This is the mail archive of the cygwin mailing list for the Cygwin project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: file-transfer-over-soundcard

On May 21, 2015, at 1:03 AM, Spet <> wrote:
> When i execute from dos console:

Why arenât you using Cygwin Terminal instead?

The Cygwin Terminal has a huge number of features missing from the Windows console, some of which allow it to work better with Cygwin programs.  (UTF-8 support, for example.)

Meanwhile, the DOS console offers zero advantage over the Cygwin Terminal when it comes to running Cygwin executables.

> ./generate.exe -b 25 -r 48000 -o /etc/fstab

Youâre trying to send *Cygwinâs* /etc/fstab to another box?  Why?  Typically the contents of that file are so trivial that itâs faster to just transcribe any changes in it when setting up a new box.

Does "cat /etc/fstab" show something from the Cygwin Terminal?

Anyway, it works here.  I get a 25 MiB sound output file for a 531 byte /etc/fstab.

> generate.exe -b 25 -r 48000 -o "k:\TEMP\a.txt" c:\cygwin\etc\fstab

Are you sure you built it with the Cygwin C compiler, and not something else, like the MinGW or Visual Studio compilers?  Thatâs the only explanation I can think of for why the POSIX path (/etc/fstab) would fail, but a DOS path would succeed.

I have to ask: Why do you even want this software?  Itâs a neat hack, but Iâm having trouble imagining a situation where you have two computers that each have Internet access in order to fetch Cygwin and this file transfer software, but they have no way to mutually communicate other than via a highly inefficient [*] communication link.

Iâve run into real-world situations where one box doesnât have Internet access and no way to attach an external storage device, so youâre left with hacks like uudecode and serial cables, but that doesnât seem to be whatâs going on here.

[*] The Nyquist limit for 48 kHz sampling is 24,000 bps, but the article is recommending 25 bps, three orders of magnitude lower.  Even if you could approach the Nyquist limit with better encoding, youâre still going to be transferring slower than a good RS-232 serial link.
Problem reports:
Unsubscribe info:

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]