The REXX language doesn't support callback; the API does. When an application calls IRXINIT, each of MODNAMET, SUBCOMTB and PACKTB can specify call-back routines.
-- Shmuel (Seymour J.) Metz http://mason.gmu.edu/~smetz3 ________________________________________ From: IBM Mainframe Discussion List [[email protected]] on behalf of David Crayford [[email protected]] Sent: Thursday, June 24, 2021 9:45 PM To: [email protected] Subject: Re: Coding for the future On 24/06/2021 9:33 pm, Seymour J Metz wrote: > That wiki article describes exactly the mechanism I was referring to. That's interesting. REXX doesn't support functions as first class objects so how can it support callbacks? I can image some monstrous concoction using "interpret". Maybe there are gaps in my REXX knowledge. > > > -- > Shmuel (Seymour J.) Metz > http://mason.gmu.edu/~smetz3 > > ________________________________________ > From: IBM Mainframe Discussion List [[email protected]] on behalf of > David Crayford [[email protected]] > Sent: Thursday, June 24, 2021 9:29 AM > To: [email protected] > Subject: Re: Coding for the future > > On 24/06/2021 8:51 pm, Seymour J Metz wrote: >> The only meaning that I'm familiar with is the one that I get on a search >> for 'callback definition programming'. Or were you thinking of a >> non-computer contest? > Callback is is a well known pattern > https://secure-web.cisco.com/1QSdJn1tHFo3fLnTHXKVuzco3ZyGmLW-vJIU-4BgBoJ6fnHEdkkNxvL7PgoLaBvpo0I8lemVuZ75OvusonUGyzxr1CV5i0Ww9nI1WZn3XFUso-xHRUl4FLtHqIGQqw0Q_U0FyxioHkGZUHg_5CwjKzl7I0tSi7y6qWdIvgG-u6fxFAJDUmIM6B1NKqqUXSHjK-SbQJOj1LAnHqNpTKdZueBEe6xZ9T-X0sazYHQQzyDEkdghGd_v8hwKJ7IfhJ7j1z1QOgZGFloyGsbaWtcL4yQWVqdTdPB-maTaKnYn8_7KxGVM-0zYdM0dYs9fgNJrebgDY_x4uksThdeKPMJSSwPBBYhutBegj1RG3PDcP8QFP8fy0xl-4EbJYe09lNW4C3k_VT77137C2DdIGZNS34cvFYUaPMP7R3M8AsTnOYZIOUKJsFvX4d4YJ6TXIaneJ/https%3A%2F%2Fen.wikipedia.org%2Fwiki%2FCallback_%28computer_programming) > > The ,array class goes all the way back to OREXX in OS/2. 5.0 adds a > regex class, but it doesn't support captures, and I rely on them heav > > I must admit I don't really care > >> -- >> Shmuel (Seymour J.) Metz >> http://mason.gmu.edu/~smetz3 >> >> ________________________________________ >> From: IBM Mainframe Discussion List [[email protected]] on behalf of >> David Crayford [[email protected]] >> Sent: Thursday, June 24, 2021 5:00 AM >> To: [email protected] >> Subject: Re: Coding for the future >> >> On 24/06/2021 9:23 am, Seymour J Metz wrote: >>> You invoke ISPF. You call a REXX script for ISPF. The script does ADDRESS >>> ISPEXEC foo and ISPF gets control back. It's well documented in the TSO/E >>> REXX Reference, and other implementations have essentially the same API. >>> Basically, when a REXX-aware application creates an environment, REXX >>> associates context with the environment name; when a REXX script routes a >>> command to that environment, the environment block contains the context so >>> that the application can locate its control blocks. It's quite elegant, and >>> REXX does all of the heavy lifting. >> Yes, I'm familiar with the environment block as I have used IRXSUBCM to >> create host command environments. In my case I had to use CEEPIPI to >> create an LE environment for the C++ regex library. The environment >> block has a user context pointer which is very useful. I was confused by >> the term "call-back" which means something quite different to me. >> >> >>> On