365 lines
18 KiB
Plaintext
365 lines
18 KiB
Plaintext
America Online
|
|
APPLE II DEVELOPMENT FORUM CONFERENCE LOG
|
|
Tuesday, November 6, 1990 10:00 p.m. Eastern Time
|
|
Topic: Apple IIGS System Software 5.0.3
|
|
Forum Leader: Dave Sugar (AFL Dyfet)
|
|
|
|
AFL Dyfet Good evening and welcome to the Apple II Development Forum. We
|
|
will be
|
|
AFL Dyfet discussing and trying to answer questions in regards to System
|
|
5.0.3 tonight,
|
|
AFL Dyfet and we will be using PROTOCOL starting now. This means if you
|
|
have a question,
|
|
AFL Dyfet enter a '?' on a line by itself, and if you have a comment on a
|
|
question, enter
|
|
AFL Dyfet a '!' on a line by itself. You will each be called in turn.
|
|
There are no
|
|
AFL Dyfet doubt a number of questions in regards to changes in 5.0.3, and if
|
|
we can give
|
|
AFL Dyfet Matt a minute to summerize the important points, I think we will
|
|
be able to
|
|
AFL Dyfet move more smoothly from there....
|
|
ShanoJ So, what's new and cool in the 5.0.3 toolbox?!
|
|
Matt DTS Would you care to be more specific, Jonah? There are about 18
|
|
pages of release notes.
|
|
DaniF What is so specal that no one can bleave?!?
|
|
ShanoJ Well, for 5.0 you could say "there's a new toolset called TextEdit
|
|
that allows
|
|
ShanoJ you to do word processors really easily, and there's a new
|
|
Resource Manager."
|
|
ShanoJ Guess I was looking for a 'major points' summary like that,
|
|
Matt....
|
|
DaniF where is the release notes? Can they be found here?
|
|
Matt DTS The release notes were not sent electronically to licensees, so
|
|
they're not going to be available
|
|
Matt DTS online at first.
|
|
Matt DTS Let me page through my set of 5.0.3 documents and come up with
|
|
some changes:
|
|
Matt DTS Many things have been moved from one disk to another to accomodate
|
|
changed sizes in things like
|
|
Matt DTS the SCSI Manager (which supports DMA SCSI; 5.0.2's did not) and
|
|
ACE (which has been rewritten).
|
|
Matt DTS Bugs outlined in Tech Notes as "a bug in 5.0.2" are mostly fixed,
|
|
and many of the TNs have already
|
|
Matt DTS been corrected. GS/OS TN #1 shows which files on the disk have
|
|
changed, with two corrections:
|
|
Matt DTS 1) The LaserWriter driver did change for 5.0.3; the TN says it
|
|
did not, and
|
|
Matt DTS 2) The TN indicates there is an HP DeskJet Plus driver on the
|
|
disk; there is not.
|
|
Matt DTS The Window Manager has a new TaskMaster bit to let you feed it an
|
|
event record instead of it always
|
|
Matt DTS calling GetNextEvent - kind of an application version of
|
|
TaskMasterDA.
|
|
Matt DTS Standard File is very much changed. It is much more bug-free
|
|
(almost totally), and has a new
|
|
Matt DTS interface component. The "disk" button is replaced by "Volumes",
|
|
which puts a list of online volumes
|
|
Matt DTS in the file list area. The <tab> key still goes to the next
|
|
volume.
|
|
Matt DTS ACE has been rewritten to improve performance - it's now two to
|
|
three times faster.
|
|
Matt DTS The Resource Manager has a few nasty bugs fixed, as does TextEdit
|
|
- both are much more solid now.
|
|
Matt DTS You can now tell QuickDraw not to use FastFont at QDStartUp
|
|
time.
|
|
Matt DTS The ImageWriter and ImageWriter LQ drivers are totally new, from
|
|
scratch.
|
|
Matt DTS The IW driver is about two to three times faster, uses better
|
|
dialog boxes and has better color.
