From: Rick M. <obj...@gm...> - 2008-03-02 16:00:31
|
I've got to make a number of changes to the rexxutil code to clean it up for the new package-loader. This seems like an excellent time to split out the functions that are not truly system dependent. I'd like to make rexxutil a dll that's built from a set of common extension functions (e.g. SysAddRexxMacro) linked with a functions that do require a platform-specific implementation (most of them, really). The Windows version of rexxutil implements a set of math functions (SysPi(), etc.) that are not included in the unix version of rexxutil. These functions largely duplicate the functions implemented by the RxMath package, so there's not an overwhelming need for these to be added to the unix rexxutil. On the other hand, there's nothing at all platform-specific in these functions, so it would be a trivial matter to move these over into the common code module. One part of me says it would be good to make things more common between the two implementations. Another part of me says these functions should be deprecated in favor of the RxMath versions, so we shouldn't propagate these to other platforms. I'm interested in hearing how everybody else feels. If the consensus appears split, I might ask for a formal vote on this. Rick |
From: Mike C. <MF...@uk...> - 2008-03-02 17:07:00
|
Re math functions ... the new 754r standard will be done soon, and includes math functions and their special cases. Might be good to have a new library (perhaps a simple mod of RxMath) that conforms to that. A chance for a fresh start/cleaner match to ooRexx. Mike - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Mike Cowlishaw, IBM Fellow IBM UK (MP8), PO Box 31, Birmingham Road, Warwick, CV34 5JL mailto:mf...@uk... -- http://www2.hursley.ibm.com/mfcsumm.html "Rick McGuire" <obj...@gm...> Sent by: oor...@li... 02/03/2008 16:00 Please respond to Open Object Rexx Developer Mailing List <oor...@li...> To "Open Object Rexx Developer Mailing List" <oor...@li...> cc Subject [Oorexx-devel] Reworking rexxutil a little... I've got to make a number of changes to the rexxutil code to clean it up for the new package-loader. This seems like an excellent time to split out the functions that are not truly system dependent. I'd like to make rexxutil a dll that's built from a set of common extension functions (e.g. SysAddRexxMacro) linked with a functions that do require a platform-specific implementation (most of them, really). The Windows version of rexxutil implements a set of math functions (SysPi(), etc.) that are not included in the unix version of rexxutil. These functions largely duplicate the functions implemented by the RxMath package, so there's not an overwhelming need for these to be added to the unix rexxutil. On the other hand, there's nothing at all platform-specific in these functions, so it would be a trivial matter to move these over into the common code module. One part of me says it would be good to make things more common between the two implementations. Another part of me says these functions should be deprecated in favor of the RxMath versions, so we shouldn't propagate these to other platforms. I'm interested in hearing how everybody else feels. If the consensus appears split, I might ask for a formal vote on this. Rick ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Oorexx-devel mailing list Oor...@li... https://lists.sourceforge.net/lists/listinfo/oorexx-devel Unless stated otherwise above: IBM United Kingdom Limited - Registered in England and Wales with number 741598. Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU |
From: Mark M. <mie...@gm...> - 2008-03-02 18:39:10
|
Rick, I think it would be better to remove the math functions from RexxUtil and have them just in one place, the RxMath package. Mike's idea for a fresh start/ cleaner match sounds good to me. Of course, I'm sure there will be some complaint from someone whose program(s) rely on the functions being in RexxUtil. I would prefer to try and convince those people that we need to be able to move forward - not be tied too tightly to decisions made in the previous century. <grin> -- Mark Miesfeld On Sun, Mar 2, 2008 at 8:00 AM, Rick McGuire <obj...@gm...> wrote: > I've got to make a number of changes to the rexxutil code to clean it up for > the new package-loader. This seems like an excellent time to split out the > functions that are not truly system dependent. I'd like to make rexxutil a > dll that's built from a set of common extension functions (e.g. > SysAddRexxMacro) linked with a functions that do require a platform-specific > implementation (most of them, really). The Windows version of rexxutil > implements a set of math functions (SysPi(), etc.) that are not included in > the unix version of rexxutil. These functions largely duplicate the > functions implemented by the RxMath package, so there's not an overwhelming > need for these to be added to the unix rexxutil. On the other hand, there's > nothing at all platform-specific in these functions, so it would be a > trivial matter to move these over into the common code module. > > One part of me says it would be good to make things more common between the > two implementations. Another part of me says these functions should be > deprecated in favor of the RxMath versions, so we shouldn't propagate these > to other platforms. I'm interested in hearing how everybody else feels. If > the consensus appears split, I might ask for a formal vote on this. > > Rick > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Oorexx-devel mailing list > Oor...@li... > https://lists.sourceforge.net/lists/listinfo/oorexx-devel > > |
From: Rick M. <obj...@gm...> - 2008-03-02 18:55:33
|
