README for Informix Database Driver for Perl Version 1.00.PC1 (2000-03-03) Portions Copyright 1994-96 Alligator Descartes Portions Copyright 1996-99 Jonathan Leffler Portions Copyright 2000 Informix Software You may distribute under the terms of either the GNU General Public License or the Artistic License, as specified in the Perl README file. --------------------------------------------------------------------------- PREREQUISITES You need the following five items to build Informix Database Driver for Perl: 1. Perl 5.004 or later. 2. DBI 1.13 or later. 3. A C compiler that accepts function prototypes (such as GCC 2.95.2). 4. Informix ESQL/C 5.00 or later, or Client SDK 2.00 or later. 5. A 'stores' database to which you can connect without specifying username or password and in which you can create tables. Ideally, the connection should not use shared memory (neither olipcshm nor onipcshm), and you should have DBA privileges in the database. The basic build steps are: perl Makefile.PL make make test make install If you are not sure about any of these items or if one of the build steps above fails when you run it, you need to read the information below. Read the BUILD AND TEST ENVIRONMENT section when you set your environment variables. If you run into problems during the build or test phases, read the section IF YOU HAVE PROBLEMS BUILDING DBD::INFORMIX. When you have DBD::Informix working *and* installed, use the ItWorks Perl script (previously a shell script) to report your successful installation. Be sure to review the output (especially the email address it deduces), and send it in an email to the maintainers at dbd-informix@informix.com with the subject "DBD::Informix - It Works". TECHNICAL SUPPORT For information on Informix Technical Support, please run: perldoc DBD::Informix::TechSupport If you have purchased technical support, you will be given a licence number. You can build DBD::Informix and specify that licence number when you compile it: perl Makefile.PL make INFORMIX_LICENCE="DBD#I000000" Alternatively, once you have installed the product, you can use: InformixTechSupport -l DBD#I000000 PERL If you do not have Perl Version 5.004 or later installed, you should build, test, and install it before you do anything else. You should use at least Perl Version 5.004_04 and really should use Perl 5.005_03. Be sure to compile it with the options necessary to make the C compiler accept prototypes because the DBD::Informix code uses prototypes. If it is possible on your platform (and it is not possible on some platforms, including AIX), use the C compiler to create shared libraries even if the Perl configuration script suggests that you use the 'ld' program directly. People who do not use the C compiler to create the shared libraries have often had many problems, and those who use it have generally had very few. Note also that to install DBD::Informix, you must be able to put files under the Perl lib directory. For alternative options, see the Notes/nonroot.install file. If you are working on NT, you should use the Perl binaries available from ActiveState at http://www.activestate.com. This site also provides pre-compiled versions of many Perl modules, including DBI (but not DBD::Informix at 2000-02-01). DBI If you do not have DBI Version 1.13 or later installed, you should build, test, and install it. Although some old versions of DBD::Informix allowed you to use older versions of DBI than the version it was developed with, the current versions of DBD::Informix do not. Note that if the 'perl Makefile.PL' process requires a later version of DBI, these notes contain a minor bug. Use the version the process says. Note also that "perl -MCPAN -e 'install Bundle::DBI'" gets the latest version. Similarly, if the process rejects your version of Perl, you need to upgrade your Perl. C COMPILER To build DBD::Informix, the C compiler must accept function prototypes. This is not a problem on any computer to which Informix is currently ported (though the HP-UX bundled compiler does not accept prototypes at all, and the ANSI compiler does not accept them unless told to do so). If you have problems, get the GNU C Compiler, Version 2.95.2 or later. It is available from http://egcs.cygnus.com. Use the same compiler to build Perl, DBI, and DBD::Informix. Note that Perl 5.005 requires a C compiler that handles prototypes. The Notes/hpux file contains information about how to compile the GNU C Compiler on HP-UX. ESQL/C OR CLIENT SDK You must have a version of Informix ESQL/C or Informix Client SDK installed on the computer where you wish to compile DBD::Informix. Informix Connect is not sufficient. The Notes/FAQ file contains more information about what you need. ESQL/C Versions 4.1x and earlier are not (and will not be) supported by DBD::Informix. ESQL/C Versions 5.00 and up should be OK. If you do not have ESQL/C, DBD::Informix will not work. You can probably get Informix Client SDK at a reasonable cost. If you have Informix Open Database Connectivity (ODBC) drivers available to you, consider using DBD::ODBC instead. If you are on Linux, you