Problem with apostrophes and other special characters when using aspell in windows

Question

Hi there,

We seem to be having a problem with the spell checker on our content management system where it marks the ve part of We’ve as a misspelling. The spellchecker uses aspell which is called from a script on the server which executes the cmd.exe and uses it to pipe a file into aspell (it's a long winded way I know, but our server side programming langauge (ColdFusion) doesn't support writing to stdin for executables).

Aspell is called by executing:

c:\windows\system32\cmd.exe /c type d:\path_to_file\file.txt | "C:\Program Files\Aspell\bin\aspell" --lang=en -a

Where file.txt contains the text to be spelled e.g. ^Oh have We’ve (the carat is added to prevent piping problems I believe).

Aspell then output:

@(#) International Ispell Version 3.1.20 (but really Aspell 0.50.3)
*
*
*
& ve 62 12: vie, voe, V, v, veg, vet, Be, Ce, be, Ev, E, e, vex, VA, VI, Va, Vi, vi, we, VD, VF, VG, VJ, VP, VT, Vt, vb, vs, DE, De, Fe, GE, Ge, He, IE, Le, ME, Me, NE, Ne, OE, PE, Re, SE, Se, Te, Xe, he, me, re, ye, Ave, Eve, Ive, ave, eve, VAR, var, veer, vier, view, vow

However, we have a dev site, with the same version of Aspell, and when the same file is used it outputs with no misspellings. Both servers are running Aspell 0.50.3 on Windows server 2003, but there could be other differences in configuration:

@(#) International Ispell Version 3.1.20 (but really Aspell 0.50.3)

I'm wondering if the problem is to do with the piping part of the process or something different in the Aspell configuration. Does anyone have any ideas?

Cheers,

Tom

Answer 1

Since the header is being output correctly, I would guess that the pipe part is working, and I'd start by looking at the aspell configuration, or the test setup -- are you running exactly the same thing on the dev server and the production one?

I've seen problems before (not aspell specifically) with smart quotes, where it looks like a normal single quote character but actually it's a Unicode apostrophe or close-quote character. This can happen if text is pasted from MS Word or Outlook, for example. If you typed in the string manually on the dev server, you might have different strings.