kkim@uiucdcsb.cs.uiuc.edu (08/10/87)
>> I would like to intercept calls to an Operating System or Toolbox routine >> (specifically, GetNextEvent) and do some pre- and/or post-processing of my >> own. NGetTrapAddress and NSetTrapAddress seem useful to do it. > > Actually, there is a hook for GetNextEvent which may be used. There was an > article about it in MacTutor last year or so. I believe the global is > JGNEFilter or something like that. You might want to look up that issue. I checked all tables of contents of MacTutor and got the following list: vol/no year/month title author page 2/12 86/dec Trap Speed comments ??? 6 2/10 86/oct Reduce your time in the traps Mike Morton 21 1/10 85/sep C Glue routines for Filter Procs Van Kichline 53 1/6 85/may In line traps simulate OS calls ??? 53 1/5 85/apr In line traps create windows ??? 20 1/3 85/feb traps and so FORTH ??? 23 Is there anybody out there who can tell me which article talks about "jGNEFilter" ? Or is the article missing in the above list? Also, any information about jGNEFilter/NGetTrapAddress/NSetTrapAddress will be appreciated (I have already checked Inside Mac and Mac Tech Notes). Please e-mail to me. Thanks in advance. Kyongsok KIM Dept. of Computer Science; Univ. of Illinois at Urbana-Champaign arpanet/csnet: kkim@a.cs.uiuc.edu usenet/uucp : {seismo, pur-ee, ihnp4}!uiucdcs!kkim
tim@hoptoad.uucp (Tim Maroney) (08/11/87)
All such filters are not long for this world and should be avoided. I
believe that you never need to patch GetNextEvent or intercept it in any
case, since you can always do whatever extra processing is needed normally
inside your main event loop, and it is nearly trivial to install a filter
procedure in all your modal dialogs which will handle events as neccessary.
In TOPS Terminal, I have done just this, with the result that all background
type actions, such as receiving data on a terminal connection, take place
behind any modal dialogs that come up. Don't ever patch a trap unless you
absolutely have to! And *never* use a low-memory-global-based hook, because
it will stop working very soon if it hasn't already.
In fact (it should be kosher to say this now that MultiFinder is all over
the trades and being announced in Boston), all low-memory globals are going
away pretty soon. This should cause some amusement among those who remember
that an Apple employee viciously flamed me here last year for my suggestion
that Apple should decommission all the low-memory globals over the space of
a year. Don't worry, I don't expect an apology.
--
Tim Maroney, {ihnp4,sun,well,ptsfa,lll-crg,frog}!hoptoad!tim (uucp)
hoptoad!tim@lll-crg (arpa)kkim@uiucdcsb.cs.uiuc.edu (08/12/87)
> /* Written Aug 11, 1987 by tim@hoptoad.uucp in uiucdcsb:comp.sys.mac */ > you never need to patch GetNextEvent or intercept it in any case, since you > can always do whatever extra processing is needed normally inside your main > event loop... Don't ever patch a trap unless you absolutely have to! my main objective is to modify event queue elements within a DESK ACCESSORY so that an application can retrieve modified keydown events. somebody suggested me to use NGetTrapAddress/NSetTrapAddress. i looked up the Inside Mac but could not figure how to patch GetNextEvent using these two traps. then i posted a question on comp.sys.mac how to patch GetNextEvent and somebody suggested to use jGNEFilter saying that there was an article about it in MacTutor. that is my story. therefore, if i can modify keydown events within a desk acc without patching GetNextEvent, i will be happy to hear the method. if i have to patch, i would like to know how to patch it (probably using these two traps). thanks. Kyongsok KIM arpanet/csnet: kkim@a.cs.uiuc.edu usenet/uucp : {seismo, pur-ee, ihnp4}!uiucdcs!kkim
jwhitnel@csib.UUCP (08/14/87)
In article <164500043@uiucdcsb> kkim@uiucdcsb.cs.uiuc.edu writes: | therefore, if i can modify keydown events within a desk acc without patching |GetNextEvent, i will be happy to hear the method. if i have to patch, i |would like to know how to patch it (probably using these two traps). I'm writting this without reference to IM, so take it with a small grain of salt. This is in C, translation to Pascal or other lanaguages is left as an exercise for the reader. #define GNE_TRAP 0xa??? /* Trap number for GetNextEvent (from IM) */ typedef int (*funcptr)(); /* Pointer to function returning an int */ functptr oldGetNextEvent; /* initlization code */ oldGetNextEvent = NGetTrapAddr( GNE_TRAP ); NSetTrapAddr( myGetNextEvent ); /* This is the patch itself, parameters must match GetNextEvent */ pascal Boolean myGetNextEvent( event, mask ) EventRecord *event; int mask; { Boolean retCode; /* First call real GetNextEvent to get the event. CallPascalW calls ** indirectly a pascal function from C. Your glue may differ here */ retCode = CallPascalW( event, mask, oldGetNextEvent ); if ( retCode ) { /* We have an event, so preprocess it */ } return( retCode ); } /* termination code. Must be executed before application termination or ** bad things will happen, not the least of which is your mac will crash */ SetTrapAddr( GNE_TRAP, oldGetNextEvent ); That's it. The initilzation code is executed once when your DA initlizes, the termination code is executed once when your DA is closed and the patch code is what is executed when the application calls GetNextEvent. Caveat, this method works only on Mac+ or newer system, it will not work on 64K ROM systems (and no flames about that, please). | | thanks. | |Kyongsok KIM |arpanet/csnet: kkim@a.cs.uiuc.edu |usenet/uucp : {seismo, pur-ee, ihnp4}!uiucdcs!kkim Jerry Whitnell It's a damn poor mind that can only Communication Solutions, Inc. think of one way to spell a word. -- Andrew Jackson
glennr@cognos.uucp (Glenn D. Rasmussen) (08/17/87)
In article <164500042@uiucdcsb> kkim@uiucdcsb.cs.uiuc.edu writes: > >Is there anybody out there who can tell me which article talks about >"jGNEFilter" ? Or is the article missing in the above list? > jGNEFilter was described in the article that implemented a screen blanker. I can't remember which issue it was in, but I do know that it is in the Best of MacTutor Volume 1. -- Glenn D. Rasmussen 3755 Riverside Dr. Cognos Incorporated Ottawa, Ontario {allegra,decvax,ihnp4,linus,pyramid} (613) 738-1440 CANADA K1G 3N3 !utzoo!dciem!nrcaer!cognos!glennr