| AWARE [SYSTEMS] | Imaging expertise for the Delphi developer | |||||||
![]() |
TIFF and LibTiff Mailing List Archive | |||||||
LibTiff Mailing List
TIFF and LibTiff Mailing List Archive Contact
The TIFF Mailing List Homepage |
Thread1994.06.24 18:23 "Re: Think C Libtiff ???", by Dan Mccoy"David W. Leland" <dleland@maroon.tc.umn.edu> writes: >Would anyone out there happen to have already created a Think C version of >Libtiff for the Mac? > >Well,I'm sure it's been done, so what I'm really asking is if if anyone reading >this would be willing to let me get my greedy little hands on it, so I can be >lazy. ndr@tazboy.JPL.NASA.GOV (Niles Ritter) writes: >A lot depends on which version of libtiff, and which version of >Think/Symantec C. Are you using the new 3.3 ANSI-compliant version >of libtiff? Think 5.0? 6.0? Symantec C++ 7.0+? They all have to >be tweeked differently. > >For the sake of argument, I will assume you are on the bleeding edge >and using the latest and greatest of both. If you want lzw compression, >a few changes need to be made because of the large static arrays that >make it gag. ???? lzw? static arrays? We use the libtiff 3.3 tif_lzw.c unchanged, it works with Think 5,6,7,... tif_lzw.c hasn't had any static arrays since libtiff 3.0, even then there were two arrays of nine bytes each. There is a data structure malloc'ed by tif_lzw.c that is about 72k bytes, but Think can handle that just fine. > ... Also,the fax3 fax4 and g3states >stuff wont work at all, so leave out the fax compressions. You can >disable these either by editing "tiffconf.h" directly, or by >defining COMPRESSION_SUPPORT in the "Prefix" preferences. I agree here. We can just leave the fax stuff out because it would be dumb to put the output of a photorealistic renderer into a fax file. Others may not have that option. Personally, it mystifies me why otherwise sane people put up with a toy compiler like Think. >Basically, all you have to do to make independent libraries is to >split it up into two projects (so you wont have to worry about 32K >limitations). Then in the Think C Preferences, turn on the >"ANSI" button, and then activate the "Think C" checkbox, the >"4-byte int" checkbox (it might work with two-byte ints, but you >can't be too careful). I believe libtiff 3.3 does work with 16-bit int's, but I also think 16-bit int's are a lame idea. >You can optionally turn on the 68020 box for >speed, but if you turn on the 68881 box your code will no longer >be PowerPC compatible. I have heard that you can run the soft 68881 emulator (name escapes me at the moment) on the PowerMac. > ... >(I'm pretty sure that Sam has integrated the lzw changes into his >next release of libtiff, so there's nothing in these archives that >he would care about). > ... >The fixed version of the >LZW code Again, what's this about lzw changes. We use 3.3 tif_lzw.c unchanged with no problems. > ... and a little more "maclike" tiffinfo utility project is >provided as an example. Good luck! The mac-tiffinfo sounds like it might be a good addition to the library tools. I'm always having to tell the mac users to move their problem tiff files over to a unix box so I can look at them and tell them what's wrong. Dan McCoy Pixar mccoy@pixar.com |
|||||||