should investigate the software available from the following Informix Web sites: http://www.informix.com/ http://www.informix.com/idn You must also be able to compile, link, and run ESQL/C programs with your setup. Makefile.PL will test that you can do this, but you can save time if you ensure this beforehand. If you cannot compile, link and run free-standing ESQL/C programs, you certainly cannot link DBD::Informix into Perl. For help with environment variable settings, consult the information below and also the Notes/environment.variables file. (If you are not yet familiar with how to set environment variables, be sure to get and read a UNIX primer such as "Learning the Unix System, 4th Edition" from O'Reilly, http://www.oreilly.com/, and be prepared for a major learning exercise). DBD::Informix, Version 1.00, provides limited support for user-defined data types (UDTs), treating them as CHAR(255). To handle BLOBs and CLOBs, use LOTOFILE() when you fetch the data and FILETOBLOB() or FILETOCLOB() when you insert data. To handle nonblob UDTs that exceed 255 characters in length, use server-side cast to lvarchar, as in select mycol::lvarchar from mytab; Although you can compile DBD::Informix under Client SDK 2.01.UC1 (ESQL/C 9.14.UC1), some tests will fail because of a bug in this version of Client SDK. If you use the same code under a later version of Client SDK, the failures in t/t31nulls.t and t/t32nulls.t now will pass. Most versions of ESQL/C that support shared libraries have shared linking as the default, which is correct. Other versions reportedly have static linking as the default, which is a nuisance. The Makefile.PL will add the '-shared' flag to the ESQL/C command line to try to force shared libraries for ESQL/C Versions 7.20 and up. If this does not work for you or if you are building a static Perl, you will need to set the environment variable DBD_INFORMIX_ESQLC_LINKAGE either to nothing if your version of ESQL/C does not support the '-shared' option or to '-static' to force static linkage. You could also use this environment variable to bootstrap any special ESQL/C compiler options into the build process (such as '-thread' if you experiment with threaded Perl and threaded ESQL/C); you are advised to set '-static' or '-shared' as well. STORES DATABASE Unless you have a 'stores' database that you can connect to without specifying a username or password (and in which you can create tables), you will need to set various environment variables to tell the build and test code for DBD::Informix which database to use for testing and exactly how to connect to it. For more details on the environment variables that you can set, see the BUILD AND TEST ENVIRONMENT section. You *must* have a fully working Informix environment before you try to build and test DBD::Informix. This means you must have access to at least one database where you have at least RESOURCE level privileges. If you do not understand what this means, refer to the "Informix Guide to SQL: Syntax" manual and read the discussion of the GRANT statement. You can obtain a PDF version of any Informix manual from: http://www.informix.com/answers If you do not have RESOURCE (preferably DBA) privileges on a database, consider creating a database called 'stores' for testing. If you do not have DBA privileges, the test t/t55mdata.t may fail but, unless you have other problems, you can disregard this failure. Note that the tests for DBD::Informix create and drop their own tables. Most of the tests use temporary tables. It does not matter whether the test database has database logging, though a logged database allows more features to be tested than an unlogged database. You can use a brand new, empty database for testing. When DBD::Informix creates any database object, the name begins with "dbd_ix_". After running the test t/t99clean.t, nothing should remain from the testing. If you find any leftovers, report them to the maintenance team. DBD::Informix has one test that creates a database and then drops it. SHARED MEMORY CONNECTIONS The multiple connection tests use two databases for preference (though the tests will use the same database twice if you do not specify two separate databases). As of version 0.95, the esqltest program will report that both connections use shared memory and will allow you to proceed after writing a message. The actual test scripts attempt to detect that the two connections both use shared memory connections and skip the tests. However, if you run into problems with shared memory connections (for example, error -27000 from the esqltest program), read Notes/olipcshm. If your databases are not on the computer where you build DBD::Informix, be sure that you have the necessary privileges the necessary privileges to connect to the machine where the databases are. This may be as simple as setting DBD_INFORMIX_USERNAME and DBD_INFORMIX_PASSWORD (see BUILD AND TEST ENVIRONMENT) or might require you to get your system administrators to set up a login account for you on the computer. BUILD AND TEST ENVIRONMENT Be sure to set $INFORMIXDIR even if