|
|
Matt DTS The new LQ driver is a NATIVE LQ driver, using all 27 pins, and
|
|
the same good interface and colors of
|
|
Matt DTS the new IW driver. The port drivers were changed to go up to
|
|
19,200 baud to help the LQ driver (a
|
|
Matt DTS full page of LQ data is a lot of stuff to be sending out the
|
|
serial port!)
|
|
Matt DTS >64K unclaimed interrupts is not a fatal error - it gives a dialog
|
|
that says "see your dealer".
|
|
Matt DTS Unloading segments by address now works again.
|
|
Matt DTS The UniDisk driver now correctly detects disk-switched condition
|
|
after returning from P8.
|
|
Matt DTS A new parameter was added to "Flush" to allow you to only flush
|
|
data blocks and not system blocks.
|
|
AFL Marty Yo Jonah!
|
|
Matt DTS A new call was added to the Miscellaneous Tools, "ConvSeconds" to
|
|
convert various date formats back
|
|
Matt DTS and forth.
|
|
Matt DTS There are lots more bug fixes, etc., but them's the highlights.
|
|
(No, I will NOT repeat them.)
|
|
Matt DTS :) GA
|
|
AFL Dyfet GA DaniF...
|
|
DaniF Is the over all speed with a HD and the plain SCSI card be
|
|
increased?
|
|
Matt DTS Not significantly, but a little bit I think. This is not a major
|
|
release, otherwise it would be
|
|
Matt DTS numbered 6.0. It probably could qualify as 5.1, but that would
|
|
require reprinting a whole lot of
|
|
Matt DTS manuals and passing the cost along, and we chose not to do that.
|
|
:) <GA>
|
|
AFA Gary J :)
|
|
AFL Dyfet Okay, the floor is now open for questions...
|
|
AFL Dyfet Go ahead Coach...
|
|
Coach101 Were there any changes in the AppleTalk boot over network process?
|
|
(particularly speed)
|
|
Coach101 ga
|
|
Matt DTS No, not in speed. A few bugs were fixed, though. <GA>
|
|
Coach101 Thanks.... ga
|
|
ShanoJ Matt'll probably tell me to read one of the tech notes, but you
|
|
mentioned a
|
|
ShanoJ problem with unloading a segment by address. What exactly was
|
|
wrong here? I
|
|
ShanoJ used it in Twilight.... :(
|
|
AFA Gary J :)
|
|
AFL Marty Read a technote Jonah.
|
|
Matt DTS It didn't work with ExpressLoad. See IIgs TN #22 (Ha! You were
|
|
right!) for more details.
|
|
Matt DTS (Jonah, it just gets to me because we spend a lot of time creating
|
|
a lot of documentation that you
|
|
Matt DTS don't seem to spend a lot of time reading. I'd rather do new
|
|
stuff than repeat the stuff I've already
|
|
Matt DTS done, and I'm sure you would too.) <GA>
|
|
ShanoJ Okay, thanks! (okay, everone tell me to read the technotes. Get it
|
|
out of your
|
|
ShanoJ system... Good boy, Marty.... :)
|
|
ShanoJ I know you do, Matt. I hope I'm not ACTUALLY as bad about reading
|
|
stuff as I
|
|
ShanoJ appear... I _DO_ read most of it, even though it probably doen't
|
|
look that way.
|
|
AFL Marty (read -all- of it :)
|
|
ShanoJ Oh, wait I've got another one---> ? :)
|
|
AFL Dyfet GA, Jonah
|
|
ShanoJ Okay, after I call UnloadSeg on a code segment, is it okay for me
|
|
to call
|
|
ShanoJ DisposeHandle on the memory it used to be in, or am I going to
|
|
dispose part
|
|
ShanoJ of the system instead of the code?
|
|
Matt DTS Don't go disposing of that stuff.
|
|
Matt DTS The Loader keeps the handle around and will try to reallocate it
|
|
if necessary later. The Loader makes
|
|
Matt DTS it purge level three, so it will be disposed as soon as any memory
|
|
request can't be immediately
|
|
Matt DTS fulfilled. Just Unload it and go about your business. <GA>
|
|
ShanoJ I was having a problem that after calling UnloadSeg, when I tried
|
|
to go to P8
|
|
ShanoJ I would get the $0201 error dialog. I checked with NiftyList, and
|
|
the code was
|
|
ShanoJ getting loaded into bank $01, leading me to believe that this was
|
|
causing the
|
|
ShanoJ problem...
