1INSTALLATION INSTRUCTIONS for the Independent JPEG Group's JPEG software 2 3Copyright (C) 1991-2010, Thomas G. Lane, Guido Vollbeding. 4This file is part of the Independent JPEG Group's software. 5For conditions of distribution and use, see the accompanying README file. 6 7 8This file explains how to configure and install the IJG software. We have 9tried to make this software extremely portable and flexible, so that it can be 10adapted to almost any environment. The downside of this decision is that the 11installation process is complicated. We have provided shortcuts to simplify 12the task on common systems. But in any case, you will need at least a little 13familiarity with C programming and program build procedures for your system. 14 15If you are only using this software as part of a larger program, the larger 16program's installation procedure may take care of configuring the IJG code. 17For example, Ghostscript's installation script will configure the IJG code. 18You don't need to read this file if you just want to compile Ghostscript. 19 20If you are on a Unix machine, you may not need to read this file at all. 21Try doing 22 ./configure 23 make 24 make test 25If that doesn't complain, do 26 make install 27(better do "make -n install" first to see if the makefile will put the files 28where you want them). Read further if you run into snags or want to customize 29the code for your system. 30 31 32TABLE OF CONTENTS 33----------------- 34 35Before you start 36Configuring the software: 37 using the automatic "configure" script 38 using one of the supplied jconfig and makefile files 39 by hand 40Building the software 41Testing the software 42Installing the software 43Optional stuff 44Optimization 45Hints for specific systems 46 47 48BEFORE YOU START 49================ 50 51Before installing the software you must unpack the distributed source code. 52Since you are reading this file, you have probably already succeeded in this 53task. However, there is a potential for error if you needed to convert the 54files to the local standard text file format (for example, if you are on 55MS-DOS you may have converted LF end-of-line to CR/LF). You must apply 56such conversion to all the files EXCEPT those whose names begin with "test". 57The test files contain binary data; if you change them in any way then the 58self-test will give bad results. 59 60Please check the last section of this file to see if there are hints for the 61specific machine or compiler you are using. 62 63 64CONFIGURING THE SOFTWARE 65======================== 66 67To configure the IJG code for your system, you need to create two files: 68 * jconfig.h: contains values for system-dependent #define symbols. 69 * Makefile: controls the compilation process. 70(On a non-Unix machine, you may create "project files" or some other 71substitute for a Makefile. jconfig.h is needed in any environment.) 72 73We provide three different ways to generate these files: 74 * On a Unix system, you can just run the "configure" script. 75 * We provide sample jconfig files and makefiles for popular machines; 76 if your machine matches one of the samples, just copy the right sample 77 files to jconfig.h and Makefile. 78 * If all else fails, read the instructions below and make your own files. 79 80 81Configuring the software using the automatic "configure" script 82--------------------------------------------------------------- 83 84If you are on a Unix machine, you can just type 85 ./configure 86and let the configure script construct appropriate configuration files. 87If you're using "csh" on an old version of System V, you might need to type 88 sh configure 89instead to prevent csh from trying to execute configure itself. 90Expect configure to run for a few minutes, particularly on slower machines; 91it works by compiling a series of test programs. 92 93Configure was created with GNU Autoconf and it follows the usual conventions 94for GNU configure scripts. It makes a few assumptions that you may want to 95override. You can do this by providing optional switches to configure: 96 97* Configure will build both static and shared libraries, if possible. 98If you want to build libjpeg only as a static library, say 99 ./configure --disable-shared 100If you want to build libjpeg only as a shared library, say 101 ./configure --disable-static 102Configure uses GNU libtool to take care of system-dependent shared library 103building methods. 104 105* Configure will use gcc (GNU C compiler) if it's available, otherwise cc. 106To force a particular compiler to be selected, use the CC option, for example 107 ./configure CC='cc' 108The same method can be used to include any unusual compiler switches. 109For example, on HP-UX you probably want to say 110 ./configure CC='cc -Aa' 111to get HP's compiler to run in ANSI mode. 112 113* The default CFLAGS setting is "-g" for non-gcc compilers, "-g -O2" for gcc. 114You can override this by saying, for example, 115 ./configure CFLAGS='-O2' 116if you want to compile without debugging support. 117 118* Configure will set up the makefile so that "make install" will install files 119into /usr/local/bin, /usr/local/man, etc. You can specify an installation 120prefix other than "/usr/local" by giving configure the option "--prefix=PATH". 121 122* If you don't have a lot of swap space, you may need to enable the IJG 123software's internal virtual memory mechanism. To do this, give the option 124"--enable-maxmem=N" where N is the default maxmemory limit in megabytes. 125This is discussed in more detail under "Selecting a memory manager", below. 126You probably don't need to worry about this on reasonably-sized Unix machines, 127unless you plan to process very large images. 128 129Configure has some other features that are useful if you are cross-compiling 130or working in a network of multiple machine types; but if you need those 131features, you probably already know how to use them. 132 133 134Configuring the software using one of the supplied jconfig and makefile files 135----------------------------------------------------------------------------- 136 137If you have one of these systems, you can just use the provided configuration 138files: 139 140Makefile jconfig file System and/or compiler 141 142makefile.manx jconfig.manx Amiga, Manx Aztec C 143makefile.sas jconfig.sas Amiga, SAS C 144makeproj.mac jconfig.mac Apple Macintosh, Metrowerks CodeWarrior 145mak*jpeg.st jconfig.st Atari ST/STE/TT, Pure C or Turbo C 146makefile.bcc jconfig.bcc MS-DOS or OS/2, Borland C 147makefile.dj jconfig.dj MS-DOS, DJGPP (Delorie's port of GNU C) 148makefile.mc6 jconfig.mc6 MS-DOS, Microsoft C (16-bit only) 149makefile.wat jconfig.wat MS-DOS, OS/2, or Windows NT, Watcom C 150makefile.vc jconfig.vc Windows NT/95, MS Visual C++ 151make*.vc6 jconfig.vc Windows NT/95, MS Visual C++ 6 152make*.v10 jconfig.vc Windows NT/95, MS Visual C++ 2010 (v10) 153makefile.mms jconfig.vms Digital VMS, with MMS software 154makefile.vms jconfig.vms Digital VMS, without MMS software 155 