[sword-devel] Xiphos : copy'n'paste to facebook produces Chinese !

Matthew Talbert ransom1982 at gmail.com
Wed Jun 30 11:57:11 MST 2010


> Karl, Xiphos's special copy setup is just that, special. Either way, no need to go berserk about it. Even if another system is to blame it is still a good idea to prove it to be such instead of "shooting the messenger" as it were.
> -Wes

There is no "special copy setup". In this sort of copy/paste
operation, the burden is typically on the receiver to interpret the
bytes correctly. In this case, it's Firefox/Facebook/Javascript which
are misinterpreting the data. The *same* data, pasted into Word,
pastes correctly.

This has been explained before in this thread, but the way copy/paste
works is that ^C merely signifies to the clipboard that there is data
available. It does not actually "copy" the data. When you hit ^V,
*then* the receiving application gets a pointer to the data, and
interprets it however it likes (though there are hints to tell it what
sort of data it is). So, no data is transferred until the receiving
application does it. The receiving application bears the
responsibility of interpreting the data correctly, except that the
original app is supposed to mark what kind of data it is. That is
obviously working, as the data gets interpreted correctly by Word,
Chrome, even Internet Explorer correctly, just not Firefox (when
pasting into Facebook).

We've been through this already in very technical detail. It will be
my last response to it. If there *is* a problem in the Xiphos stack,
it is in gtkhtml which I am working very hard to get rid of. I find
the possibility that there is an actual problem somewhere in Xiphos
very hard to accept*, however.

Matthew

Naturally, I just mean in regards to this issue; there are no doubt
other issues in Xiphos.



More information about the sword-devel mailing list