From: Leyne, S. <sl...@at...> - 2001-03-02 18:31:55
|
All, A message in the Borland newsgroups prompted this question. Is there a Firebird port for AIX? I've have seen references to people building an AIX version, even seen something in the v0.9 release notes, but can't find anything listed in the Firebird - Files area. Sean |
From: Paul R. <pr...@ib...> - 2001-03-02 18:59:15
|
"Leyne, Sean" wrote: > > All, > > A message in the Borland newsgroups prompted this question. Is there a > Firebird port for AIX? > > I've have seen references to people building an AIX version, even seen > something in the v0.9 release notes, but can't find anything listed in > the Firebird - Files area. > It's there somewhere - try the ftp site. Paul -- Paul Reeves http://www.ibphoenix.com taking InterBase further |
From: Paul R. <pr...@ib...> - 2001-03-02 19:46:03
|
Paul Reeves wrote: > > "Leyne, Sean" wrote: > > > > All, > > > > A message in the Borland newsgroups prompted this question. Is there a > > Firebird port for AIX? > > > > I've have seen references to people building an AIX version, even seen > > something in the v0.9 release notes, but can't find anything listed in > > the Firebird - Files area. > > > > It's there somewhere - try the ftp site. > It should be here to be more precise: ftp://firebird.sourceforge.net/pub/firebird/FirebirdCS-0.9-4.AIX.bff.Z with useful notes here: ftp://firebird.sourceforge.net/pub/firebird/FirebirdCS-0.9-4.AIX.readme Paul -- Paul Reeves http://www.ibphoenix.com taking InterBase further |
From: Tom C. <tco...@au...> - 2001-03-02 19:02:43
|
> > A message in the Borland newsgroups prompted this question. Is there a > Firebird port for AIX? > There must indeed be a firebird port for AIX. The following is a summary of one developer's experience with Firebird/AIX. He's had some problems with the firebird mailing list, so he contacted me directly for some help with resolving problems with his UDF's. I asked him to summarize his experience since it might be helpful to others. I believe the 'rant' has to do with the fact that the expected location of UDF's changed between IB V5 and V6. There's something for both Firebird developers and AIX users here. Tom - - - - - ....... here is a brief synopsis of what I went through to get this stuff working. You can post this to the list as is or edited if you feel it is appropriate. [ Posted Unedited ] 1) Downloaded FirebirdCS-0.9-4.AIX.bff.Z and installed in. No problems. 2) Tried to compile my app using xlc. Everything compiles fine but can't link. 3) Find, download and install gcc for AIX. 4) Recompile using gcc. Now some of my embedded SQL apps won't compile because the source file is too big (I think). I get errors from the assembler about unacceptably large offsets. 5) Compile embedded SQL with xlc, everything else with gcc. Everything links except UDF. 6) Looked for but could not find an example of how to build a UDF for AIX. The stuff in the examples directory does not seem to have an AIX version. 7) Spent a day or so rummaging through the firebird makefiles to find out how those shared libraries are made. 8) Copied that stuff from the firebird makefile to mine and hey presto, the UDF links. Now I think I'm in good shape. Little did I know. 9) Install the UDF into /usr/interbase/lib just like I have always done for IB5. Used a DECLARE FUNCTION ... MODULE_NAME '/usr/interbase/lib/libudf.so' ... to define the function. 10) Test the UDF by issuing a SELECT and Interbase tells me that it can't find the entry point or module. 11) Spend a day trying various fruitless variations of the above and verifying that I had indeed built the UDF correctly. 12) Contacted you and found out that the shared library must be in the UDF directory. 13) Move the shared library and changed the DECLARE FUNCTION to remove the absolute path and now everything works. 14) Now that I know what the problem is, I go back and read the manual and find that this behaviour is actually documented. begin rant { I really wish that software designers would design in new features without breaking code that has worked for years. It is all well and good to improve the product by adding new features, but deliberately breaking existing code is unacceptable. Simply telling people that it is broken by burying a small reference to the fact deep in the manual is also not acceptable. Speaking just for myself I don't have the time or the inclination to re-read an entire manual set every time I get a new release. It also seems bizarre that sticking an absolute path in front of the library name doesn't override the default setting. All this ranting is not directed at the firebird developers specifically since I know that this stuff was broken before it was open sourced, but I believe that this is something we should all bear in mind before we start hacking in new features. } end rant |
From: reed m. <rf...@cr...> - 2001-03-02 22:25:22
|
Tom Coleman wrote: > > > > > A message in the Borland newsgroups prompted this question. Is there a > > Firebird port for AIX? > > > > There must indeed be a firebird port for AIX. > > The following is a summary of one developer's experience with Firebird/AIX. > He's had some problems with the firebird mailing list, so he contacted me > directly for some help with resolving problems with his UDF's. > > I asked him to summarize his experience since it might be helpful to others. > > I believe the 'rant' has to do with the fact that the expected location of > UDF's changed between IB V5 and V6. If you explain to the ranter the implications of any user being able to define the 'system' function from the C runtime library as a UDF (as well as ANY other function of ANY shared library on the system), perhaps he might be less upset that the UDF directory is now restricted. This is not simply changing old behaviour for the sake of it, but fixing a vary serious bug. If he really wants the old behaviour back, he can add /usr/lib to the UDF path. This should be mentioned in our readme/release notes as a change from previous versions. (* reed goes back to lurking *) [...] > begin rant { > > I really wish that software designers would design in new features without > breaking code that has worked for years. It is all well and good to > improve the product by adding new features, but deliberately breaking > existing code is unacceptable. Simply telling people that it is broken by > burying a small reference to the fact deep in the manual is also not > acceptable. Speaking just for myself I don't have the time or the > inclination to re-read an entire manual set every time I get a new release. > > It also seems bizarre that sticking an absolute path in front of the > library name doesn't override the default setting. > > All this ranting is not directed at the firebird developers specifically > since I know that this stuff was broken before it was open sourced, but I > believe that this is something we should all bear in mind before we start > hacking in new features. > > } end rant > > -- Reed Mideke email: rfm(at)cruzers.com -If that fails: rfm(at)portalofevil.com |