Re: REXX problem
- From: Marty <net.comcast@martyamodeo>
- Date: Thu, 20 Apr 2006 20:26:27 -0700
Paul Ratcliffe wrote:
On Mon, 17 Apr 2006 19:43:47 -0700, Marty <net.comcast@martyamodeo> wrote:
You can try that by yourself. Write 2 procedures loading sysloadfuncs,
let one use one of them in a loop and the other unload it - try to find a situation where the other can use the function after SysDropFuncs has run by the other.
Yup. I know. I've been bitten by it first-hand and lost many hours and hairs trying to figure it out. I was just commenting on how puzzling it is that they implemented things this way, given that DLLs operate in a much more robust way. It almost seems like they would have had to go out of there way to make it work the way it does today.
It's nothing to do with how DLLs work. It's to do with matching procedure
names to DLL names and entrypoints - a name registration issue if you like.
They should have put a usage count on registrations and deregistrations,
but they didn't for whatever reason. It would have made the code marginally
more difficult to write, but would have saved all the confusion...
That makes more sense now, thanks. It probably does use the normal referencing technique to actually load and unload the DLLs, but registering the procedure names in the language is the all or nothing approach.
--
[Reverse the parts of the e-mail address to reply.]
.
- References:
- REXX problem
- From: Fred Blau
- Re: REXX problem
- From: Doug Bissett
- Re: REXX problem
- From: James J. Weinkam
- Re: REXX problem
- From: Marty
- Re: REXX problem
- From: Herbert Rosenau
- Re: REXX problem
- From: Marty
- Re: REXX problem
- From: Paul Ratcliffe
- REXX problem
- Prev by Date: Re: Internet Adventurer
- Next by Date: Re: Clear All
- Previous by thread: Re: REXX problem
- Next by thread: Re: REXX problem
- Index(es):
Relevant Pages
|
Loading