Re: [mpg123-devel] Please test new network code in/out of mpg123 (now with HTTPS support)!

[Dave Y. <dav...@gm...>]

Hi, back home.
On 05/17/22 09:05 AM, Thomas Orgis wrote:
> Am Tue, 17 May 2022 07:52:41 -0700
> schrieb Dave Yeo <dav...@gm...>:
>
>> Those were all one line. Widening the terminal to 90 characters gives
>> (one line);
>>  > 00104+10114  00:02.71+04:24.20 --- 100=100 112 kb/s  366 B acc    0
>> clip p+0.000
>>
>> Notice that previously the last 00 was cutoff.
>
> So with widened terminal, the line fits, but it does still scroll?

Had to double check as I previously pressed q too quick. No it doesn't 
scroll and works much as expected, though really I should test on Linux.
The cut off is 83 chars, so with a 82 char wide terminal it scrolls, at 
83 char it doesn't unless have -vvvvvvv

>
>>> One possibility is that \r doesn't do carriage return (move cursor back
>>> to beginning) but a line end instead. Do we need to do some custom
>>> cursor positioning?
>>
>> I'd assume it gets translated into CR/LF, perhaps needs to be in binary
>> mode?
>
> Hm. Maybe. But this maybe could mess up other parts. You can just try a
>
> 	compat_binmode(STDERR_FILENO, TRUE);
>
> early in main() in mgp123.c. We'd have to insert some \r then, though,
> for metadata printout. They won't hurt on other systems.
>
> But should textmode in OS/2 really translate \r to \r\n? That does not
> match my recall that says that it's only about what \n means internally
> and externally.
>

Binary mode makes no difference and you're right about \r.
Dave