On Tue, 2002-12-17 at 10:13, Jeffrey Stedfast wrote: > On Tue, 2002-12-17 at 02:04, Joe Marcus Clarke wrote: > > Hi. I have filed Ximian bug 34911 stating that FreeBSD users cannot > > copy and paste text from the Mozilla/Galeon main window into Evolution. > > Linux users do not seem to be affected. I did some digging, and I think > > I'm close to solving this problem. I just need a little help. > > > > If I edit gtkhtml-1.1.7/src/gtkhtml.c, and remove the text/html format > > from the formats array, this problem goes away. Therefore, the cause > > seems to be in the new text/html pasting code (i.e. the code that didn't > > exist in gtkhtml 1.0.4). Now, I played with the new html selection > > functions, but I am unable to make things work. I see that conversion > > is being done between ucs2 and utf-8 character sets. I'm curious as > > ucs2, or UCS2 for that matter, is not a registered character set with > > libiconv-1.8, nor is it statically mapped in gal's e-iconv.c. How does > > Linux handle this? If I add a mapping from ucs2 to UCS-2 in e-iconv.c, > > the problem is not solved. And I'm not sure why gtkhtml assumes > > Mozilla's character set is UCS2. I personally use ISO-8859-1 (though no > > matter which charset I choose in Mozilla, the pasting problem remains). > > Mozilla's export charset is always UCS2 (aka Unicode16 aka > iso-10646-1?). And it's internal charset is likely also unicode, despite > what charset preferences you have chosen (although their internal > charset *might* be utf-8, just like evolution's). > > Anyways, my iconv seems to be able to handle the following: > > UCS-2, UCS-2LE, UCS-2BE, UCS2, > UTF-16LE, UTF-16BE, UTF16, UTF16LE, UTF16BE
Mine seems to be able to handle all but UCS2. Are you saying you see UCS2 in your libiconv's aliases.gperf? Maybe I need to add this. > > My guess is that you actually want one of the big-endian representations > if one of the aliases such as UCS-2 or UCS2 don't exist, but Larry would > know for sure. I noticed in gtkhtml.c that you check the order of the UCS2, and store that in tocode. However, you always use the encoding "ucs2" instead of what is stored in tocode. This has a tendency to produce a '?' as the first character when pasting certain text into Evo. > > > > > I have also tried to debug some of this the old fashioned way, but have > > found printf's nor g_warning's generated by the html-editor component > > are written to stdout/err. How can I print out debugging info from the > > html editor? > > you'll have to run the editor by hand, since if evolution has oafd start > it then oafd will redirect stdout/err to /dev/null. How would I do that? I started the test component, but the same problem occurred. How do I run a bonobo component by hand? > > > > > I'd really like to get this resolved, and I feel I'm close. I just need > > a little guidance as to what is going on or what is being assumed of the > > host OS. Thanks. > > hope that helps, Thanks. It does give me some things to look at. Joe > > Jeff > > > > > Joe -- Joe Marcus Clarke FreeBSD GNOME Team :: [EMAIL PROTECTED] http://www.FreeBSD.org/gnome
signature.asc
Description: This is a digitally signed message part
