Re: [timexdatalinkusbdevelop] Re: wBuilder Version 0.6
- View SourceHi Tons,
Selon tonsofquestions <tonsofquestions@...>:
> I've been using your tool a bit now, a couple of questions/comments:I will add this in the next version.
> It might be useful to have a "total" line somewhere that displays the
> total size of the app, number of files, states, etc. all in one handy
>Why not. Personnaly I always have an editor (psPad) opened when I use wBuilder.
> It would be really cool (though admittedly perhaps a little difficult)
> to be able to click on a file in a state, and have it open in an
> editor. It would be harder if the state has multiple files (because
> then the box would have several names to distinguish between), but if
> it's just one file, I think it might not be too bad.
>I can add a check box to allow the code generation even if there is an oversized
> If one of the states is too large, an invalid *_code.bin is generated.
> Invalid as in the VDLK can't handle it. The original tool generated a
> working app, even if it would potentially cause problems, and
> overwrite needed memory. This was useful for testing, you could see
> that code was working properly (for the most part), and then optimize
> out extra bytes later. While this way is perhaps technically more
> correct, it might be better to not generate a *_code.bin at all,
> though I think generating the oversized one would be preferable. It
> would also be very important to be able to do this when Paulo gets his
> "larger effective code space" modification working properly.
> A question: at what point are the .ers/.elk files checked for errors?
> I ask this because I added a filesize check in a helper .bat file - ISend me your project in other to test it.
> use it as a sanity-check/warning for when I modify the browser in
> modebrowser. (We use unallocated memory at the end of the state as an
> extra buffer, but if we make the state too large from changes, we
> might not have enough room. So if the filesize changes I drop in a
> warning to make sure the developer double-checks that there's still
> the necessary memory left). What I do right now is move the .sre file
> away, and add in a .ers with a warning (all at the end of
> compilation). It worked with the Timex builder (the second step
> couldn't find the sre for the state and brought up it's usual error
> message), but with your tool no errors are reported (the text of the
> .ers file I create doesn't appear in the box), and no other messages
> appear, but the wristapp is not generated. Is there a better way I
> should be doing this? Perhaps something small could be changed to make
> this possible?
>Actually the load feature is a hack. It's far to be perfect. Actually the best
> Finally, one additional observation/request related to the VDLK
> (instead of wBuilder). On the watch, before the COMM mode is entered,
> all of the apps' background handlers get sent a "SHUTDOWN_FOR_COMM"
> event. Could the emulator send these as well, before entering it's
> COMM mode (and loading in new wristapps)? It would be very helpful
> when working with TSRs (like neolink), both in testing that it would
> work properly on a real watch, and to avoid the repeated crashes that
> ocurr without it after every sync (the SHUTDOWN event clears up some
> modified internal data, which causes problems if other wristapps - or
> the same on - are loaded into memory). For example, loading the
> sv_night app and then removing it works fine on the watch, but on the
> emulator it can never be removed without resetting (or loading a
thing to do when you want to add a new wristapp it's to first reset to watch and
then to load the new wristapp.
- View Source
> Why not. Personnaly I always have an editor (psPad) opened when Iuse wBuilder.
Yes, I have an editor open, too (I'm lazy and use notepad *hangs head
in shame*), but sometimes there's an error in a file I was working on
a while before, that I might have closed. Or maybe I'm looking at the
list of files and think of something cool to add. It's more of a
convenience thing, not a huge deal-breaker.
> I can add a check box to allow the code generation even if there isan oversized
> > A question: at what point are the .ers/.elk files checked for errors?Sorry, I guess I was a little unclear. I didn't have any doubts that
they did get checked, I wasn't sure when.
> Send me your project in other to test it.Actually, the one I'm working on is pretty big, I'll just make a tiny
one to demonstrate what I mean. I'll send it to you a bit later,
either today or tomorrow.
> Actually the load feature is a hack. It's far to be perfect.Actually the best
> thing to do when you want to add a new wristapp it's to first resetto watch and
> then to load the new wristapp.Hmm. Good to know. Unfortunately, I was hoping to be able to test the
unloading of the wristapp. It's actually important in Neolink because
of how I modify the TOD state handlers. I'm fairly certain it's
correct, but it would still be nice to test/debug on the emulator
instead of the watch.