156Copy the proper jconfig file to jconfig.h and the makefile to Makefile (or 157whatever your system uses as the standard makefile name). For more info see 158the appropriate system-specific hints section near the end of this file. 159 160 161Configuring the software by hand 162-------------------------------- 163 164First, generate a jconfig.h file. If you are moderately familiar with C, 165the comments in jconfig.txt should be enough information to do this; just 166copy jconfig.txt to jconfig.h and edit it appropriately. Otherwise, you may 167prefer to use the ckconfig.c program. You will need to compile and execute 168ckconfig.c by hand --- we hope you know at least enough to do that. 169ckconfig.c may not compile the first try (in fact, the whole idea is for it 170to fail if anything is going to). If you get compile errors, fix them by 171editing ckconfig.c according to the directions given in ckconfig.c. Once 172you get it to run, it will write a suitable jconfig.h file, and will also 173print out some advice about which makefile to use. 174 175You may also want to look at the canned jconfig files, if there is one for a 176system similar to yours. 177 178Second, select a makefile and copy it to Makefile (or whatever your system 179uses as the standard makefile name). The most generic makefiles we provide 180are 181 makefile.ansi: if your C compiler supports function prototypes 182 makefile.unix: if not. 183(You have function prototypes if ckconfig.c put "#define HAVE_PROTOTYPES" 184in jconfig.h.) You may want to start from one of the other makefiles if 185there is one for a system similar to yours. 186 187Look over the selected Makefile and adjust options as needed. In particular 188you may want to change the CC and CFLAGS definitions. For instance, if you 189are using GCC, set CC=gcc. If you had to use any compiler switches to get 190ckconfig.c to work, make sure the same switches are in CFLAGS. 191 192If you are on a system that doesn't use makefiles, you'll need to set up 193project files (or whatever you do use) to compile all the source files and 194link them into executable files cjpeg, djpeg, jpegtran, rdjpgcom, and wrjpgcom. 195See the file lists in any of the makefiles to find out which files go into 196each program. Note that the provided makefiles all make a "library" file 197libjpeg first, but you don't have to do that if you don't want to; the file 198lists identify which source files are actually needed for compression, 199decompression, or both. As a last resort, you can make a batch script that 200just compiles everything and links it all together; makefile.vms is an example 201of this (it's for VMS systems that have no make-like utility). 202 203Here are comments about some specific configuration decisions you'll 204need to make: 205 206Command line style 207------------------ 208 209These programs can use a Unix-like command line style which supports 210redirection and piping, like this: 211 cjpeg inputfile >outputfile 212 cjpeg <inputfile >outputfile 213 source program | cjpeg >outputfile 214The simpler "two file" command line style is just 215 cjpeg inputfile outputfile 216You may prefer the two-file style, particularly if you don't have pipes. 217 218You MUST use two-file style on any system that doesn't cope well with binary 219data fed through stdin/stdout; this is true for some MS-DOS compilers, for 220example. If you're not on a Unix system, it's safest to assume you need 221two-file style. (But if your compiler provides either the Posix-standard 222fdopen() library routine or a Microsoft-compatible setmode() routine, you 223can safely use the Unix command line style, by defining USE_FDOPEN or 224USE_SETMODE respectively.) 225 226To use the two-file style, make jconfig.h say "#define TWO_FILE_COMMANDLINE". 227 228Selecting a memory manager 229-------------------------- 230 231The IJG code is capable of working on images that are too big to fit in main 232memory; data is swapped out to temporary files as necessary. However, the 233code to do this is rather system-dependent. We provide five different 234memory managers: 235 236* jmemansi.c This version uses the ANSI-standard library routine tmpfile(), 237 which not all non-ANSI systems have. On some systems 238 tmpfile() may put the temporary file in a non-optimal 239 location; if you don't like what it does, use jmemname.c. 240 241* jmemname.c This version creates named temporary files. For anything 242 except a Unix machine, you'll need to configure the 243 select_file_name() routine appropriately; see the comments 244 near the head of jmemname.c. If you use this version, define 245 NEED_SIGNAL_CATCHER in jconfig.h to make sure the temp files 246 are removed if the program is aborted. 247 248* jmemnobs.c (That stands for No Backing Store :-).) This will compile on 249 almost any system, but it assumes you have enough main memory 250 or virtual memory to hold the biggest images you work with. 251 252* jmemdos.c This should be used with most 16-bit MS-DOS compilers. 253 See the system-specific notes about MS-DOS for more info. 254 IMPORTANT: if you use this, define USE_MSDOS_MEMMGR in 255 jconfig.h, and include the assembly file jmemdosa.asm in the 256 programs. The supplied makefiles and jconfig files for 257 16-bit MS-DOS compilers already do both. 258 259* jmemmac.c Custom version for Apple Macintosh; see the system-specific 260 notes for Macintosh for more info. 261 262To use a particular memory manager, change the SYSDEPMEM variable in your 263makefile to equal the corresponding object file name (for example, jmemansi.o 264or jmemansi.obj for jmemansi.c). 265 266If you have plenty of (real or virtual) main memory, just use jmemnobs.c. 267"Plenty" means about ten bytes for every pixel in the largest images 268you plan to process, so a lot of systems don't meet this criterion. 269If yours doesn't, try jmemansi.c first. If that doesn't compile, you'll have 270to use jmemname.c; be sure to adjust select_file_name() for local conditions. 271You may also need to change unlink() to remove() in close_backing_store(). 272 273Except with jmemnobs.c or jmemmac.c, you need to adjust the DEFAULT_MAX_MEM 274setting to a reasonable value for your system (either by adding a #define for 275DEFAULT_MAX_MEM to jconfig.h, or by adding a -D switch to the Makefile). 276This value limits the amount of data space the program will attempt to 277allocate. Code and static data space isn't counted, so the actual memory 278needs for cjpeg or djpeg are typically 100 to 150Kb more than the max-memory 279setting. Larger max-memory settings reduce the amount of I/O needed to 280process a large image, but too large a value can result in "insufficient 281memory" failures. On most Unix machines (and other systems with virtual 282memory), just set DEFAULT_MAX_MEM to several million and forget it. At the 283other end of the spectrum, for MS-DOS machines you probably can't go much 284above 300K to 400K. (On MS-DOS the value refers to conventional memory only. 285Extended/expanded memory is handled separately by jmemdos.c.) 286 287 288BUILDING THE SOFTWARE 289===================== 290 291Now you should be able to compile the software. Just say "make" (or 292whatever's necessary to start the compilation). Have a cup of coffee. 