tile in a string-assignment phrase (e.g. in "client's firstName <- Melvin") should appear *after* the string constant, not before. If you expand (i.e. hit the right-facing green arrowhead) and then contract (i.e. hit the left-facing green arrowhead,) the correct effect happens. [This same bug occurs with the handling of numeric-valued tiles as well, so if you happened to be fixing it for the string case, perhaps you could fix it for the numeric case as well ;-)]

I noticed this. Hopefully it's something as simple as using addMorphBack: instead of addMorph: somewhere :-)

Karl

Cheers,

-- Scott

At 11:19 PM +0200 9/22/03, Karl Ramberg wrote:

...

"Change Set: ExtendableText Date: 22 September 2003 Author: Karl Ramberg

Adds #, (comma) as an operator in Etoys strings so strings can grow."!

Scott can you have a look at the fix of karl? So that I approve it:)

Stef

Hi guys

I moved this change in the unstable stream even if there were two conflicts, but it seems to me that the code was equivalent. Could you check the two following methods:

Checking extendableText6-kfr.cs (#5963) for method conflicts with changesets after The Worlds of Squeak ... ...checking TextMorph>>getFirstCharacter CONFLICT found for TextMorph class>>getFirstCharacter with newer changesets 6012-1016nonASCIIs 6056-40236firstLastChar ...checking TileMorph>>showSuffixChoices ...checking CategoryViewer>>makeSetter:event:from: ...checking StandardScriptingSystem>>arithmeticalOperatorsAndHelpStrings ...checking StandardScriptingSystem>>doesOperatorWantArrows: ...checking StandardScriptingSystem>>helpStringOrNilForOperator: ...checking StandardScriptingSystem>>helpStringForOperator: CONFLICT found for String class>>newTileMorphRepresentative with newer changeset 5992-002-bootstrap 2 conflicts found.

excellent! We move....

I don't know the algorithm of the conflict checker, but these are equivalent but in different encoding. In the other words, they are under control.