-----Original Message----- Perhaps, but then the general/gcc-style parsing would have to be duplicated into the Java case. I don't use Java myself, but I imagine that when compiling with gcj the error message style is the same as gcc.
It shouldn't do. The if statement checks to see if it's Java AND if we have ant style output. Since the "break" is inside the if, if it's not Java AND ant (ie. just Java) it'll carry on through until it reaches the default case.
Not that it really matters from a functionality point of view, but since you've created a java case, you might as well use it. :-)
The output from javac is syntactically the same as that of gcc. I've never used gcj but I would assume it's the same.
Jon
On 08/16/2007 09:14:08 PM, Jon Senior wrote:
-----Original Message----- Perhaps, but then the general/gcc-style parsing would have to be duplicated into the Java case. I don't use Java myself, but I
imagine
that when compiling with gcj the error message style is the same as gcc.
It shouldn't do. The if statement checks to see if it's Java AND if we have ant style output. Since the "break" is inside the if, if it's not Java AND ant (ie. just Java) it'll carry on through until it reaches the default case.
Not that it really matters from a functionality point of view, but since you've created a java case, you might as well use it. :-)
Well there's a lot of filetypes that use gcc-style error messages. You're suggesting moving the Java case above all the others so it can fall through for the javac/gcj case. This would be neater at the moment, but what happens when one of the other 6 filetypes needs special behaviour - we can't have more than one fall through. That's why I was thinking about duplicating the default parsing.
Anyway, hopefully sometime we'll switch to using regexes.
Regards, Nick