the software is installed in /usr/informix and to have $INFORMIXDIR/bin on your PATH. The build no longer works unless these environment variables are set. Also, if you are using ESQL/C version 6.x or later, you may be using ESQL/C shared libraries which are found in the directories $INFORMIXDIR/lib and $INFORMIXDIR/lib/esql. With Version 0.95, the absolute pathnames of the Informix shared libraries will be built into your DBD::Informix library by default, which means that you do not need to worry about LD_LIBRARY_PATH, LD_RUN_PATH, SHLIB_PATH, or LIBPATH at runtime. The downside of using absolute pathnames is that if you move your Informix software, you also need to rebuild and reinstall DBD::Informix. If rebuilding DBD::Informix is unacceptable, you need to set the environment variable DBD_INFORMIX_RELOCATABLE_INFORMIXDIR to a value such as "yes". The library will be built using relative names to identify the Informix shared libraries. You will be warned that this is happening. Both at test time and at run time, you need to ensure that the Informix shared libraries will be found when you run Perl with DBD::Informix. On SVR4 and Linux computers, this means adding these directories to LD_LIBRARY_PATH; on HP-UX, the variable is SHLIB_PATH; other systems may have other variable names. Note that web servers, in particular, do not propagate most environment variables unless you tell them to do so (use SetEnv or PassEnv in Apache). You need to set INFORMIXSERVER correctly unless you are using version 5.x ESQL/C. You may need to set other Informix environment variables too. Consult the Informix documentation and the Notes/environment.variables file. The documentation available from 'perldoc DBD::Informix::TestHarness' tells you how to set the DBD_INFORMIX_DATABASE, DBD_INFORMIX_USERNAME, and DBD_INFORMIX_PASSWORD environment variables for your system. The parallel environment variables with suffix 2 (DBD_INFORMIX_DATABASE2, DBD_INFORMIX_USERNAME2, DBD_INFORMIX_PASSWORD2) specify the second test database completely independently of the first. If the defaults are OK, you do not have to set any of these six environment variables. The default database is 'stores'; no username and password are supplied if none is specified. If you set the username, you must also set the password to have any effect. Although the testing does as little damage as possible, it is not a good idea to use the production database for this. The stores database is a good bet. Note that these variables have significance only when you run the DBD::Informix tests. These variables are not used by DBD::Informix itself, only by the code in DBD::Informix::TestHarness. Ideally, you should set the variables before you start the build and you should not change them until after you complete the testing. If you do change them, you should check that the esqltest program run by 'perl Makefile.PL' still gives your new environment a clean bill of health. One step in the setup process tests that you have permissions on the databases that will be used by the testing. The step compiles and runs a relatively simple ESQL/C program that opens a few databases, creates and drops some tables, and exits. If the test fails, you do you do not get a Makefile so you cannot build DBD::Informix. Note that if you set the DELIMIDENT environment variable, some tests will fail, notably t/t56tabinfo.t and t57tables.t. BUILDING DBD::INFORMIX WITH BUNDLES If you have preconfigured the Perl CPAN module and correctly set up your Informix environment, you can install DBD::Informix simply DBD::Informix by simply typing: perl -MCPAN -e 'install Bundle::DBD::Informix' This command gets the latest version of DBI (and its prerequisite modules) and the latest version of DBD::Informix, and compiles, tests, and install them all completely automatically. Before doing this, you need to be confident that things will work correctly (or that you've got good backups of your Perl installation). On the other hand, it is an extremely convenient method of updating your Perl software. When you first use the CPAN module, it will ask you many questions, including the name of the CPAN site from which to download the material, but the CPAN module saves this information for the next time and offers you a choice of sites based on continent and First consider installing the latest CPAN bundle: perl -MCPAN -e 'install Bundle::CPAN' BUILDING DBD::Informix After you install Perl, DBI, and ESQL/C, run: perl Makefile.PL The script tries to work out how to build the module. Then run: make The make command should run without errors and ideally without warnings either. If you get warnings, let us know what they are and how they how they could be fixed generically. If it fails horribly, see below. Do NOT hand edit the generated Makefile unless you are completely sure you understand the implications and are willing to make those changes manually every time the Makefile is regenerated! To make changes, always try to edit Makefile.PL, which is extensively annotated. Also refer to should also read the section on ExtUtils::MakeMaker in the 