[Toybox] [New Toys] - tcpsvd, udpsvd

[Rob Landley]

On 09/25/2013 03:57:49 AM, Ashwini Sharma wrote:
> Hi Rob, list,
> 
> Attached is the implementation of tcpsvd and udpsvd.
> Both the commands are handled in the same file and support __IPv4__  
> and
> __IPv6__.

Sorry I've been incommunicado for a few days. I was trying to tackle  
the giant heap of code needing cleanup in pending (half-finished nl.c  
cleanup that I haven't been able to get back to in weeks, how find is  
defining magic constants so two halves of the same program can talk to  
each other, the long-delayed ifconfig cleanup that's 80% done and just  
needs the ipv6 bits dealt with...)

And then this came in. I've had a todo item to add udp support to  
netcat for a while because I want to make netconsole work (and the USB3  
maintainer's netconsole tutorial has a netcat udp command line as the  
I/O mechanism on the host)...

And here's a new command, adding yet more to pending, not on the  
roadmap, reimplementing stuff that netcat already does but with a  
different UI and sharing zero code, with eight command line options to  
micromanage stuff like the localhost name not being the same as the  
machine's localhost name... I'm sure it's vital to somebody's internal  
use case that they have every single one of those command line options  
and they can't _possibly_ stop using busybox without it, even though  
the version of busybox I maintained didn't have it (recent addition),  
android hasn't got it, not needed to build linux from scratch, it's not  
present in the ubuntu base install... no, until I've copied every  
single piece of bloat busybox added including that weird set of init  
things Denys likes (not sysvinit, not upstart, not systemd, not android  
init, no a _fifth_ implementation. Sixth if you count oneit), then  
toybox can never be done.

And honestly, I haven't got a good reason to say no to it. I don't want  
to come _up_ with a good reason, I don't want to argue about it. I  
don't want to debate design philosophy with people who demonstrably  
have no idea what I would want to add to the project. This thing works,  
it's small, the easy way would be to merge it and just not argue. Same  
for the next 50 random commands that do who knows what.

And I just... stopped caring for a while. It's no fun anymore.

I haven't worked on "mount" or "sed" in ages because writing new code  
is fun and cleaning up other people's implementations (often at the  
design level) is work, and I should do work before I do fun. I can't  
really enjoy the fun stuff with work hanging over me. Especially if the  
pile of work is growing the longer I ignore it, so that if I don't  
buckle down I'll never see the end of it...

I should really just delete this message rather than posting it. (Job  
hunting always puts me in a bad mood.) But that just makes the  
incommunicado longer.

Rob