|
|
Matt DTS If you're trying to stick around during P8, that will cause the
|
|
problem. Just set the "no special
|
|
Matt DTS memory" bit in your OMF and you won't get loaded in bank 1.
|
|
<GA>
|
|
AFL Dyfet For the late comers, we are discussing System 5.0.3 and changes to
|
|
system
|
|
AFL Dyfet tools. System 5.0.3 will be available on Thur. for downloading
|
|
:)....
|
|
AFA Gary J (12:00:01 A.M., to be exact :)
|
|
ShanoJ Well, I wasn't trying to stick around. I was done with that code,
|
|
so I tried
|
|
ShanoJ to unload it, but it didn't seem to work (even though I wasn't
|
|
getting an
|
|
ShanoJ error on the UnloadSeg call....)
|
|
Matt DTS Maybe it wasn't working properly! That's fixed now. <GA>
|
|
ShanoJ Okay, thanks!
|
|
AFL Dyfet Okay, Coach, GA...
|
|
Coach101 Though DisposeHandle would be a disaster, a PurgeHandle would be
|
|
legal would it not? ga
|
|
Matt DTS This is true. <GA>
|
|
ShanoJ Okay, that's nice. Thanks!
|
|
Matt DTS I will point out again, though, that you don't *need* to purge or
|
|
dispose of anything. The system
|
|
Matt DTS will kick it out as soon as it needs to. Really!
|
|
Matt DTS <GA>
|
|
Coach101 Understood, but if one wanted to get an estimate of the amount of
|
|
memory that might be available on
|
|
Coach101 a memory request, then a PurgeHandle would be needed so that the
|
|
memory would be freed, right?
|
|
Matt DTS Even then, Coach, nothing like MaxBlock or RealFreeMem will give
|
|
you a reasonable estimate.
|
|
Matt DTS The best thing to do is just try to allocate what you want and try
|
|
something smaller if you don't get
|
|
Matt DTS it. Trying to pre-flight NewHandle calls will only cause your
|
|
users grief in the end.
|
|
AFL Dyfet Okay, do we have any other takers for questions about IIgs tools
|
|
and/or 5.0.3?
|
|
AFL Dyfet GA Scott...
|
|
AFL Scott When drawing text in 640 mode, does one have to set up a special
|
|
pen mask to draw text in color?....
|
|
AFL Scott All I need is Red, green, and blue text. GA
|
|
AFL Dyfet GA Dave
|
|
Matt DTS (saved by the bang)
|
|
Dave Lyons The pen mask is always irrelevant for text--the
|
|
Dave Lyons foreground and background colors are relevant, as is the text mode
|
|
(see SetForeColor, SetBackColor,
|
|
Dave Lyons SetTextMode, etc, all in QuickDraw). There is a special bit in
|
|
either
|
|
Dave Lyons Text Mode or Font Flags (I think it's Font Flags, still QD) that
|
|
makes 4 bits of color
|
|
Dave Lyons significant in 640 mode instead of just 2. ga
|
|
AFL Scott I've tried SetFore and Back as well as the FontFlags bit 1... All
|
|
I get is Sick green, pink, white and
|
|
AFL Scott Black. :(
|
|
AFL Scott It's a mess, I'll tell ya that much... It seems as thought the
|
|
only colors one can get are just the'
|
|
AFL Scott ones I've mentioned... On a white background. Seems like the
|
|
lower two bits of the color word are
|
|
AFL Scott the only bits used and that FontFlags has no effect that I can
|
|
see. :(
|
|
AFL Scott I know it can be done, though... I just don't know how to do it.
|
|
:(
|
|
Matt DTS Hold on, I'll try calling him and see where he is...
|
|
Matt DTS I can't find him. We'd better move on.
|
|
AFL Dyfet GA Mac...
|
|
Mac Hater how much faster are the new ace tools? In decompression only?
