Routines - When must they be recompiled?

[ray.wurlod]

That's what i'm trying to prove to tgmedear.

DataStage DOES preserve case sensitivity. It's Windows that doesn't.

[tgmedear]

Ray and ArndW,
Thank you for your replies. Unfortunately, much of your replies are cut off, because you are premium posters. To me, this is a significant DataStage bug. If DataStage is able to store separate source code files for routines that are in different categories, then it should be able to store separate executables for routines that are in different categories. If it can not do that, then it should not allow one to store separate source code files for routines in different categories that happen to have the same name. I will take this up with DataStage service, as it was a tricky problem to detect and resolve.

Terry

[ray.wurlod]

Premium membership is only a few cents per day (corporate discounts are available - see link from home page) and helps to fund the bandwidth charges that DSXchange incurs.

Read my previous post again. This is the answer you will get from support. Then get yourself a less flaky operating system (any UNIX or Linux) in which this "bug" does not occur.

[tgmedear]

Ray... I do not control the purse strings, and going through all the hoops to get it done is frankly too much of a pain.
If you can just answer this, I will be happy. Why does DataStage allow me to store two different routines with, ignoring case sensitivity, the same name, in two different categories but not produce two different executables. It appears that DS is storing source files in separate directories and executables in a single directory... thus creating the problem for executables but not source files. That seems inconsistent to me. I've worked on UNIX and Windows/DOS for 20 years and I've never seen an application behave like this. If someone else thinks this is justifiable, please explain it to me.

[ray.wurlod]

The short answer is that UniVerse (the original underlying engine) was out and about well before Windows appeared. It's originally a UNIX-only application, where case-sensitivity is assured.

The only "fix" you're likely to get is a documentation fix extending the prohibition on duplicate routine names to include differently-cased duplicates.