293 294Here are some things that could go wrong: 295 296If your compiler complains about undefined structures, you should be able to 297shut it up by putting "#define INCOMPLETE_TYPES_BROKEN" in jconfig.h. 298 299If you have trouble with missing system include files or inclusion of the 300wrong ones, read jinclude.h. This shouldn't happen if you used configure 301or ckconfig.c to set up jconfig.h. 302 303There are a fair number of routines that do not use all of their parameters; 304some compilers will issue warnings about this, which you can ignore. There 305are also a few configuration checks that may give "unreachable code" warnings. 306Any other warning deserves investigation. 307 308If you don't have a getenv() library routine, define NO_GETENV. 309 310Also see the system-specific hints, below. 311 312 313TESTING THE SOFTWARE 314==================== 315 316As a quick test of functionality we've included a small sample image in 317several forms: 318 testorig.jpg Starting point for the djpeg tests. 319 testimg.ppm The output of djpeg testorig.jpg 320 testimg.bmp The output of djpeg -bmp -colors 256 testorig.jpg 321 testimg.jpg The output of cjpeg testimg.ppm 322 testprog.jpg Progressive-mode equivalent of testorig.jpg. 323 testimgp.jpg The output of cjpeg -progressive -optimize testimg.ppm 324(The first- and second-generation .jpg files aren't identical since JPEG is 325lossy.) If you can generate duplicates of the testimg* files then you 326probably have working programs. 327 328With most of the makefiles, "make test" will perform the necessary 329comparisons. 330 331If you're using a makefile that doesn't provide the test option, run djpeg 332and cjpeg by hand and compare the output files to testimg* with whatever 333binary file comparison tool you have. The files should be bit-for-bit 334identical. 335 336If the programs complain "MAX_ALLOC_CHUNK is wrong, please fix", then you 337need to reduce MAX_ALLOC_CHUNK to a value that fits in type size_t. 338Try adding "#define MAX_ALLOC_CHUNK 65520L" to jconfig.h. A less likely 339configuration error is "ALIGN_TYPE is wrong, please fix": defining ALIGN_TYPE 340as long should take care of that one. 341 342If the cjpeg test run fails with "Missing Huffman code table entry", it's a 343good bet that you needed to define RIGHT_SHIFT_IS_UNSIGNED. Go back to the 344configuration step and run ckconfig.c. (This is a good plan for any other 345test failure, too.) 346 347If you are using Unix (one-file) command line style on a non-Unix system, 348it's a good idea to check that binary I/O through stdin/stdout actually 349works. You should get the same results from "djpeg <testorig.jpg >out.ppm" 350as from "djpeg -outfile out.ppm testorig.jpg". Note that the makefiles all 351use the latter style and therefore do not exercise stdin/stdout! If this 352check fails, try recompiling with USE_SETMODE or USE_FDOPEN defined. 353If it still doesn't work, better use two-file style. 354 355If you chose a memory manager other than jmemnobs.c, you should test that 356temporary-file usage works. Try "djpeg -bmp -colors 256 -max 0 testorig.jpg" 357and make sure its output matches testimg.bmp. If you have any really large 358images handy, try compressing them with -optimize and/or decompressing with 359-colors 256 to make sure your DEFAULT_MAX_MEM setting is not too large. 360 361NOTE: this is far from an exhaustive test of the JPEG software; some modules, 362such as 1-pass color quantization, are not exercised at all. It's just a 363quick test to give you some confidence that you haven't missed something 364major. 365 366 367INSTALLING THE SOFTWARE 368======================= 369 370Once you're done with the above steps, you can install the software by 371copying the executable files (cjpeg, djpeg, jpegtran, rdjpgcom, and wrjpgcom) 372to wherever you normally install programs. On Unix systems, you'll also want 373to put the man pages (cjpeg.1, djpeg.1, jpegtran.1, rdjpgcom.1, wrjpgcom.1) 374in the man-page directory. The pre-fab makefiles don't support this step 375since there's such a wide variety of installation procedures on different 376systems. 377 378If you generated a Makefile with the "configure" script, you can just say 379 make install 380to install the programs and their man pages into the standard places. 381(You'll probably need to be root to do this.) We recommend first saying 382 make -n install 383to see where configure thought the files should go. You may need to edit 384the Makefile, particularly if your system's conventions for man page 385filenames don't match what configure expects. 386 387If you want to install the IJG library itself, for use in compiling other 388programs besides ours, then you need to put the four include files 389 jpeglib.h jerror.h jconfig.h jmorecfg.h 390into your include-file directory, and put the library file libjpeg.a 391(extension may vary depending on system) wherever library files go. 392If you generated a Makefile with "configure", it will do what it thinks 393is the right thing if you say 394 make install-lib 395 396 397OPTIONAL STUFF 398============== 399 400Progress monitor: 401 402If you like, you can #define PROGRESS_REPORT (in jconfig.h) to enable display 403of percent-done progress reports. The routine provided in cdjpeg.c merely 404prints percentages to stderr, but you can customize it to do something 405fancier. 406 407Utah RLE file format support: 408 409We distribute the software with support for RLE image files (Utah Raster 410Toolkit format) disabled, because the RLE support won't compile without the 411Utah library. If you have URT version 3.1 or later, you can enable RLE 412support as follows: 413 1. #define RLE_SUPPORTED in jconfig.h. 414 2. Add a -I option to CFLAGS in the Makefile for the directory 415 containing the URT .h files (typically the "include" 416 subdirectory of the URT distribution). 417 3. Add -L... -lrle to LDLIBS in the Makefile, where ... specifies 418 the directory containing the URT "librle.a" file (typically the 419 "lib" subdirectory of the URT distribution). 420 421Support for 12-bit-deep pixel data: 422 423The JPEG standard allows either 8-bit or 12-bit data precision. (For color, 424this means 8 or 12 bits per channel, of course.) If you need to work with 425deeper than 8-bit data, you can compile the IJG code for 12-bit operation. 426To do so: 427 1. In jmorecfg.h, define BITS_IN_JSAMPLE as 12 rather than 8. 428 2. In jconfig.h, undefine BMP_SUPPORTED, RLE_SUPPORTED, and TARGA_SUPPORTED, 429 because the code for those formats doesn't handle 12-bit data and won't 430 even compile. (The PPM code does work, as explained below. The GIF 431 code works too; it scales 8-bit GIF data to and from 12-bit depth 432 automatically.) 