joeb@hpmwtd.HP.COM (Joe Ballantyne) (03/17/89)
I am writing an application that creates some very large contiguous files using File Manager routines. I want to be sure that the files are always contiguous even when they are copied by the Finder from one disk to another. Will the Finder always copy a contiguous file as a contiguous file? Is it possible to force the Finder to keep a file a contiguous file? If so, how is this done? I had the brilliant? idea that if I set the file clump size of my contiguous file to the size of the file itself, then perhaps the finder would use the original file's clump size when allocating space on the disk for the copy. Is that true? (I'm beginning to have my doubts.) Unfortunately, I have found it impossible (so far) to even set the file clump size. I first tried using PBSetCatInfo. I set the file clump size feild to the value that I wanted, called PBSetCatInfo, recalled PBGetCatInfo, and the clump size had been conveniently set to zero for me. I then decided that Pascal wouldn't do it for me, so I tried machine language. I found the files file control block, set the clump size long word directly in the file control block, set the dirty bit (bit 15) of the ioflags word, set the file has its own clump size bit (bit 6) of the ioflags word, jumped back to my program, (everything seemingly ok) called flushvol, and naturally, the system crashed. Is it possible to set a file's clump size to something other than zero? If I didn't set bit 15 in my assembly language subroutine, the flushvol and subsequent file manager calls worked fine, BUT when I checked the file clump size with FEdit Plus, it was zero. Do I need to set bit 7 of the ioflags word? How does one set the clump size of a file in a permanant manner? Do you have to access the B* trees directly? Thanks in advance for your help Joe Ballantyne Hewlett Packard Microwave Technology Division Santa Rosa, CA 95405 joeb@hpmwjcb.HP.COM
resnick@cogsci.uiuc.edu (Pete Resnick) (10/23/90)
Two questions, one which I had asked long ago but have forgotten the answer to: 1. I am reading a file and sending it over a TCP connection. That means I have to convert newline's to CR/LF and not take any non-ASCII characters. Assuming that the file I am reading has been typed in TeachText or some other text editor, what will be the easiest way to read it in the right format? (I am assuming that autowraps don't count here and that they are not in the file I am reading. Correct me if this is mistaken). 2. This file reading is going to be done by a device driver during accRun time. Is it possible to leave the file open between accRun calls? In other words, can I open the file into the system heap so that it wont be closed when an application exits? Sorry if these are repeats of old questions. pr -- Pete Resnick (...so what is a mojo, and why would one be rising?) Graduate assistant - Philosophy Department, Gregory Hall, UIUC System manager - Cognitive Science Group, Beckman Institute, UIUC Internet/ARPAnet/EDUnet : resnick@cogsci.uiuc.edu BITNET (if no other way) : FREE0285@UIUCVMD
resnick@cogsci.uiuc.edu (Pete Resnick) (06/03/91)
Two questions about the File Manager: 1. What routines that are documented in IM VI (if any) are available in System 6.0.x? I am especially interested in whether PBHOpenDF will work under 6.0.x. (As an aside, what about Resource Manager routines like HOpenResFile?) 2. When a completion routine gets called for, say, PBRead, I know I get the paramBlock in A0. Do I get it on the stack by any chance also? If not, is there any way to get at it besides my C code having a "asm {move A0,myPB}" in the middle of my code? How do people do it in Pascal? And what should the declaration of the completion routine look like? I couldn't find any of this in either IM IV or IM VI. pr -- Pete Resnick (...so what is a mojo, and why would one be rising?) Graduate assistant - Philosophy Department, Gregory Hall, UIUC System manager - Cognitive Science Group, Beckman Institute, UIUC Internet/ARPAnet/EDUnet : resnick@cogsci.uiuc.edu BITNET (if no other way) : FREE0285@UIUCVMD
d88-jwa@byse.nada.kth.se (Jon W{tte) (06/03/91)
@ux1.cso.uiuc.edu> resnick@cogsci.uiuc.edu (Pete Resnick) writes:
1. What routines that are documented in IM VI (if any) are available in
System 6.0.x? I am especially interested in whether PBHOpenDF will work
under 6.0.x. (As an aside, what about Resource Manager routines like
HOpenResFile?)
These kinds of routines (especially HOpenResFile) are just glue
routines for the PBH calls. They are avaiailable in the development
environment (THINK C 4, for instance) and work well with system 6.
This is documented in technote 200-something (14 ?)
--
Jon W{tte
h+@nada.kth.se
- Speed !