home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.os2.misc
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!sgiblab!sdd.hp.com!caen!batcomputer!lynx@msc.cornell.edu!tommy
- From: tommy@lynx.msc.cornell.edu.UUCP (Mike Thompson,332,54714,2734927)
- Subject: Re: OS/2 Crashproof? NOT!
- Message-ID: <1992Nov17.170913.23292@msc.cornell.edu>
- Originator: tommy@snoopy.msc.cornell.edu
- Sender: news@msc.cornell.edu
- Organization: Cornell-Materials-Science-Center
- References: <BUHR.92Nov16184637@ccu.umanitoba.ca>
- Date: Tue, 17 Nov 1992 17:09:13 GMT
- Lines: 70
-
- From article <BUHR.92Nov16184637@ccu.umanitoba.ca>, by buhr@umanitoba.ca (Kevin Andrew Buhr):
-
- >> Example 2:
- >> From programming, it is TRIVIAL to hang the OS. Write a PM
- >> application and before starting the StdWindow, create a named pipe
- >> and wait for another process to access the pipe. Hang city...
- >
- > Clearly, this is a bug. On the other hand, it is a rather obscure
- > bug. I think one can reasonably differentiate between a lack of
- > .....
- > This is a bug, not a major shortcoming. Bugs can usually be worked
- > around: start your "StdWindow" before creating your named pipe;
- > ...
- > Major shortcomings aren't nearly as easy to deal with.
-
- I would disagree strongly. This bug, as I now understand, is related
- to the single message thread in OS/2. In the long term, this will be
- a major limitation and weakness in making OS/2 a robust operating system
- (or perhaps I should say making WPS a robust environment).
-
- >> try this one after loading the CSD. SubClass the FrameWindow in a
- >> PM application to modify the RESIZE message. Pass minimize and
- >> maximize message unchanged to original routine. Resize works
- > ....
- > Report it to IBM, and it'll probably get fixed.
-
- Let say something about support. I paid the subscription and hourly
- charges to access IBM's OS/2 bulletin board (as a developer, I really
- wanted upto date information and knowledge that there might be real
- developers listening). However, I currently can't even report bugs
- to IBM through their bulletin board. The CSD broke THEIR program
- PMHPO used to access the board. I am currently waiting for version
- 1.1b before I can see if these problems are already APAR's. I would
- think that inside IBM one would at least test their own support codes
- before releasing an official CSD (at least a warning if they knew).
-
- > talk to IBM and Northgate to see if you can narrow down the problem.
-
- Northgate! - they don't want to hear of OS/2. Said I needed an updated
- BIOS and possibly motherboard (strange for a system less than 6 months
- old). Called AMI and lo/behold, we have the latest of both. Faith in
- Northgate is now ZERO, which is why I am looking for a new supplier.
-
- IBM technical support has been very supportive, but not necessarily
- helpful. There seems to be little communication between the different
- support centers. IBM has worked hard to localize the crashes from
- the TCP/IP NFS problem, and at least we know that the faults are
- occuring in file read/write requests (unlikely to be memory or other
- hardware problems). However, I know long before IBM tech support of
- CSD's for the TCP package. Even on the base OS, they recently shipped
- me an "updated" kernal and WPS to try by overnight mail. However, the
- same day I downloaded the golden CSD which was several revisions
- "newer". (Unfortunately, neither eliminated the problems).
-
- > In addition, realize that operating system bugs may be infuriating and
- > they may fly in the face of zero-tolerance programming principles, but
- > in the PC world, the relative stability of OS/2 2.0 (and *not* OS/2
- > 2.1 or OS/2 2.2!) is impressive to some of us.
-
- Finally, I will be positive for a moment. Despite all my grumblings,
- I am totally committed to OS/2. There is no way I am returning to DOS
- or Windows programming - either OS/2 makes it or I will be looking for
- a real-time UNIX/MOTIF environment. IBM did a fantastic job developing
- the code and environment. I just hope that rash of teething problems
- will not be enough to drive off all support developers and kill the
- system before it is a year old. As for NT, Microsoft lost all of my
- respect after fighting with them over the rash of useless FORTRAN tools
- (up to and including the present version). Even in C, I'm much more
- impressed by CSET/2 than by MSC 6.1 (no reason to buy 7.0 since it
- doesn't support OS/2).
-