Re: [STUMP] Stumpwm crashes when there is incorrect utf-8 in windows title

[Alexey Lebedeff]

On Sat, 21 Jun 2008 02:50:17 -0700, Shawn Betts wrote:

 >> The app that created incorrect title was Licq.
 >> I made a small change in sbcl variant, so now invalid utf-8 is ignored.
 >>
 >> It's not perfect, because it replaces whole title with "<Invalid>" when
 >> encounters incorrect utf-8, but this is enough for me.

 SB> Is clisp affected by incorrect utf-8 from licq?

 SB> A slightly better fix might be to have a fallback function that just
 SB> converts each octet to a character and for any octet that doesnt have
 SB> a character use ?. It'll still look corrupted but at least something
 SB> half readable might come in..depending on the string. Anyone have
 SB> better ideas?

  Here is better version of utf8-to-string (sbcl part):

  #+sbcl (handler-bind
                         ((sb-impl::octet-decoding-error #'(lambda (c) 
(invoke-restart 'use-value "?"))))
                   (sb-ext:octets-to-string octets '(vector (unsigned-byte 8))) 
:external-format :utf-8)))


 SB> Maybe the sbcl error has enough data in it to deal with the corruption?