|
|
Matt DTS Two to three times, and I believe it's both ways. <GA>
|
|
AFA Gary J (Crud, I'll have to put a delay loop in my program now, so that
|
|
the startup
|
|
AFA Gary J pic will display long enough :(
|
|
Mac Hater make a bigger sound sample ;)
|
|
AFA Gary J :)
|
|
AFL Dyfet Go ahead, David...
|
|
FeinDavid Has there been any changes to the finder? Like icons that are
|
|
showing up on
|
|
FeinDavid the desktop being automatically placed where they belong when you
|
|
do a clean
|
|
FeinDavid up?
|
|
Matt DTS The Finder did not change from 5.0.2.
|
|
Matt DTS <GA>
|
|
FeinDavid Damn.
|
|
AFL Dyfet As Matt said :)...
|
|
FeinDavid Sorry, I missed that.
|
|
Coach101 What is David's problem, CleanUp works for me.
|
|
Mac Hater it should begin to change dramatically, since Andy Nicholas now
|
|
works on the
|
|
Mac Hater Finder team! ()()()()
|
|
AFA Gary J :)
|
|
AFL Dyfet ShrinkIcons :)
|
|
FeinDavid Is the system much faster? I saw a beta that was faster than
|
|
FeinDavid 5.0.2.
|
|
Matt DTS It's a little faster, but nothing to write home about.
|
|
FeinDavid Load your IIgs in under 5 seconds to desktop off of a 3.5!
|
|
AFL Dyfet I like the changes to standard file in 5.0.3...
|
|
AFL Dyfet Go ahead Coach...
|
|
Coach101 Any changes in the dispatching/handling of SIGNALs in 5.0.3?
|
|
Matt DTS No. What did you want to change?
|
|
Coach101 If BusyFlag is set, then a SIGNAL is deferred...
|
|
Coach101 But, when the BusyFlag is reset, the SIGNAL queue is not checked
|
|
to see if anything needs to be
|
|
Coach101 dispatched... Not nice... He who blocks, should queue when the
|
|
unblock takes place... ga
|
|
Matt DTS This didn't change, Coach. Sorry.
|
|
Coach101 Well, I could alwasy hope :)
|
|
Matt DTS Any other takers?
|
|
Coach101 How about the Machine bit in a CDev?
|
|
Matt DTS No changes to the Control Panel, ADU, etc. An Installer bug was
|
|
fixed, though.
|
|
AFA Gary J What is the version of P8 on 5.0.3?
|
|
Mac Hater 1.9
|
|
Matt DTS Version 1.9. You wanted it, you got it. :)
|
|
AFA Gary J Actually, 1.8 was fine for me :)
|
|
EdM14 Anyone know anything about the new sound tools?
|
|
EdM14 Oops, I mean where they are or how can we get them?
|
|
ShanoJ Which sound tools? The new MIDISynth ones?
|
|
EdM14 Yes!!!
|
|
ShanoJ Nope, don't know anything about that.... :) (check APDA,
|
|
though...)
|
|
Matt DTS MIDI Synth should be at APDA very soon now, I hear.
|
|
AFA Gary J That's good news.
|
|
EdM14 I'll believe it when I see it.........
|
|
ShanoJ Okay, I've got to ask this before everyone leaves....
|
|
ShanoJ Okay, when Twilight goes to blank the screen, it's possible that
|
|
it can't get
|
|
ShanoJ some memory it needs, and if so, it won't be able to blank the
|
|
screen. What
|
|
ShanoJ then?! I guess my question is, what would be the best thing to do
|
|
here? I was
|
|
ShanoJ thinking of flasking the menu bar if the desk manager was started,
|
|
and the
|
|
ShanoJ border if it wasn't.... Thoughts, ideas, comments, socks to tape
|
|
in mouth so
|
|
ShanoJ I'll shut up? :)
|
|
Matt DTS The best thing to do is not to blank the screen. Doing something
|
|
else weird won't help.
|
|
AFL Dyfet Why not pre-allocate the memory you need at startup/initial
|
|
running?
|
|
ShanoJ Well, I'm not going to blank the screen. I thought it might be
|
|
nice to have
|
|
ShanoJ some way of telling the user that, though....