2nd Edition of 'Programming Perl'. You should never need to make any changes to the generated Makefile, nor to Makefile.PL. If you do, *let us know* so that we can try to make it automatic in a later release. Then run: make test Note that testing DBD::Informix does create some database objects (tables, views, synonyms, types, etc). The database is called 'dbd_ix_db', and the other object names start with 'dbd_ix_'. Some of the tables are permanent; most are temporary. The tests are designed to work whether the tables and database are present when the tests start or not; that means they get dropped. Do not run the tests if you have precious tables or databases that begin 'dbd_ix_'! Starting with DBD::Informix Version 0.61, the cleanup script t/t99clean.t is run at the end of the testing. It removes the tables, views, synonyms, etc and so on that DBD::Informix might have created. Running it manually ("test.one.sh t/t99clean.t") also cleans up the database objects created by testing DBD::Informix. The tests should be 100% clean if you run the cleanup script, but if you don't run that, the tests can leave the odd table or stored procedure (or user-defined data types and so on) in the database. On the reference system (Sun Ultra-10 running Solaris 7, compiling with GCC 2.95.2 or SUNWspro SC4.0.1, using either Perl 5.00503), all tests succeed when DBD::Informix is built with ESQL/C Versions 5.10.UC7, 7.24.UC1, and 9.30.UC1. It is a good idea to run: make test TEST_VERBOSE=1 You should inspect the results as not every test compares the actual data returned with the data that should be returned (though many of them do check the data which is returned). Note that the blob tests pass on SE, because the test scripts detect that the engine is SE and do not try to test blob handling. Additionally, for a really thorough scrutiny of DBD::Informix, you need to test it with at least three different databases: one created with MODE ANSI, one created with a transaction log but not MODE ANSI, and one created without any transaction logs at all: DBD_INFORMIX_DATABASE=mode_ansi make test DBD_INFORMIX_DATABASE=logged make test DBD_INFORMIX_DATABASE=unlogged make test Different tests will be skipped depending on the version of ESQL/C, the version of the database, and the logging modes of the databases you are connecting to. If you are concerned about both OnLine and SE, then the connection tests will use two different databases if you set the environment variable DBD_INFORMIX_DATABASE2 (and possibly DBD_INFORMIX_USERNAME2 and DBD_INFORMIX_PASSWORD2). You can also use one SE database and one OnLine database. You can also test with different server versions (eg using 7.2x ESQL/C to connect to a 5.0x OnLine) if you have the software available. Once you are satisfied that DBD::Informix is working correctly, you should install it: make install If you ever need to remove it, possibly as a preamble to installing a new version, you should use the old version's makefile and run: make uninstall You can then use the makefile of the new version to install. It is important to use the correct old or new makefile because the installed may be different, and if some file is made obsolete by the new version (is not used by the new version), its makefile will not uninstall the obsolete file; over time and multiple versions, this could, eventually, fill your disk completely. If you run into problems that suggest that the ESQL/C you have will not work as dynamically loaded libraries (such as on HP-UX or SCO), you should create a statically linked version of Perl with DBD::Informix linked to it. Use: make perl make test_static Please consult the Notes/Working.Versions file for information about known working versions of the software (and specific problem versions). If you are using a combination of versions which is different from any previously recorded, please send me (j.leffler@acm.org) the details for your new, successful port. If you run into major problems even getting the esqltest program to compile, you can try to compile the esqlbasic.ec program with a plain ESQL/C command: esql -o esqlbasic esqlbasic.ec If even this command will not compile, concentrate on fixing your ESQL/C ESQL/C environment before doing anything else with DBD::Informix. If it compiles but does not run, then you need to ensure that you fix the Informix environment so that you can access databases. Once this test both compiles and runs, you will probably be able to compile and test DBD::Informix. To suppress the esqltest code in Makefile.PL, you can set DBD_INFORMIX_NO_ESQLTEST=yes in your environment before you run 'perl Makefile.PL'. If you do that, however, no problem reports will be accepted; the esqltest code is critical to ensuring that DBD::Informix has some chance of compiling successfully. You can see how the esqltest code is compiled if you set DBD_INFORMIX_DEBUG_ESQLTEST=yes in your environment before you run 'perl Makefile.PL'. For more information on environment variables for both DBD:Informix and Informix, see the Notes/environment.variables file. KNOWN PROBLEMS: * DBD::Informix