Hollywood forums

I found from Even Library following command:

CtrlCQuit(enable)

You could try to disable that CTRLC Quit. Maybe that would help. But do remember that you wont be able to quit using CTRL-C after that.

evil wrote:hello!
Isn't there any workaround in hollywood for that?
Why can It get and reckon these break-signals? Is it, because hollywood-execs are commodities? In this case, it may be possible to change the codes to not be commodities??

As a workaround you could try to avoid calling WaitEvent(), i.e. do the following as a main loop:
But this is very bad practice because it keeps polling for events! One should never do that. Better wait for an OS4 fix.

Hello
@Andreas:
Isn't it possible to Create an event for this Ctrl-C Signal?? Maybe in Future versions of Hollywood? So the coder can decide himself to permet these Breaks or not.

@Bugula: What the hell is that Even Library??? I do not know it at all.

Greets

Jörg

@Evil

It was a misspelling. It was actually Event Library.

And by that I mean the section under which i found the command in Hollywood manual on windows in case you wish to look at the manual explanation of the command.

I opened a thread on hyperions os4-forum and got some interesting answers:

1. I tested my problem on os4, os3(uae) and winxp. On winxp and uae it went well. BUT:
ThomasRapp told me, that disconnecting an internet-connection on winxp does not mean, that the tcp-stack is closed at all (Like with miami and roadshow).
And on uae I only tested my program by disconnecting the internet-connection from windows (so uae isn't really disconnected, too...)

That means, that the problem should be the same on original-68k and winxp, if I really shutdown the net...

2. Roadshow has to send these break-signals, because every software, that has opened (and not closed) the bsd-socket.library, must exit. Otherwise NetShutDown would fail. IBrowse and YAM close the bsd-socker.library after every connection.
That menas: Yam just opens the bsd-socket, if it is going to fetch or send mail. Afterwards it will close the lib immediatley.
IBrowse Opens the bsd-socket for every single connection. And it closes it afterwards, too.
That's why those programs to not have to quit, when NetShutDown was run.

OWB or Jabberwocky for example open the bsd-socket.library at programstart and close it at programend. So they have to exit, when netshutdown is run.

Back to my problem.
In Order to get a solution for this problem with Hollywood, it must be possible to Open/Close the bsd-socket.libry inside the Program.
a) Hollywood has to do this automatically or
b) There should be commands like OpenTCP and CloseTCP to do this manually.

@Andreas: Would this be possible in future versions of Hollywood?

Best regards

Jörg

@evil

Do you use dial-up connection?
Connected to a router I tried to netshutdown and it closed mui-owb and waiting the network to shutdown. Meanwhile the hollywood script is gone unresponsible even to ctrl-c.

This seems to be a bad issue, so it would be nice to solve it somehow.

@evil:
Ok, this explanation makes sense. The TCP stack really needs to find a way to get rid of its clients when it tries to kill bsdsocket.library. But I think it's not worth the effort to change Hollywood's behaviour to a more sophisticated handling of sockets because nowadays it can be expected that bsdsocket.library is always there. The thing with bsdsocket.library going away is a relic from the 68k days. I think OS4 should switch to a static bsdsocket.library that is always accessible as it is on other platforms as well.

@Andreas!
You are absolutely right. Nowadays almost everyone will have a 24/7 connection...
In my first post a thought, it was a bug in Hollywood.
But after asking in the hyperion-os4-forum, I now know, that this is an absolut normal behaviour of Hollywood, and no bug.
So leave things as they are, and everything is alright

Best regards!!
Jörg