|
|
Matt DTS You should display a message in the CDev that says you can't get
|
|
the memory. If the screen doesn't
|
|
Matt DTS blank they're most likely to go check the CDev to see if
|
|
something's wrong.
|
|
ShanoJ Hmm... Might try getting it at startup, but I'm not sure if people
|
|
would like
|
|
ShanoJ a 32k chunk permanently allocatd... :(
|
|
AFL Dyfet They're gonna lose it everytime you blank anyway...
|
|
ShanoJ Ah! Now THERE'S a good idea! Thanks, Matt!
|
|
AFL Dyfet GA Scott...
|
|
AFL Scott Why not make your blankers use a different palette than the
|
|
screen... then add what ever is on the
|
|
AFL Scott screen to so the Refresh desktop call can fix it up when the mouse
|
|
is moved?
|
|
ShanoJ Ya, but they might not like it PERMANENTLY gone.... Dunno....
|
|
A2GS Why not require a TWILIGHT user to have a minimum of 5 megs?
|
|
Mac Hater lol
|
|
ShanoJ I thought about that, Albert, but Jim and Scott didn't like it
|
|
very much... :)
|
|
A2GS Picky, aren't they? ;)
|
|
ShanoJ Well, I actually don't call RefreshDesktop. I just save the screen
|
|
and then
|
|
ShanoJ restore it myself. I found RefreshDesktop was messing a LOT of
|
|
things up... :(
|
|
AFL Dyfet If they dont wan't it gone, they could always shut down the
|
|
blanker, right?
|
|
AFL Scott Hey, this might just be the reason why Twilight causes problem on
|
|
some systems and not on others.
|
|
ShanoJ What's that, Scott?
|
|
ShanoJ Well, they'd have to disable Twilight and then reboot, as far as I
|
|
can tell...
|
|
AFL Scott Jonah, cant you add what is on the desktop to some region so the
|
|
RefreshDesktop call will redraw the
|
|
AFL Scott whole thing? I mean sorta like before HideMenu was
|
|
AFL Scott available?
|
|
ShanoJ Or I could add a shutdown button to the CDev that would make
|
|
Twilight release
|
|
ShanoJ all of it's memory, but then it would have to be gone for the rest
|
|
of the time
|
|
ShanoJ the system is active....
|
|
AFL Dyfet No, just a shutdown button that deactivated it until it is
|
|
restarted (with a
|
|
AFL Dyfet restart button) and release the 32K in the interum...
|
|
AFL Dyfet And grab a new 32K area when the user restarts...
|
|
AFL Scott I mean it's six or one half dozen or the other at this stage,
|
|
Jonah... sooner or later you're going
|
|
AFL Scott to have to trap an error in low memory situations.
|
|
ShanoJ My first version of Twilight DID call RefreshDesktop, with a
|
|
parameter of 0
|
|
ShanoJ to refresh the whole thing, but I was getting WEIRD problems,
|
|
because the
|
|
ShanoJ application wasn't expecting to have it's desktiop refreshed right
|
|
then....
|
|
ShanoJ Maybe with the modifications I've made I'll be able to use it now,
|
|
but I'd
|
|
ShanoJ still have to do something else when the DeskManager isn't
|
|
started....
|
|
AFL Scott Jonah, the point that I'm trying to make is that sooner or later
|
|
you're going to lose that memory.
|
|
A2GS I don't see what the big deal is...if the user doesn't have the
|
|
memory then
|
|
A2GS don't blank the screen...period.
|
|
AFL Scott So, you're gonna have to either kill the blanker, or "what A2GS
|
|
said"... :)
|
|
ShanoJ Ya, but my original problem was that if I don't have some nice way
|
|
to tell the
|
|
ShanoJ user that I can't blank the screen. Otherwise I'm going against
|
|
the HIG, and
|
|
ShanoJ Matt'll drop a Mac II on me :(
|
|
A2GS Ok. then cycle the border colors if the user doesn't have enough
|
|
memory, that
|
|
A2GS should get the users attention. :)
|
|
ShanoJ (so would the Mac II... :)
|
|
A2GS Matt....our Matt??? He wouldn't do that...nothing less then a Mac
|
|
Luggable :)
|
|
|