I'm not sure we can get away with just removing them. There's little cost in keeping them in. I will take that as a vote in favor of keeping them a Windows exclusive though. Rick On Sun, Mar 2, 2008 at 1:37 PM, Mark Miesfeld <mie...@gm...> wrote: > Rick, > > I think it would be better to remove the math functions from RexxUtil > and have them just in one place, the RxMath package. > > Mike's idea for a fresh start/ cleaner match sounds good to me. > > Of course, I'm sure there will be some complaint from someone whose > program(s) rely on the functions being in RexxUtil. I would prefer to > try and convince those people that we need to be able to move forward > - not be tied too tightly to decisions made in the previous century. > <grin> > > -- > Mark Miesfeld > > On Sun, Mar 2, 2008 at 8:00 AM, Rick McGuire <obj...@gm...> > wrote: > > I've got to make a number of changes to the rexxutil code to clean it up > for > > the new package-loader. This seems like an excellent time to split out > the > > functions that are not truly system dependent. I'd like to make > rexxutil a > > dll that's built from a set of common extension functions (e.g. > > SysAddRexxMacro) linked with a functions that do require a > platform-specific > > implementation (most of them, really). The Windows version of rexxutil > > implements a set of math functions (SysPi(), etc.) that are not included > in > > the unix version of rexxutil. These functions largely duplicate the > > functions implemented by the RxMath package, so there's not an > overwhelming > > need for these to be added to the unix rexxutil. On the other hand, > there's > > nothing at all platform-specific in these functions, so it would be a > > trivial matter to move these over into the common code module. > > > > One part of me says it would be good to make things more common between > the > > two implementations. Another part of me says these functions should be > > deprecated in favor of the RxMath versions, so we shouldn't propagate > these > > to other platforms. I'm interested in hearing how everybody else feels. > If > > the consensus appears split, I might ask for a formal vote on this. > > > > Rick > > > > > ------------------------------------------------------------------------- > > This SF.net email is sponsored by: Microsoft > > Defy all challenges. Microsoft(R) Visual Studio 2008. > > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > > _______________________________________________ > > Oorexx-devel mailing list > > Oor...@li... > > https://lists.sourceforge.net/lists/listinfo/oorexx-devel > > > > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Oorexx-devel mailing list > Oor...@li... > https://lists.sourceforge.net/lists/listinfo/oorexx-devel > |
From: Mark M. <mie...@gm...> - 2008-03-02 21:54:31
|
On Sun, Mar 2, 2008 at 10:55 AM, Rick McGuire <obj...@gm...> wrote: > I'm not sure we can get away with just removing them. There's little cost > in keeping them in. I will take that as a vote in favor of keeping them a > Windows exclusive though. Okay, I re-read your first post a little closer. > One part of me says it would be good to make things more common between the > two implementations. Another part of me says these functions should be > deprecated in favor of the RxMath versions, so we shouldn't propagate these > to other platforms. So, I guess I have two parts to this too. 1.) I really would like to see the functions in RexxUtil work on all the platforms, at least as a goal. If the (Windows only) math functions are destined to *forever* remain in RexxUtil, then I think moving them to a common module so that they are not Windows exclusive is preferable. 2.) However, if deprecated means that at some point in the foreseeable future they can and will be removed, then yeah - don't move them to unix. Let's mark them as deprecated. And by foreseeable future I don't necessary mean near future. I just mean that deprecated functions will eventually be removed. I don't see any point in saying something is deprecated, but then turning around and never removing it. -- Mark Miesfeld |
From: Rick M. <obj...@gm...> - 2008-03-02 22:14:54
|
You've done a very good job of summarizing my position....and I'm pretty evenly split between option 1) and 2). I think if we believe we really can deprecate this functions and make a statement as to when they will be removed, I'd prefer option 2. Rick On Sun, Mar 2, 2008 at 4:54 PM, Mark Miesfeld <mie...@gm...> wrote: > On Sun, Mar 2, 2008 at 10:55 AM, Rick McGuire <obj...@gm...> > wrote: > > > I'm not sure we can get away with just removing them. There's little > cost > > in keeping them in. I will take that as a vote in favor of keeping them > a > > Windows exclusive though. > > Okay, I re-read your first post a little closer. > > > One part of me says it would be good to make things more common between > the > > two implementations. Another part of me says these functions should be > > deprecated in favor of the RxMath versions, so we shouldn't propagate > these > > to other platforms. > > So, I guess I have two parts to this too. > > 1.) I really would like to see the functions in RexxUtil work on all > the platforms, at least as a goal. If the (Windows only) math > functions are destined to *forever* remain in RexxUtil, then I think > moving them to a common module so that they are not Windows exclusive > is preferable. > > 2.) However, if deprecated means that at some point in the foreseeable > future they can and will be removed, then yeah - don't move them to > unix. Let's mark them as deprecated. And by foreseeable future I > don't necessary mean near future. I just mean that deprecated > functions will eventually be removed. I don't see any point in saying > something is deprecated, but then turning around and never removing > it. > > -- > Mark Miesfeld > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Oorexx-devel mailing list > Oor...@li... > https://lists.sourceforge.net/lists/listinfo/oorexx-devel > |
