From: Nicholas Clark Date: 15:01 on 01 Aug 2006 Subject: Spreadsheet::ParseExcel return pack('C*', unpack('n*', $sTxt)); No. That's not a valid way to convert ucs2 text to anything, let alone ISO-8859-1 Urge To Kill Rising [It's converting network order 16 bit values into a list of numbers, then converting that list to 8 bit characters, with an implicit truncation] Nicholas Clark
From: Jeremy Stephens Date: 17:20 on 01 Aug 2006 Subject: Re: Spreadsheet::ParseExcel Nicholas Clark wrote: > return pack('C*', unpack('n*', $sTxt)); > > No. That's not a valid way to convert ucs2 text to anything, let alone > ISO-8859-1 > > Urge > To > Kill > Rising > > [It's converting network order 16 bit values into a list of numbers, then > converting that list to 8 bit characters, with an implicit truncation] > > Nicholas Clark > > I ran into problems with UCS2 a couple weeks ago. I exported some data from MS SQL into a CSV file, which I then read from a scripting language (R, a statistical language) to do some stuff. R doesn't know how to read UCS2, which is hateful (since there's a Windows binary of it, you would think it would know how). Anyway, I ended up just doing the conversion with vim (using the fenc option), since it was a one-shot deal. That isn't really all that relevant to your hate, but I feel your pain nonetheless. Jeremy
From: Sean Conner Date: 18:34 on 01 Aug 2006 Subject: Re: Spreadsheet::ParseExcel It was thus said that the Great Jeremy Stephens once stated: > > I ran into problems with UCS2 a couple weeks ago. I exported some data > from MS SQL into a CSV file, which I then read from a scripting language > (R, a statistical language) to do some stuff. R doesn't know how to > read UCS2, which is hateful (since there's a Windows binary of it, you > would think it would know how). Anyway, I ended up just doing the > conversion with vim (using the fenc option), since it was a one-shot deal. You didn't have iconv installed? [1] -spc (iconv isn't all that hateful either [2]) [1] Is useful information here allowed? Forgive me if it isn't. [2] So I guess I really need to beg for forgiveness then.
From: Sean Conner Date: 18:31 on 01 Aug 2006 Subject: Re: Spreadsheet::ParseExcel It was thus said that the Great Nicholas Clark once stated: > return pack('C*', unpack('n*', $sTxt)); > > No. That's not a valid way to convert ucs2 text to anything, let alone > ISO-8859-1 "But ... but ... but ... it works for me!" [1] -spc ("We don't need to steeeking accented letters!") [1] Probably a programmer using US-ASCII only.
Generated at 10:26 on 16 Apr 2008 by mariachi