In SMath Solver 7921, it seems that some characters typed into string constants are coded even if they are in the basic ascii set.
In SMath Studio 7822 (latest installed version on my computer) this problem does not appear.

Here is an example which drives the user mad if he doesn't suspect about string encoding.

Even inspecting of the xml source in notepad++ doesn't reveal the difference, obviously the encoding is hidden from the user.
Only pasting from clipboard to notepad++ did help.
In the maxima translation process, a difference is seen on maxima side, yet it is hidden when back-translated.

IMHO string comparison or identity check in general should account for such encoding issues.

findrows.sm (11 KiB) downloaded 135 time(s).

Can this be a Microsoft character encoding issue?
I am also on v7921 and opened your file directly.
There is no issue on linux mono.

Regards