On Mon, Oct 13, 2008 at 12:14:44PM +0200, Uwe Stöhr wrote: > Andre Poenitz schrieb: > >>> Maybe I am wrong, but you cannot assume an encoding from the length >>> of a comment. >> >> There are a lot more serious assumptions in the tex2lyx code. Given >> any data in an unknown encoding you will have to guess what it is. > > But your guess is not correct. You assume an encoding by a comment > given in a file, this is an ugly hack solving only the import of one > single file. The comment you check is not automatically inserted by > XeTeX, so your "magic" only works for one special file. What if the > user doesn't insert a comment, what when he only uses "Use XeTeX for > this file!" or "XeTeX is required for this" ...?
According to the bug description that's not a single file but all files that come from a XeTeX/kile combination. Rest assured that I will put in any ten line hack that's "mostly harmless" and lets us read all files that come from MS Word ;-} >>> So you introduced a bug when I use the same comment but use plain >>> ASCII. >> >> Plain ASCII is a subset of UTF8. It's hard to see what can go wrong >> here. Also note that the assumption only kicks in when the encoding >> is "auto". > > I see, so you are right that it doesn't hurt, but it only helps for > one single file, not in general. See above. Price/value ratio is good enough for me in this case. >>> Besides this, XeTeX support is planned for LyX 1.7, LyX 1.6 is >>> feature frozen. So I don't understand what you tried to solve with >>> your patch. >> >> Bug #<whatever>, marked as "serious showstopper" by someone? > > You rever to bug 3035, right? This is the showstopper, but you don't > fix it. Bug 4299 was marked as duplicate of this. I'm still confused > what you intention was. Fix a bug? And #4299 was in the line above #3035 in your mail. Sorry for reading top-down... Apart from that I still don't understand what #3035 is actually refering to. That's a parser problem leading to some additional whitespaces in the produced output. Not nice. But not a showstopper, and not really related to unicode unless the intention is to use unicode for stuff outside the 7bit range (which would be fine, but which we haven't done so far, and which isn't hinted at in the bu description either). Andre'
