First of all, I assume (since you did not mention) that the setup screen did not show any error (when all fields were set)?
Errors would be shown in red, right below the 4 top edits. You can easily check that they appear, by e.g. setting language to "fr" when you don't have a french dict.
I tested with English and German.
I forgot to mention one important detail.
Because there isn't a full Unicode processor in SynEdit yet, it doesn't know what Unicode chars are "word chars" and wich are "word break chars" (e.g. all kind of spaces, or dashes that Unicode has)
So if you have chars outside A..Z then you need to enter them in the fields "Unicode upper/lower chars"
E.g.:
For German there should be: "äöü" and "ÄÖÜ"
For French would need all the accented chars...
It also checks source code. But you have settings to affect how. (and those apply for comments (and strings) too, since comments for many people contain identifier names. (e.g. class or procedure names)
It does ignore words that the HL knows. (Currently only the Pascal HL does report support that, and "known" are keywords and modifiers (stdcall). There are still some missings, e.g. AnsiString should be known (but is not), if the option to highlight "all string types" is off. (And the HL does not know those inside of comments or strings, though that is actually correct imho, as it can't tell if a comment is commented source or other text => stdcall is not a word in normal text)
As for checking comments/string only. That may still get added.
Settings:
It can check TSpellDictionary as "T" (ignored), "Spell" (ok) and "Dictionary" (ok)
=> The "T" can be in the field "List of chars (prefix) to ignore at the start of word"
Ignoring is not forced, but tested to match either: TRACTION is tested to be correct as "T","RACTION" or "TRACTION". (if either returns ok, then good)
Then ChkAnswer => Split into "Chk" and "Answer" (if it isn't a word in its entire): You can set to ignore parts that are 3 or shorter.
Recent