433 3. Compile. Don't expect "make test" to pass, since the supplied test 434 files are for 8-bit data. 435 436Currently, 12-bit support does not work on 16-bit-int machines. 437 438Note that a 12-bit version will not read 8-bit JPEG files, nor vice versa; 439so you'll want to keep around a regular 8-bit compilation as well. 440(Run-time selection of data depth, to allow a single copy that does both, 441is possible but would probably slow things down considerably; it's very low 442on our to-do list.) 443 444The PPM reader (rdppm.c) can read 12-bit data from either text-format or 445binary-format PPM and PGM files. Binary-format PPM/PGM files which have a 446maxval greater than 255 are assumed to use 2 bytes per sample, MSB first 447(big-endian order). As of early 1995, 2-byte binary format is not 448officially supported by the PBMPLUS library, but it is expected that a 449future release of PBMPLUS will support it. Note that the PPM reader will 450read files of any maxval regardless of the BITS_IN_JSAMPLE setting; incoming 451data is automatically rescaled to either maxval=255 or maxval=4095 as 452appropriate for the cjpeg bit depth. 453 454The PPM writer (wrppm.c) will normally write 2-byte binary PPM or PGM 455format, maxval 4095, when compiled with BITS_IN_JSAMPLE=12. Since this 456format is not yet widely supported, you can disable it by compiling wrppm.c 457with PPM_NORAWWORD defined; then the data is scaled down to 8 bits to make a 458standard 1-byte/sample PPM or PGM file. (Yes, this means still another copy 459of djpeg to keep around. But hopefully you won't need it for very long. 460Poskanzer's supposed to get that new PBMPLUS release out Real Soon Now.) 461 462Of course, if you are working with 12-bit data, you probably have it stored 463in some other, nonstandard format. In that case you'll probably want to 464write your own I/O modules to read and write your format. 465 466Note that a 12-bit version of cjpeg always runs in "-optimize" mode, in 467order to generate valid Huffman tables. This is necessary because our 468default Huffman tables only cover 8-bit data. 469 470Removing code: 471 472If you need to make a smaller version of the JPEG software, some optional 473functions can be removed at compile time. See the xxx_SUPPORTED #defines in 474jconfig.h and jmorecfg.h. If at all possible, we recommend that you leave in 475decoder support for all valid JPEG files, to ensure that you can read anyone's 476output. Taking out support for image file formats that you don't use is the 477most painless way to make the programs smaller. Another possibility is to 478remove some of the DCT methods: in particular, the "IFAST" method may not be 479enough faster than the others to be worth keeping on your machine. (If you 480do remove ISLOW or IFAST, be sure to redefine JDCT_DEFAULT or JDCT_FASTEST 481to a supported method, by adding a #define in jconfig.h.) 482 483 484OPTIMIZATION 485============ 486 487Unless you own a Cray, you'll probably be interested in making the JPEG 488software go as fast as possible. This section covers some machine-dependent 489optimizations you may want to try. We suggest that before trying any of 490this, you first get the basic installation to pass the self-test step. 491Repeat the self-test after any optimization to make sure that you haven't 492broken anything. 493 494The integer DCT routines perform a lot of multiplications. These 495multiplications must yield 32-bit results, but none of their input values 496are more than 16 bits wide. On many machines, notably the 680x0 and 80x86 497CPUs, a 16x16=>32 bit multiply instruction is faster than a full 32x32=>32 498bit multiply. Unfortunately there is no portable way to specify such a 499multiplication in C, but some compilers can generate one when you use the 500right combination of casts. See the MULTIPLYxxx macro definitions in 501jdct.h. If your compiler makes "int" be 32 bits and "short" be 16 bits, 502defining SHORTxSHORT_32 is fairly likely to work. When experimenting with 503alternate definitions, be sure to test not only whether the code still works 504(use the self-test), but also whether it is actually faster --- on some 505compilers, alternate definitions may compute the right answer, yet be slower 506than the default. Timing cjpeg on a large PGM (grayscale) input file is the 507best way to check this, as the DCT will be the largest fraction of the runtime 508in that mode. (Note: some of the distributed compiler-specific jconfig files 509already contain #define switches to select appropriate MULTIPLYxxx 510definitions.) 511 512If your machine has sufficiently fast floating point hardware, you may find 513that the float DCT method is faster than the integer DCT methods, even 514after tweaking the integer multiply macros. In that case you may want to 515make the float DCT be the default method. (The only objection to this is 516that float DCT results may vary slightly across machines.) To do that, add 517"#define JDCT_DEFAULT JDCT_FLOAT" to jconfig.h. Even if you don't change 518the default, you should redefine JDCT_FASTEST, which is the method selected 519by djpeg's -fast switch. Don't forget to update the documentation files 520(usage.txt and/or cjpeg.1, djpeg.1) to agree with what you've done. 521 522If access to "short" arrays is slow on your machine, it may be a win to 523define type JCOEF as int rather than short. This will cost a good deal of 524memory though, particularly in some multi-pass modes, so don't do it unless 525you have memory to burn and short is REALLY slow. 526 527If your compiler can compile function calls in-line, make sure the INLINE 528macro in jmorecfg.h is defined as the keyword that marks a function 529inline-able. Some compilers have a switch that tells the compiler to inline 530any function it thinks is profitable (e.g., -finline-functions for gcc). 531Enabling such a switch is likely to make the compiled code bigger but faster. 532 533In general, it's worth trying the maximum optimization level of your compiler, 534and experimenting with any optional optimizations such as loop unrolling. 535(Unfortunately, far too many compilers have optimizer bugs ... be prepared to 536back off if the code fails self-test.) If you do any experimentation along 537these lines, please report the optimal settings to jpeg-info@uc.ag so we 538can mention them in future releases. Be sure to specify your machine 539and compiler version. 540 541 542HINTS FOR SPECIFIC SYSTEMS 543========================== 544 545We welcome reports on changes needed for systems not mentioned here. Submit 546'em to jpeg-info@uc.ag. Also, if configure or ckconfig.c is wrong about how 547to configure the JPEG software for your system, please let us know. 548 549 550Acorn RISC OS: 551 552(Thanks to Simon Middleton for these hints on compiling with Desktop C.) 553After renaming the files according to Acorn conventions, take a copy of 554makefile.ansi, change all occurrences of 'libjpeg.a' to 'libjpeg.o' and 555change these definitions as indicated: 556 557CFLAGS= -throwback -IC: -Wn 558LDLIBS=C:o.Stubs 559SYSDEPMEM=jmemansi.o 560LN=Link 561AR=LibFile -c -o 562 563Also add a new line '.c.o:; $(cc) $< $(cflags) -c -o $@'. Remove the 564lines '$(RM) libjpeg.o' and '$(AR2) libjpeg.o' and the 'jconfig.h' 565dependency section. 