From: Mike C. <MF...@uk...> - 2008-03-03 08:22:01
|
> You've done a very good job of summarizing my position....and I'm > pretty evenly split between option 1) and 2). I think if we believe > we really can deprecate this functions and make a statement as to > when they will be removed, I'd prefer option 2. There's a part-way interpretation of 'deprecated' -- that at the designated time they are removed from the documentation but remain in the code. That way there's no burden on new users (and smaller documentation) but no 'ancient' code breaks. Mike Unless stated otherwise above: IBM United Kingdom Limited - Registered in England and Wales with number 741598. Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU |
From: Rick M. <obj...@gm...> - 2008-03-03 10:41:59
|
Hmmm, It would appear we've already achieved that level of deprecation! None of the rexxutil math functions are in the docs. This sent me back to the original IBM code drop to see if these had been accidentally enabled during the transition. No, the code was there, but these were never documented. Very curious indeed. Rick On Mon, Mar 3, 2008 at 3:21 AM, Mike Cowlishaw <MF...@uk...> wrote: > > You've done a very good job of summarizing my position....and I'm > > pretty evenly split between option 1) and 2). I think if we believe > > we really can deprecate this functions and make a statement as to > > when they will be removed, I'd prefer option 2. > > There's a part-way interpretation of 'deprecated' -- that at the > designated time they are removed from the documentation but remain in the > code. That way there's no burden on new users (and smaller documentation) > but no 'ancient' code breaks. > > Mike > > > > > > Unless stated otherwise above: > IBM United Kingdom Limited - Registered in England and Wales with number > 741598. > Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU > > > > > > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Oorexx-devel mailing list > Oor...@li... > https://lists.sourceforge.net/lists/listinfo/oorexx-devel > |
From: Mike C. <MF...@uk...> - 2008-03-03 11:02:43
|
> Hmmm, It would appear we've already achieved that level of > deprecation! None of the rexxutil math functions are in the docs. > This sent me back to the original IBM code drop to see if these had > been accidentally enabled during the transition. No, the code was > there, but these were never documented. Very curious indeed. Indeed. Forgotten, but not yet gone. :-)) Mike Unless stated otherwise above: IBM United Kingdom Limited - Registered in England and Wales with number 741598. Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU |
From: Gil B. <gi...@be...> - 2008-03-03 13:33:19
|
It seems to me that since they aren't/weren't (ever) documented, then no one could be dependent on them! I'd leave them in the source but commented out so they could not get accidently invoked. Mike Cowlishaw wrote: >>Hmmm, It would appear we've already achieved that level of >>deprecation! None of the rexxutil math functions are in the docs. >>This sent me back to the original IBM code drop to see if these had >>been accidentally enabled during the transition. No, the code was >>there, but these were never documented. Very curious indeed. > > > Indeed. Forgotten, but not yet gone. :-)) > > Mike > > > > > > Unless stated otherwise above: > IBM United Kingdom Limited - Registered in England and Wales with number > 741598. > Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU > > > > > > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Oorexx-devel mailing list > Oor...@li... > https://lists.sourceforge.net/lists/listinfo/oorexx-devel > > -- Gil Barmwater |
From: Rick M. <obj...@gm...> - 2008-03-03 14:15:49
|
Oh wonderful. A little googling has revealed that someone has ported those sections back into a close of rexxutil for OS/2 :-) Rick On Mon, Mar 3, 2008 at 9:03 AM, Mike Cowlishaw <MF...@uk...> wrote: > > It seems to me that since they aren't/weren't (ever) documented, > > How do you know that? :-) Or maybe they were documented in OS/2 > days and people who used that have been happily continuing to use > them in later versions, never knowing they weren't documented. (Chip?) > > > then no > > one could be dependent on them! I'd leave them in the source but > > commented out so they could not get accidentally invoked. > > I'd at least google for them, first. > > mfc > > > > > > Unless stated otherwise above: > IBM United Kingdom Limited - Registered in England and Wales with number > 741598. > Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU > > > > > > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Oorexx-devel mailing list > Oor...@li... > https://lists.sourceforge.net/lists/listinfo/oorexx-devel > |
From: Mike C. <MF...@uk...> - 2008-03-03 15:14:22
|
> Oh wonderful. A little googling has revealed that someone has > ported those sections back into a close of rexxutil for OS/2 :-) :-)) Once the genie is out of the bottle ... Unless stated otherwise above: IBM United Kingdom Limited - Registered in England and Wales with number 741598. Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU |
From: Mike C. <MF...@uk...> - 2008-03-03 14:03:17
|
> It seems to me that since they aren't/weren't (ever) documented, How do you know that? :-) Or maybe they were documented in OS/2 days and people who used that have been happily continuing to use them in later versions, never knowing they weren't documented. (Chip?) > then no > one could be dependent on them! I'd leave them in the source but > commented out so they could not get accidentally invoked. I'd at least google for them, first. mfc Unless stated otherwise above: IBM United Kingdom Limited - Registered in England and Wales with number 741598. Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU |