NetNews Usenet Archive 1993 #1/NN_1993_1.iso/spool/comp/sys/amiga/programm/18331

Internet Message Format | 1993-01-08 | 3.1 KB

Path: sparky!uunet!gatech!usenet.ins.cwru.edu!agate!darkstar.UCSC.EDU!cats.ucsc.edu!davids
From: [e-mail] (Dave Schreiber)
Newsgroups: comp.sys.amiga.programmer
Subject: Re: Flushing Devices; C= doco (mutter!)
Date: 8 Jan 1993 07:06:27 GMT
Organization: University of California; Santa Cruz
Lines: 64
Message-ID: <[e-mail]>
References: <[e-mail]> <[e-mail]> <crystal.726454269@glia>
NNTP-Posting-Host: si.ucsc.edu


In article <crystal.726454269@glia> [e-mail] (Crystal) writes:
>In <[e-mail]> [e-mail] (Dave Schreiber) writes:
>
>
>>In article <[e-mail]> [e-mail] (Gary Duncan) writes:
>>>Passing thought ; why the heck aren't unused libraries/devices
>>>flushed automatically anyway? Sounds like a legacy of the days
>>>when hard disks were rare and dragging a device off floppy was
>>>time-consuming. Those days are gone...
>
>>Why should they be flushed?  An unused library essentially takes up
>>zero memory, since the memory it uses can be demanded from it at
>
>Then how come running "avail flush" returned 75k of memory to me if they
>take up ZERO memory...?

I said "_essentially_ takes up zero memory."  See below for the reason.

>
>>any time (and will be demanded from it before a memory allocation
>>is allowed to fail).  The only people for whom auto-removal of
>>libraries is important are developers who are checking for memory
>
>Warning: you are dealing in absolutes here... developers are not the ONLY
>people for whom auto-removal of libraries is important.  Those with
>little ram trying to run DTP programs or Database programs that access
>large textfiles or graphics likewise need to be able to have as much memory
>available to them as possible for their work, too.

Reread what I said above.  If AllocMem() cannot allocate a block of memory
of requested size, it will flush all unused libraries and devices from
memory and try again.  A memory allocation _cannot_ fail because needed
memory is being occupied by a shared library or device.

>
>>leaks; there are plenty of ways of flushing libraries on those
>>rare occassions when it is needed.
>
>Ok, so, other than "avail flush", what other ways are there?  I want to
>recover as much as I can from those programs that don't clean up after
>themselves.

The following will work since the AllocMem() will always fail and thus
will always try and flush unused libraries and devices from memory
(unless, of course, the user has a chunk of memory 4G-16K large).

APTR *foo;
foo=AllocMem((unsigned long)(-16384),0L);

Also, if you run Workbench with the -debug option, you get a menu
called "Debug" that contains an item called "flushlibs" that does
the same thing.

BTW, since the retaining of unused libraries in memory is an OS
service, I'd hardly call it a case of programs "not cleaning up
after themselves."

>>>Gary Duncan            [e-mail]
>>Dave Schreiber  "Look.  Don't touch."  [e-mail] (until 6/20/93)


--
Dave Schreiber  "Look.  Don't touch."  [e-mail] (until 6/20/93)