566 567Copy jconfig.txt to jconfig.h. Edit jconfig.h to define TWO_FILE_COMMANDLINE 568and CHAR_IS_UNSIGNED. 569 570Run the makefile using !AMU not !Make. If you want to use the 'clean' and 571'test' makefile entries then you will have to fiddle with the syntax a bit 572and rename the test files. 573 574 575Amiga: 576 577SAS C 6.50 reportedly is too buggy to compile the IJG code properly. 578A patch to update to 6.51 is available from SAS or AmiNet FTP sites. 579 580The supplied config files are set up to use jmemname.c as the memory 581manager, with temporary files being created on the device named by 582"JPEGTMP:". 583 584 585Atari ST/STE/TT: 586 587Copy the project files makcjpeg.st, makdjpeg.st, maktjpeg.st, and makljpeg.st 588to cjpeg.prj, djpeg.prj, jpegtran.prj, and libjpeg.prj respectively. The 589project files should work as-is with Pure C. For Turbo C, change library 590filenames "pc..." to "tc..." in each project file. Note that libjpeg.prj 591selects jmemansi.c as the recommended memory manager. You'll probably want to 592adjust the DEFAULT_MAX_MEM setting --- you want it to be a couple hundred K 593less than your normal free memory. Put "#define DEFAULT_MAX_MEM nnnn" into 594jconfig.h to do this. 595 596To use the 68881/68882 coprocessor for the floating point DCT, add the 597compiler option "-8" to the project files and replace pcfltlib.lib with 598pc881lib.lib in cjpeg.prj and djpeg.prj. Or if you don't have a 599coprocessor, you may prefer to remove the float DCT code by undefining 600DCT_FLOAT_SUPPORTED in jmorecfg.h (since without a coprocessor, the float 601code will be too slow to be useful). In that case, you can delete 602pcfltlib.lib from the project files. 603 604Note that you must make libjpeg.lib before making cjpeg.ttp, djpeg.ttp, 605or jpegtran.ttp. You'll have to perform the self-test by hand. 606 607We haven't bothered to include project files for rdjpgcom and wrjpgcom. 608Those source files should just be compiled by themselves; they don't 609depend on the JPEG library. You can use the default.prj project file 610of the Pure C distribution to make the programs. 611 612There is a bug in some older versions of the Turbo C library which causes the 613space used by temporary files created with "tmpfile()" not to be freed after 614an abnormal program exit. If you check your disk afterwards, you will find 615cluster chains that are allocated but not used by a file. This should not 616happen in cjpeg/djpeg/jpegtran, since we enable a signal catcher to explicitly 617close temp files before exiting. But if you use the JPEG library with your 618own code, be sure to supply a signal catcher, or else use a different 619system-dependent memory manager. 620 621 622Cray: 623 624Should you be so fortunate as to be running JPEG on a Cray YMP, there is a 625compiler bug in old versions of Cray's Standard C (prior to 3.1). If you 626still have an old compiler, you'll need to insert a line reading 627"#pragma novector" just before the loop 628 for (i = 1; i <= (int) htbl->bits[l]; i++) 629 huffsize[p++] = (char) l; 630in fix_huff_tbl (in V5beta1, line 204 of jchuff.c and line 176 of jdhuff.c). 631[This bug may or may not still occur with the current IJG code, but it's 632probably a dead issue anyway...] 633 634 635HP-UX: 636 637If you have HP-UX 7.05 or later with the "software development" C compiler, 638you should run the compiler in ANSI mode. If using the configure script, 639say 640 ./configure CC='cc -Aa' 641(or -Ae if you prefer). If configuring by hand, use makefile.ansi and add 642"-Aa" to the CFLAGS line in the makefile. 643 644If you have a pre-7.05 system, or if you are using the non-ANSI C compiler 645delivered with a minimum HP-UX system, then you must use makefile.unix 646(and do NOT add -Aa); or just run configure without the CC option. 647 648On HP 9000 series 800 machines, the HP C compiler is buggy in revisions prior 649to A.08.07. If you get complaints about "not a typedef name", you'll have to 650use makefile.unix, or run configure without the CC option. 651 652 653Macintosh, generic comments: 654 655The supplied user-interface files (cjpeg.c, djpeg.c, etc) are set up to 656provide a Unix-style command line interface. You can use this interface on 657the Mac by means of the ccommand() library routine provided by Metrowerks 658CodeWarrior or Think C. This is only appropriate for testing the library, 659however; to make a user-friendly equivalent of cjpeg/djpeg you'd really want 660to develop a Mac-style user interface. There isn't a complete example 661available at the moment, but there are some helpful starting points: 6621. Sam Bushell's free "To JPEG" applet provides drag-and-drop conversion to 663JPEG under System 7 and later. This only illustrates how to use the 664compression half of the library, but it does a very nice job of that part. 665The CodeWarrior source code is available from http://www.pobox.com/~jsam. 6662. Jim Brunner prepared a Mac-style user interface for both compression and 667decompression. Unfortunately, it hasn't been updated since IJG v4, and 668the library's API has changed considerably since then. Still it may be of 669some help, particularly as a guide to compiling the IJG code under Think C. 670Jim's code is available from the Info-Mac archives, at sumex-aim.stanford.edu 671or mirrors thereof; see file /info-mac/dev/src/jpeg-convert-c.hqx. 672 673jmemmac.c is the recommended memory manager back end for Macintosh. It uses 674NewPtr/DisposePtr instead of malloc/free, and has a Mac-specific 675implementation of jpeg_mem_available(). It also creates temporary files that 676follow Mac conventions. (That part of the code relies on System-7-or-later OS 677functions. See the comments in jmemmac.c if you need to run it on System 6.) 678NOTE that USE_MAC_MEMMGR must be defined in jconfig.h to use jmemmac.c. 679 680You can also use jmemnobs.c, if you don't care about handling images larger 681than available memory. If you use any memory manager back end other than 682jmemmac.c, we recommend replacing "malloc" and "free" by "NewPtr" and 683"DisposePtr", because Mac C libraries often have peculiar implementations of 684malloc/free. (For instance, free() may not return the freed space to the 685Mac Memory Manager. This is undesirable for the IJG code because jmemmgr.c 686already clumps space requests.) 687 688 689Macintosh, Metrowerks CodeWarrior: 690 691The Unix-command-line-style interface can be used by defining USE_CCOMMAND. 692You'll also need to define TWO_FILE_COMMANDLINE to avoid stdin/stdout. 