Re: Shell Access with C++ Recommendations
On 2008-11-16 21:10, Sam wrote:
Erik Wikstr??m writes:
On 2008-11-16 15:44, Sam wrote:
xander.grespesky@gmail.com writes:
Install Linux on a spare machine. Now you have your own "shell access".
That's a really dumb answer Sam, obviously i have a few boxes I'm
playing with locally. But looking at the caliber of your past post
responses I'm not the least surprised.
If you already have "a few boxes", then you have no need for a unix "shell
access" services in order to test "out some networking related code over the
internet". I manage to develop quite a bit of "networking related code" this
way, just fine. And I don't really need the other machines at all.
"Networking related code" can be built and tested on a single machine, just
fine. So, who's the real the dummy, around here?
While it is possible to simulate stuff like latency, low bandwidth, low
MTU, etc. using stuff like dummynet or such no amount of simulation can
replace running live in the intended environment. Believing anything
else just shows who the "real dummy" is.
It's been my experience that only the "real dummies" insist that the only
way to test something is to let it loose in the indended environment,
without any testing. Because, after all, no real testing can possibly come
close to the "real thing".
If you count yourself in the "to hell with testing, just deploy" group, then
good luck with deploying what you cobbled together, with no real testing.
Without real experts like you, I wouldn't be able to make a nice living.
If you had read my reply more carefully you would notice that I never
said that you should not test your code, in fact I think it is very
important that you carefully test your code. What I did say, however,
was that no matter how much testing you perform, it will never be the
same as running the application live in the production environment.
Of course, having said that, it is important to notice that the kind of
testing the OP wants to do is still just testing. But by running as
realistic tests as possible, in as realistic environments as possible,
the chance of failure in production is minimised.
--
Erik Wikstr??m