does not work, and never will work, with ESQL/C Version 4.0x or 4.1x. A DBD::Informix4 package is available from Goran Thyni (version 0.23 is current at 1999-12-31 and is available from CPAN). * Trying to build DBD::Informix with 5.x and 6.x ESQL/C on HP-UX (both 9.x and 10.x) seems to result in problems because the Informix products either have no shared libraries or have shared libraries that the HP-UX compilers and loader do not accept. The only way to build DBD::Informix on HP-UX is to build a statically linked version of Perl that includes the Informix libraries. This is a defect in ESQL/C rather than Perl, DBI, or DBD::Informix. If you know how to build DBD::Informix on HP-UX with ESQL/C Version 5.x or 6.x without doing it statically, be sure to tell everyone. * Compiling on HP-UX 10.20 has often been problematic. See the guidelines in the Notes/hpux file, some of which were supplied by Reed White (REED_WHITE@HP-Corvallis-om1.om.hp.com). Note that the non-ANSI C compiler that is distributed with the HP-UX base package is not suitable for compiling DBD::Informix because it does not accept prototypes. However, the file also contains information on how to compile GCC (and the other GNU software required) on HP-UX, using the bundled C compiler to start the process. If you are using ESQL/C 7.24.UC8A or Client SDK 2.30 or 2.40, you should be able to build a dynamically loadable DBD::Informix using either the ANSI C compiler or GCC without any special actions on your part. * Some problems have been encountered on SCO and seem to be related to the equivalent of LD_LIBRARY_PATH on System V Release 4 computers. Again, help or reassurance is needed on this. * If you are using GCC to compile the software (on Solaris and HP-UX 10.20 in particular, but probably on other platforms too), be sure to use GCC to build the shared libraries if at all possible, rather than trying to use 'ld' directly. If you use ld instead, you are likely to run into problems with an undefined symbol __eprintf. Please read the Notes/eprintf file contributed by Bernd Gill (bernd@dsi.co.nz). To work around the problems, set DBD_INFORMIX_USE_EPRINTF_CODE in the environment before you run 'perl Makefile.PL'. You will only run into problems with this if you have DBD_INFORMIX_ENABLE_ASSERT set in the environment. Note that earlier versions used the inverse variable DBD_INFORMIX_DISABLE_ASSERT to disable assertions. * 1996-12-31: Do not use the GNU assembler or loader with GCC on Solaris; neither works properly with shared libraries. Remove them, and rebuild and reinstall GCC immediately. IF YOU HAVE PROBLEMS BUILDING DBD::INFORMIX Read the Notes/bug.reports file, which describes what to do and where to send the failure report. Please ensure that any email message has DBD::Informix in the subject line -- thanks! IF YOU HAVE PROBLEMS USING DBD::INFORMIX If you have a problem with your own code and all the DBD::Informix tests succeed, in your initial message give the version information (see the Notes/bug.reports file), a description of the problem, a minimal test script, and the results of running the test script on your machine, along with an explanation of why the result is wrong (it may not be obvious to me) and what the correct result should be. Be sure to use DBD::Informix in the subject line of any email -- thanks! The minimal test script should preferably: 1. Use the stores database, with empty username and password fields. If the test needs a particular type of database (eg with transactions) to demonstrate the problem, alternative convenient names are 'logged', 'unlogged' and 'mode_ansi'. If you are using SE, please mention that. 2. Use temporary tables rather than permanent ones. 3. Load just enough data to show the problem. 4. Test every statement that uses a DBI function for success. 5. Clearly indicate when it fails. 6. Clearly indicate when the test succeeds. 7. The test script should not use DBI->install_driver(). If your test is failing with a core dump, the stack trace may be useful if it lists function names. The stack trace is not useful if it does not list them. The tests which come with DBD::Informix show a variety of ways of using DBD::Informix. Do not use the connection method shown in the code for test t/t83oconn.t; it is officially obsolete. DBD::Informix needs to test it, but you should not. The examples subdirectory contains some simple examples of DBI scripts for examples sub-directory. Read the examples/README file for more details. --------------------------------------------------------------------------- REMEMBER IT IS SUPPOSED TO BE FUN! Jonathan Leffler (j.leffler@acm.org) Michael Kopchenov (myk@informix.com) @(#)$Id: README version /main/63 2000-02-17 13:57:04 $ --------------------------------------------------------------------------- Original document: Copyright 1994 Tim Bunce Amendments 1: Copyright 1994-96 Alligator Descartes Amendments 2: Copyright 1996-98 Jonathan Leffler Amendments 3: Copyright 1998 Jonathan Leffler Amendments 4: Copyright 1998-99 Jonathan Leffler Amendments 5: Copyright 2000 Informix Software Inc