693This means that when using the cjpeg/djpeg programs, you'll have to type the 694input and output file names in the "Arguments" text-edit box, rather than 695using the file radio buttons. (Perhaps USE_FDOPEN or USE_SETMODE would 696eliminate the problem, but I haven't heard from anyone who's tried it.) 697 698On 680x0 Macs, Metrowerks defines type "double" as a 10-byte IEEE extended 699float. jmemmgr.c won't like this: it wants sizeof(ALIGN_TYPE) to be a power 700of 2. Add "#define ALIGN_TYPE long" to jconfig.h to eliminate the complaint. 701 702The supplied configuration file jconfig.mac can be used for your jconfig.h; 703it includes all the recommended symbol definitions. If you have AppleScript 704installed, you can run the supplied script makeproj.mac to create CodeWarrior 705project files for the library and the testbed applications, then build the 706library and applications. (Thanks to Dan Sears and Don Agro for this nifty 707hack, which saves us from trying to maintain CodeWarrior project files as part 708of the IJG distribution...) 709 710 711Macintosh, Think C: 712 713The documentation in Jim Brunner's "JPEG Convert" source code (see above) 714includes detailed build instructions for Think C; it's probably somewhat 715out of date for the current release, but may be helpful. 716 717If you want to build the minimal command line version, proceed as follows. 718You'll have to prepare project files for the programs; we don't include any 719in the distribution since they are not text files. Use the file lists in 720any of the supplied makefiles as a guide. Also add the ANSI and Unix C 721libraries in a separate segment. You may need to divide the JPEG files into 722more than one segment; we recommend dividing compression and decompression 723modules. Define USE_CCOMMAND in jconfig.h so that the ccommand() routine is 724called. You must also define TWO_FILE_COMMANDLINE because stdin/stdout 725don't handle binary data correctly. 726 727On 680x0 Macs, Think C defines type "double" as a 12-byte IEEE extended float. 728jmemmgr.c won't like this: it wants sizeof(ALIGN_TYPE) to be a power of 2. 729Add "#define ALIGN_TYPE long" to jconfig.h to eliminate the complaint. 730 731jconfig.mac should work as a jconfig.h configuration file for Think C, 732but the makeproj.mac AppleScript script is specific to CodeWarrior. Sorry. 733 734 735MIPS R3000: 736 737MIPS's cc version 1.31 has a rather nasty optimization bug. Don't use -O 738if you have that compiler version. (Use "cc -V" to check the version.) 739Note that the R3000 chip is found in workstations from DEC and others. 740 741 742MS-DOS, generic comments for 16-bit compilers: 743 744The IJG code is designed to work well in 80x86 "small" or "medium" memory 745models (i.e., data pointers are 16 bits unless explicitly declared "far"; 746code pointers can be either size). You may be able to use small model to 747compile cjpeg or djpeg by itself, but you will probably have to use medium 748model for any larger application. This won't make much difference in 749performance. You *will* take a noticeable performance hit if you use a 750large-data memory model, and you should avoid "huge" model if at all 751possible. Be sure that NEED_FAR_POINTERS is defined in jconfig.h if you use 752a small-data memory model; be sure it is NOT defined if you use a large-data 753model. (The supplied makefiles and jconfig files for Borland and Microsoft C 754compile in medium model and define NEED_FAR_POINTERS.) 755 756The DOS-specific memory manager, jmemdos.c, should be used if possible. 757It needs some assembly-code routines which are in jmemdosa.asm; make sure 758your makefile assembles that file and includes it in the library. If you 759don't have a suitable assembler, you can get pre-assembled object files for 760jmemdosa by FTP from ftp.uu.net:/graphics/jpeg/jdosaobj.zip. (DOS-oriented 761distributions of the IJG source code often include these object files.) 762 763When using jmemdos.c, jconfig.h must define USE_MSDOS_MEMMGR and must set 764MAX_ALLOC_CHUNK to less than 64K (65520L is a typical value). If your 765C library's far-heap malloc() can't allocate blocks that large, reduce 766MAX_ALLOC_CHUNK to whatever it can handle. 767 768If you can't use jmemdos.c for some reason --- for example, because you 769don't have an assembler to assemble jmemdosa.asm --- you'll have to fall 770back to jmemansi.c or jmemname.c. You'll probably still need to set 771MAX_ALLOC_CHUNK in jconfig.h, because most DOS C libraries won't malloc() 772more than 64K at a time. IMPORTANT: if you use jmemansi.c or jmemname.c, 773you will have to compile in a large-data memory model in order to get the 774right stdio library. Too bad. 775 776wrjpgcom needs to be compiled in large model, because it malloc()s a 64KB 777work area to hold the comment text. If your C library's malloc can't 778handle that, reduce MAX_COM_LENGTH as necessary in wrjpgcom.c. 779 780Most MS-DOS compilers treat stdin/stdout as text files, so you must use 781two-file command line style. But if your compiler has either fdopen() or 782setmode(), you can use one-file style if you like. To do this, define 783USE_SETMODE or USE_FDOPEN so that stdin/stdout will be set to binary mode. 784(USE_SETMODE seems to work with more DOS compilers than USE_FDOPEN.) You 785should test that I/O through stdin/stdout produces the same results as I/O 786to explicitly named files... the "make test" procedures in the supplied 787makefiles do NOT use stdin/stdout. 788 789 790MS-DOS, generic comments for 32-bit compilers: 791 792None of the above comments about memory models apply if you are using a 79332-bit flat-memory-space environment, such as DJGPP or Watcom C. (And you 794should use one if you have it, as performance will be much better than 7958086-compatible code!) For flat-memory-space compilers, do NOT define 796NEED_FAR_POINTERS, and do NOT use jmemdos.c. Use jmemnobs.c if the 797environment supplies adequate virtual memory, otherwise use jmemansi.c or 798jmemname.c. 799 800You'll still need to be careful about binary I/O through stdin/stdout. 801See the last paragraph of the previous section. 802 803 804MS-DOS, Borland C: 805 806Be sure to convert all the source files to DOS text format (CR/LF newlines). 807Although Borland C will often work OK with unmodified Unix (LF newlines) 808source files, sometimes it will give bogus compile errors. 809"Illegal character '#'" is the most common such error. (This is true with 810Borland C 3.1, but perhaps is fixed in newer releases.) 811 812If you want one-file command line style, just undefine TWO_FILE_COMMANDLINE. 813jconfig.bcc already includes #define USE_SETMODE to make this work. 814(fdopen does not work correctly.) 815 816 817MS-DOS, Microsoft C: 818 819makefile.mc6 works with Microsoft C, DOS Visual C++, etc. It should only 820be used if you want to build a 16-bit (small or medium memory model) program. 821 822If you want one-file command line style, just undefine TWO_FILE_COMMANDLINE. 823jconfig.mc6 already includes #define USE_SETMODE to make this work. 824(fdopen does not work correctly.) 