SIGSEGV in jc1

Mo DeJong
Sat Apr 21 17:54:00 GMT 2001

On 21 Apr 2001, Per Bothner wrote:

> Tom Tromey <> writes:
> > A lot of people run into this.  Maybe we should address this before
> > 3.0.  Anybody have a suggestion for a fix?
> I seem to remember a discussing about something similar earlier.
> I think we should be lenient about bogus characters in comments,
> at least when the encoding is not explicitly specified.  One solution
> may be to ignore iconv errors in comments, except perhaps if
> -pedantic is specified.

I thought iconv was getting run over the whole file before
things were broken up into tokens. How would the compiler know
if an encoding error showed up in a comment, if comments
are not detected until a later stage?


More information about the Java mailing list