RE: Informix DBD build on Windows - Better

[CAMPBELL, BRIAN D (BRIAN)]

No response from anyone on this subject.  So I perservered, and began to study the source code and internals of how the DBD build works.  To discover the root cause, I think I learned far more about building DBDs and Windows linking that I wanted to know.  With some consulting help from a colleague who is more experienced with the internals of Windows builds than I am, I managed to crack this problem.  It was ultimately another Makefile problem.  Then when I retested, I got another crash from the test runs.  But fortunately it was a source problem I was familiary with (I have fixed it in previous versions of Informix::DBD).  I finally got a sucessful test run and install (with some manual work) .

 

All in all I uncovered 9 bugs in the build process for DBD.  Should I go ahead and publish my findings to this mailing list (DBI users)?

 

I need to qualify "success" in the test runs:  All tests passed, except 4 were skipped and two tests had some failures:
t\t08fork.t
t\t91udts.t

How significant are failures in these particular tests?  Details available on request.

 

-----Original Message-----
From: CAMPBELL, BRIAN D (BRIAN) [mailto:[EMAIL PROTECTED]
Sent: Wednesday, August 03, 2005 4:42 PM
To: 'dbi-users@perl.org'
Subject: Informix DBD build on Windows - major test failure

I tried building the recently released Informix DBD 2005.02 today.  I apparently have a successful compile and link. But I'm having a major problem on the test.  Below is the link input and output, and the first test.  All remaining tests look like the first test.  Help and suggestions needed.  Here's the message I'm getting upon running the first test (and all the rest).

 

Can't find 'boot_DBD__Informix' symbol in ./Informix.dll

I ran into a couple of source problems and and a couple of makefile problems that I was able to figure out and fix.  Basically what I had found appear to be "Unix-isms" that don't apply to Windows or the MSVC compiler.  And one esqlc command line option bug.  I'll report the details of all that soon when I get a working DBD.  Now I'm running "nmake test".

 

Note: this is the third time I've built an Informix DBD for Windows.  The previous times were successful, but not without grief.  The process was smother this time, because of that experience.  At least up to this point.

 

Jonathan Leffler: Yes, I've looked at the source, I can't understand how the message relates to the source code.  I could get some help on that, maybe we could crack the problem.

 

My Environment:

Windows 2000 [Version 5.00.2195] SP4

ActiveState Perl v5.8.7 built for MSWin32-x86-multi-thread, Binary build 813 [148120]

DBI: 1.48

ESQL: IBM Informix CSDK Version 2.90, IBM Informix-ESQL Version 2.90.TC1

Microsoft Visual C++ 6.0

Connection to Informix DB server 9.21.FC4 on an Alpha box running Compaq Tru64 Unix.

 

My Relevant Make output:

[output snipped from original message]

Brian Campbell (E-mail).vcf
Description: Binary data