825 826Note that this makefile assumes that the working copy of itself is called 827"makefile". If you want to call it something else, say "makefile.mak", 828be sure to adjust the dependency line that reads "$(RFILE) : makefile". 829Otherwise the make will fail because it doesn't know how to create "makefile". 830Worse, some releases of Microsoft's make utilities give an incorrect error 831message in this situation. 832 833Old versions of MS C fail with an "out of macro expansion space" error 834because they can't cope with the macro TRACEMS8 (defined in jerror.h). 835If this happens to you, the easiest solution is to change TRACEMS8 to 836expand to nothing. You'll lose the ability to dump out JPEG coefficient 837tables with djpeg -debug -debug, but at least you can compile. 838 839Original MS C 6.0 is very buggy; it compiles incorrect code unless you turn 840off optimization entirely (remove -O from CFLAGS). 6.00A is better, but it 841still generates bad code if you enable loop optimizations (-Ol or -Ox). 842 843MS C 8.0 crashes when compiling jquant1.c with optimization switch /Oo ... 844which is on by default. To work around this bug, compile that one file 845with /Oo-. 846 847 848Microsoft Windows (all versions), generic comments: 849 850Some Windows system include files define typedef boolean as "unsigned char". 851The IJG code also defines typedef boolean, but we make it "int" by default. 852This doesn't affect the IJG programs because we don't import those Windows 853include files. But if you use the JPEG library in your own program, and some 854of your program's files import one definition of boolean while some import the 855other, you can get all sorts of mysterious problems. A good preventive step 856is to make the IJG library use "unsigned char" for boolean. To do that, 857add something like this to your jconfig.h file: 858 /* Define "boolean" as unsigned char, not int, per Windows custom */ 859 #ifndef __RPCNDR_H__ /* don't conflict if rpcndr.h already read */ 860 typedef unsigned char boolean; 861 #endif 862 #define HAVE_BOOLEAN /* prevent jmorecfg.h from redefining it */ 863(This is already in jconfig.vc, by the way.) 864 865windef.h contains the declarations 866 #define far 867 #define FAR far 868Since jmorecfg.h tries to define FAR as empty, you may get a compiler 869warning if you include both jpeglib.h and windef.h (which windows.h 870includes). To suppress the warning, you can put "#ifndef FAR"/"#endif" 871around the line "#define FAR" in jmorecfg.h. 872(Something like this is already in jmorecfg.h, by the way.) 873 874When using the library in a Windows application, you will almost certainly 875want to modify or replace the error handler module jerror.c, since our 876default error handler does a couple of inappropriate things: 877 1. it tries to write error and warning messages on stderr; 878 2. in event of a fatal error, it exits by calling exit(). 879 880A simple stopgap solution for problem 1 is to replace the line 881 fprintf(stderr, "%s\n", buffer); 882(in output_message in jerror.c) with 883 MessageBox(GetActiveWindow(),buffer,"JPEG Error",MB_OK|MB_ICONERROR); 884It's highly recommended that you at least do that much, since otherwise 885error messages will disappear into nowhere. (Beginning with IJG v6b, this 886code is already present in jerror.c; just define USE_WINDOWS_MESSAGEBOX in 887jconfig.h to enable it.) 888 889The proper solution for problem 2 is to return control to your calling 890application after a library error. This can be done with the setjmp/longjmp 891technique discussed in libjpeg.txt and illustrated in example.c. (NOTE: 892some older Windows C compilers provide versions of setjmp/longjmp that 893don't actually work under Windows. You may need to use the Windows system 894functions Catch and Throw instead.) 895 896The recommended memory manager under Windows is jmemnobs.c; in other words, 897let Windows do any virtual memory management needed. You should NOT use 898jmemdos.c nor jmemdosa.asm under Windows. 899 900For Windows 3.1, we recommend compiling in medium or large memory model; 901for newer Windows versions, use a 32-bit flat memory model. (See the MS-DOS 902sections above for more info about memory models.) In the 16-bit memory 903models only, you'll need to put 904 #define MAX_ALLOC_CHUNK 65520L /* Maximum request to malloc() */ 905into jconfig.h to limit allocation chunks to 64Kb. (Without that, you'd 906have to use huge memory model, which slows things down unnecessarily.) 907jmemnobs.c works without modification in large or flat memory models, but to 908use medium model, you need to modify its jpeg_get_large and jpeg_free_large 909routines to allocate far memory. In any case, you might like to replace 910its calls to malloc and free with direct calls on Windows memory allocation 911functions. 912 913You may also want to modify jdatasrc.c and jdatadst.c to use Windows file 914operations rather than fread/fwrite. This is only necessary if your C 915compiler doesn't provide a competent implementation of C stdio functions. 916 917You might want to tweak the RGB_xxx macros in jmorecfg.h so that the library 918will accept or deliver color pixels in BGR sample order, not RGB; BGR order 919is usually more convenient under Windows. Note that this change will break 920the sample applications cjpeg/djpeg, but the library itself works fine. 921 922 923Many people want to convert the IJG library into a DLL. This is reasonably 924straightforward, but watch out for the following: 925 926 1. Don't try to compile as a DLL in small or medium memory model; use 927large model, or even better, 32-bit flat model. Many places in the IJG code 928assume the address of a local variable is an ordinary (not FAR) pointer; 929that isn't true in a medium-model DLL. 930 931 2. Microsoft C cannot pass file pointers between applications and DLLs. 932(See Microsoft Knowledge Base, PSS ID Number Q50336.) So jdatasrc.c and 933jdatadst.c don't work if you open a file in your application and then pass 934the pointer to the DLL. One workaround is to make jdatasrc.c/jdatadst.c 935part of your main application rather than part of the DLL. 936 937 3. You'll probably need to modify the macros GLOBAL() and EXTERN() to 938attach suitable linkage keywords to the exported routine names. Similarly, 939you'll want to modify METHODDEF() and JMETHOD() to ensure function pointers 940are declared in a way that lets application routines be called back through 941the function pointers. These macros are in jmorecfg.h. Typical definitions 942for a 16-bit DLL are: 943 #define GLOBAL(type) type _far _pascal _loadds _export 944 #define EXTERN(type) extern type _far _pascal _loadds 945 #define METHODDEF(type) static type _far _pascal 946 #define JMETHOD(type,methodname,arglist) \ 947 type (_far _pascal *methodname) arglist 948For a 32-bit DLL you may want something like 949 #define GLOBAL(type) __declspec(dllexport) type 950 #define EXTERN(type) extern __declspec(dllexport) type 951Although not all the GLOBAL routines are actually intended to be called by 952the application, the performance cost of making them all DLL entry points is 953negligible. 