trb@stag.UUCP ( Todd Burkey ) (07/05/88)
Has anyone had problems with programs 'losing the mouse pointer' on Mega 2 machines? We recently purchased 20 Mega-2 monochrome systems at ETA (as part of our summer Superquest project) and have seen this problem on 5 or 6 of the ST's. We primarily see it when a student is using Unix Windows (the Lachman version), has several terminal sessions going on the screen and then moves the mouse up into the pulldown area. The pulldowns will start reversing when they shouldn't (i.e. messing up the top line) and when the mouse is moved out of the pulldowns and the pulldowns are dropped, the mouse disappears. I wrote a little desk accessory to allow turning off of the mouse, but then you get these 'patches' on the screen wherever you use the mouse. The mouse is very useful in the Lachman version of UW, since it allows cutting and pasting of text between windows, dynamic scaling of windows, and scrolling up and down in tall thin windows. I would normally think it is just a bug in the UW software itself, but the problems seem 1) to be intermittent on a machine basis, and 2) don't appear at all on some machines. If anyone has any ideas, please drop me a line. Some related info: We have currently set up the students using the following software: 1) UW - allows multi-session use on the Silicon Graphics workstations. This lets the student open a window up for telnetting over to our ETA10 supercomputer, and then do local compiles, edits, rogue playing, etc in other windows. Logging to up to 7 files simultaneously is great. 2) Uniterm - Unfortunately, UW doesn't support kermit or xmodem, so Uniterm gets used heavily for file transfers between the ST and the SG. 3) Absoft Fortran - Great fortran compiler. The students do most of the preliminary compiles on the ST, since that way they get the fastest feedback, have the fastest editor, and can do their work at home or when the Unix system is 'busy' (i.e. hung, rebooting, or otherwise tied up). 4) Tempus - Very, very nice editor...could use a few more key options, but its speed and power make up for a lot. 5) Micro C shell - For those who hate Gem...also, this allowed the students to become familiar with the C shell before they arrived at ETA (although how they all seemed to know about rogue I'll never know). 6) Word Writer - Nice, inexpensive word processor. 7) Publisher - Nice, DTP package by Timeworks. I would like to thank Atari (Sig in particular) for helping out in this project. We invited Sig Hartmann out to ETA about 3 weeks ago and he was so impressed with the Superquest project that Atari ended up donating two DTP systems for the project (Mega-2 ST's with Atari Laser printers). I think this type of involvement (associating Atari ST's with supercomputers) could really help out Atari's credibility in the business community. After all, MOST of us know already that it makes a very powerful development machine. Some asides: If any of you are using Silicon Graphics workstations (CDC 910's), we have a version of the unix windows server that actually works. Drop me a line if you need it. This is a modification of the 4 file uw distribution, not the more recent 20+ file mac version of the server. It is kind of half way between BSD and SYSV (i.e. we are using INET sockets, but SYSV terminal routines, and some really strange pseudo tty stuff). BTW, Sig, if you read this, thanks for inviting me to visit out there. I had a blast talking to everyone. I've tried EMAIL'ing you, but site 'sun' seems to be inaccessible (assume that is your sites' feed). And during the day, I've been run ragged helping the Superquest students out...very, very sharp kids, so playing phone tag wasn't practical. Within one week of arriving, some of the students were generating output on the supercomputers and displaying graphic displays on the Silicon Graphics workstations. -Todd Burkey "A member of STdNET-The ST developers' Network" trb@stag.UUCP -- To join, mail to ftg!dave@stag.UUCP --
federico@actisb.UUCP (Federico Heinz) (07/08/88)
In article <513@stag.UUCP> trb@stag.UUCP ( Todd Burkey ) writes: > >Has anyone had problems with programs 'losing the mouse pointer' on Mega 2 >machines? [detailed description of problem deleted] I've seen this problem on my Mega, but I thought it was a bug in the application, not in the ROM. On my machine, it happens when I accidentally drop a menu while the screen is being updated. My assumption that it was the application's fault was based on the fact that menus shouldn't drop while you update the screen, since one should "brace" the update code with a call that tells AES that you are changing the screen contents (I don't remember the function's name). I had this bug happen with some applications, most notoriously with Uniterm and DLII, both by Simon Poole, both written in what's-it's-name Pascal, so I further suspected a bug in this Pascal's GEM support. I couldn't confirm it, since I haven't it, but I reported the problem to Simon. He never answered, though, so I don't know whether he really got my mail. -- Federico Heinz "In Dubio Pro Libido" BIX: fheinz | Beusselstr. 21 UUCP: ...!unido!tub!actisb!federico | 1000 Berlin 21 Tel: (030) 396 77 92 | F.R. Germany.
kbad@atari.UUCP (07/11/88)
in article <217@actisb.UUCP>, federico@actisb.UUCP (Federico Heinz) says: > > In article <513@stag.UUCP> trb@stag.UUCP ( Todd Burkey ) writes: >> >>Has anyone had problems with programs 'losing the mouse pointer' on Mega 2 >>machines? [detailed description of problem deleted] > > [some stuff deleted] > My assumption that it was the application's fault was based on the fact > that menus shouldn't drop while you update the screen, since one should > "brace" the update code with a call that tells AES that you are changing > the screen contents (I don't remember the function's name). int wind_update(wi_ubegend) int wi_ubegend; The flag (wi_ubegend) can take on three values: 0 => end update 2 => end mouse control 1 => begin update 3 => begin mouse control Your analysis of the bug Todd described sounds quite reasonable. > > I had this bug happen with some applications, most notoriously with > Uniterm and DLII, both by Simon Poole, both written in what's-it's-name > Pascal, so I further suspected a bug in this Pascal's GEM support. CCD Pascal. I've noticed the same thing in UniTerm, running both on a Mega 4 and a 1040 ST. I have also seen this phenomenon in other PD software, so it may not be the fault of the Pascal GEM support. The moral of the story is: GEM software authors, use wind_update! > Federico Heinz "In Dubio Pro Libido" -- Ken Badertscher | Hey, umm, the stuff I said up there Atari R&D Software Test/Support | is, like, what _I_ think, okay? {portal,ames,imagen}!atari!kbad | So, y'know, don't bug Atari about it.
ron@argus.UUCP (Ron DeBlock) (07/16/88)
In article <217@actisb.UUCP>, federico@actisb.UUCP (Federico Heinz) writes: " In article <513@stag.UUCP> trb@stag.UUCP ( Todd Burkey ) writes: " > " >Has anyone had problems with programs 'losing the mouse pointer' on Mega 2 " >machines? [detailed description of problem deleted] " " I've seen this problem on my Mega, but I thought it was a bug in the " application, not in the ROM. On my machine, it happens when I " accidentally drop a menu while the screen is being updated. My " assumption that it was the application's fault was based on the fact " that menus shouldn't drop while you update the screen, since one should " "brace" the update code with a call that tells AES that you are changing " the screen contents (I don't remember the function's name). " " " -- " Federico Heinz "In Dubio Pro Libido" The problem does not only exist on the Mega. My 1040 has also exhibited the disappearing pointer syndrome while running Uniterm, WordPerfect, VIP Professional and other programs. As Federico stated, it happens if I drop a menu while the screen is updating. Moving off of the menu and clicking once will bring the pointer out of hiding. I am not sure whether the fault lies in the application or the AES. I haven't yet tried to lose the pointer on purpose, but it seens to have happened with every program that does extensive screen updating. -- Ron DeBlock KA2IKT ...!rutgers!galaxy!argus ron@wilbur.njit.edu "Oooo...that's mean!" "No, just average."