ikarus.io.ss bugs?

[Derick Eddington]

refill-bv-buffer: lines 714-716 use j but should they be using c1?

lookahead-char-utf8-mode: line 840 uses get-char but should it be? get-char will see the same decoding error and ignore it by advancing passed it and doing another get-char which will advance more, but since this was lookahead-char should it be consuming characters? I.e.:

> (define tip (open-bytevector-input-port '#vu8(#b10000000 65 66 67) (make-transcoder (utf-8-codec) (eol-style none) (error-handling-mode ignore))))
> (lookahead-char tip)
#\A
> (lookahead-char tip)
#\B
> (lookahead-char tip)
#\B
>

I know the port position stuff is not finished, but I noticed these:

refill-bv-buffer: lines 706-708: should refilling the buffer be advancing the port's position?

Should these give 0?
> (input-port-byte-position (open-string-input-port "foobar"))
1
> (input-port-byte-position (open-bytevector-input-port '#vu8(1 2 3 4 5)))
1
>

[Abdulaziz Ghuloum (aghuloum)]

> refill-bv-buffer: lines 714-716 use j but should they be using c1?

Fixed. Thanks.

> lookahead-char-utf8-mode: line 840 uses get-char but should it be?

Nop. It was supposed to be lookahead-char.

> I know the port position stuff is not finished, but I noticed these:
>
> refill-bv-buffer: lines 706-708: should refilling the buffer be
> advancing the port's position?

Yes, that's the only place where the port-position field advances.

> Should these give 0?
>> (input-port-byte-position (open-string-input-port "foobar"))
> 1
>> (input-port-byte-position (open-bytevector-input-port '#vu8(1 2 3
>> 4 5)))
> 1

I don't know. I added the 1 explicitly in line 239 (vi bias again).
BTW, this is not the port-position of R6RS, right?

Aziz,,,

[Derick Eddington (derick-eddington)]

> > Should these give 0?
> >> (input-port-byte-position (open-string-input-port "foobar"))
> > 1
> >> (input-port-byte-position (open-bytevector-input-port '#vu8(1 2 3
> >> 4 5)))
> > 1
>
> I don't know. I added the 1 explicitly in line 239 (vi bias again).
> BTW, this is not the port-position of R6RS, right?

Ah, I see where it's used now. My editor counts chars starting at 1 too, and all the character positions I've jumped to have been correct. I'm so used to thinking in 0-based indexing... :) just thought to mention it.

[Abdulaziz Ghuloum (aghuloum)]

I'm going to drop input-port-byte-position from the exports of (ikarus) because this is really an internal function.