954 955The unmodified IJG library presents a very C-specific application interface, 956so the resulting DLL is only usable from C or C++ applications. There has 957been some talk of writing wrapper code that would present a simpler interface 958usable from other languages, such as Visual Basic. This is on our to-do list 959but hasn't been very high priority --- any volunteers out there? 960 961 962Microsoft Windows, Borland C: 963 964The provided jconfig.bcc should work OK in a 32-bit Windows environment, 965but you'll need to tweak it in a 16-bit environment (you'd need to define 966NEED_FAR_POINTERS and MAX_ALLOC_CHUNK). Beware that makefile.bcc will need 967alteration if you want to use it for Windows --- in particular, you should 968use jmemnobs.c not jmemdos.c under Windows. 969 970Borland C++ 4.5 fails with an internal compiler error when trying to compile 971jdmerge.c in 32-bit mode. If enough people complain, perhaps Borland will fix 972it. In the meantime, the simplest known workaround is to add a redundant 973definition of the variable range_limit in h2v1_merged_upsample(), at the head 974of the block that handles odd image width (about line 268 in v6 jdmerge.c): 975 /* If image width is odd, do the last output column separately */ 976 if (cinfo->output_width & 1) { 977 register JSAMPLE * range_limit = cinfo->sample_range_limit; /* ADD THIS */ 978 cb = GETJSAMPLE(*inptr1); 979Pretty bizarre, especially since the very similar routine h2v2_merged_upsample 980doesn't trigger the bug. 981Recent reports suggest that this bug does not occur with "bcc32a" (the 982Pentium-optimized version of the compiler). 983 984Another report from a user of Borland C 4.5 was that incorrect code (leading 985to a color shift in processed images) was produced if any of the following 986optimization switch combinations were used: 987 -Ot -Og 988 -Ot -Op 989 -Ot -Om 990So try backing off on optimization if you see such a problem. (Are there 991several different releases all numbered "4.5"??) 992 993 994Microsoft Windows, Microsoft Visual C++: 995 996jconfig.vc should work OK with any Microsoft compiler for a 32-bit memory 997model. makefile.vc is intended for command-line use. (If you are using 998the Developer Studio environment, you may prefer the DevStudio project 999files; see below.) 1000 1001IJG JPEG 7 adds extern "C" to jpeglib.h. This avoids the need to put 1002extern "C" { ... } around #include "jpeglib.h" in your C++ application. 1003You can also force VC++ to treat the library as C++ code by renaming 1004all the *.c files to *.cpp (and adjusting the makefile to match). 1005In this case you also need to define the symbol DONT_USE_EXTERN_C in 1006the configuration to prevent jpeglib.h from using extern "C". 1007 1008 1009Microsoft Windows, Microsoft Visual C++ 6 Developer Studio: 1010 1011We include makefiles that should work as project files in DevStudio 6.0 or 1012later. There is a library makefile that builds the IJG library as a static 1013Win32 library, and application makefiles that build the sample applications 1014as Win32 console applications. (Even if you only want the library, we 1015recommend building the applications so that you can run the self-test.) 1016 1017To use: 10181. Open the command prompt, change to the main directory and execute the 1019 command line 1020 NMAKE /f makefile.vc setup-vc6 1021 This will move jconfig.vc to jconfig.h and makefiles to project files. 1022 (Note that the renaming is critical!) 10232. Open the workspace file jpeg.dsw, build the library project. 1024 (If you are using DevStudio more recent than 6.0, you'll probably 1025 get a message saying that the project files are being updated.) 10263. Open the workspace file apps.dsw, build the application projects. 10274. To perform the self-test, execute the command line 1028 NMAKE /f makefile.vc test-build 10295. Move the application .exe files from `app`\Release to an 1030 appropriate location on your path. 1031 1032 1033Microsoft Windows, Microsoft Visual C++ 2010 Developer Studio (v10): 1034 1035We include makefiles that should work as project files in Visual Studio 10362010 or later. There is a library makefile that builds the IJG library 1037as a static Win32 library, and application makefiles that build the sample 1038applications as Win32 console applications. (Even if you only want the 1039library, we recommend building the applications so that you can run the 1040self-test.) 1041 1042To use: 10431. Open the command prompt, change to the main directory and execute the 1044 command line 1045 NMAKE /f makefile.vc setup-v10 1046 This will move jconfig.vc to jconfig.h and makefiles to project files. 1047 (Note that the renaming is critical!) 10482. Open the solution file jpeg.sln, build the library project. 1049 (If you are using Visual Studio more recent than 2010 (v10), you'll 1050 probably get a message saying that the project files are being updated.) 10513. Open the solution file apps.sln, build the application projects. 10524. To perform the self-test, execute the command line 1053 NMAKE /f makefile.vc test-build 10545. Move the application .exe files from `app`\Release to an 1055 appropriate location on your path. 1056 1057Note: 1058There seems to be an optimization bug in the compiler which causes the 1059self-test to fail with the color quantization option. 1060We have disabled optimization for the file jquant2.c in the library 1061project file which causes the self-test to pass properly. 1062 1063 1064OS/2, Borland C++: 1065 1066Watch out for optimization bugs in older Borland compilers; you may need 1067to back off the optimization switch settings. See the comments in 1068makefile.bcc. 1069 1070 1071SGI: 1072 1073On some SGI systems, you may need to set "AR2= ar -ts" in the Makefile. 1074If you are using configure, you can do this by saying 1075 ./configure RANLIB='ar -ts' 1076This change is not needed on all SGIs. Use it only if the make fails at the 1077stage of linking the completed programs. 1078 1079On the MIPS R4000 architecture (Indy, etc.), the compiler option "-mips2" 1080reportedly speeds up the float DCT method substantially, enough to make it 1081faster than the default int method (but still slower than the fast int 1082method). If you use -mips2, you may want to alter the default DCT method to 1083be float. To do this, put "#define JDCT_DEFAULT JDCT_FLOAT" in jconfig.h. 1084 1085 1086VMS: 1087 1088On an Alpha/VMS system with MMS, be sure to use the "/Marco=Alpha=1" 1089qualifier with MMS when building the JPEG package. 1090 1091VAX/VMS v5.5-1 may have problems with the test step of the build procedure 1092reporting differences when it compares the original and test images. If the 1093error points to the last block of the files, it is most likely bogus and may 1094be safely ignored. It seems to be because the files are Stream_LF and 1095Backup/Compare has difficulty with the (presumably) null padded files. 1096This problem was not observed on VAX/VMS v6.1 or AXP/VMS v6.1. 1097