Subject: Re: [xsl] 99 bottles of beer|
From: "Andrew Welch" <andrew.j.welch@xxxxxxxxx>
Date: Tue, 6 Feb 2007 09:41:24 +0000
> Moreover, I found that > (1 to 2) != 10 returns false, and (1,2) != 10 returns > true.... I am really missing something here, this must be a > faq somewhere :S ).
Well, I'm missing something too because (1 to 2) and (1,2) are the same sequence.
And Andrew Welch got it wrong in his reply too...
Yup, sorry about that - it seemed a plausible enough explanation for Saxon's behaviour (I've learnt that you can usually discount a Saxon bug early on). I too read the spec but couldn't find the part explaining this.
Interestingly if you change the version number to 1.0 you get true for (1 to 2) != 10, so it does point towards a bug.
Equally a sequence of booleans *doesn't* have an EBV (or any sequence starting with a boolean), so (true(), true()) doesn't return false.