the flip side, prior to OREXX everything was a string; there was no type >>> checking, only value testing. Also, the exception handling is fragile >>> compared to PL/I, Python or Ruby. >> I noticed that ooRexx v5 has introduced array types which is long >> overdue. Stem variables are hash tables so performance is not great when >> iterating. >> >> >>> -- >>> Shmuel (Seymour J.) Metz >>> http://mason.gmu.edu/~smetz3 >>> >>> ________________________________________ >>> From: IBM Mainframe Discussion List [[email protected]] on behalf of >>> David Crayford [[email protected]] >>> Sent: Wednesday, June 23, 2021 8:57 PM >>> To: [email protected] >>> Subject: Re: Coding for the future >>> >>> Maybe I'm missing a subtlety here but can you demonstrate how REXX uses >>> callbacks on z/OS? I'm familar with command environments, I've written >>> one for REXX regex >>> https://secure-web.cisco.com/140q8fxdqURUGvmth2nK_lOrDPN09AtW0cBfgGnzMZm6MOvOho-HriWxdOrBSwiPotV2P6kacF9rpRwTORQsYeHqbOlnMFxHuUDYsjYnWlk0z0sR7CYhLEzIyd-DL6yuI6IRotLmLhmT3RbCQk9-VO0E6-NE3tyAdm1jgxYV9K6GosVOwG_tqv0j_Iv7TeRrCoUvhQ5ZNmnH2N9gR8CjU9xFydEZWlX1Y702G19rTTW9fV0qRkhU5kTGg16f16KMPIZbaB7Cqk0evdWJylMOkSMkIA5PECEmFeyiQXvrOBE2LDvGMC16AK6OZg5xdnzeWpSgm_g5F-nCgOQjteQ_82RUkquwXW_WqLZq0S0yXoz4PpyFEFJrpl2iZboK4FhuXZdrvui3BqiSrz0nKNJYIXUNRpVzfCH02p2xcc1Oqa0Lrndw3QSeal5ZFf3sRqGP-/https%3A%2F%2Fgithub.com%2Fdaveyc%2FRTK. >>> In Python, Lua etc you >>> would just create an instance of a class an communicate using message >>> passing like any OO language. >>> For example, to create and ISPF in Lua you instantiate and ISPF object >>> and then communicate with it by calling methods >>> https://secure-web.cisco.com/1qsEFCKQQ8_UfI8nvAPUHIHUMRsk8zVcv8tyCsWKzPB5x6Iqcinh6eFRTvMYUE_fu4U_RlstYoTVGhhME8x92IIYlUFy0AvsQ_osnJyM5U0ASEu77n43wTA2l6-3XA99vaRmgsICvmmQW_Wlc1RO5W5SsL6uXFqvGMNSV5d2Swy6V131ckAPNDw9a-BsfYVkO5DRvOINNh9iJZqVw466fpXNjZh1CsFmaMQcpIvnXTnndqgbQwZw4yzPuxvmWTf__4ssShTUeQnjYxO6gcW4YsvpHhQslzGOP7V45e2SIHjYK3A2EGLp3hyelRfJG9gh3h0qFBX_1a5X08xba0wB0HwMXf4VyHIzAvI5LKoqvrMA90FsZq4h6TMxefXsiXKHALdxZV9djrJXqln73WhSCoaNFiVgG9C-wOFEwYW7bB12UrfhpL6v1ZJ3Hxudjq_4kHxQJuH7g6y4Brb90haU3vA/https%3A%2F%2Flua4z.github.io%2FLua4z%2Fmodules%2Fispf.html. >>> >>> >>> On 23/06/2021 7:04 pm, Seymour J Metz wrote: >>>> Allow applications to establish environments for scripts (ADDRESS) , allow >>>> call-backs from within scripts and access the variables of the scripts. I >>>> wish that I could do that from within Perl, although I'd take the time to >>>> learn Python or Ruby if they supported that. >>>> >>>> >>>> -- >>>> Shmuel (Seymour J.) Metz >>>> http://mason.gmu.edu/~smetz3 >>>> >>>> ________________________________________ >>>> From: IBM Mainframe Discussion List [[email protected]] on behalf >>>> of David Crayford [[email protected]] >>>> Sent: Wednesday, June 23, 2021 3:19 AM >>>> To: [email protected] >>>> Subject: Re: Coding for the future >>>> >>>> On 23/06/2021 10:38 am, Seymour J Metz wrote: >>>>> None of those are relevant to what REXX does best. Again, it's the >>>>> plumbing, not the syntactic sugar. >>>> What exactly does REXX do best? What does it do better than Python? >>>> >>>> >>>>> As to large applications, I believe that there are some pretty big ones >>>>> in Perl and REXX, so why should I be surprised by large applications in, >>>>> e.g., Python, Ruby? >>>>> >>>>> >>>>> -- >>>>> Shmuel (Seymour J.) Metz >>>>> http://mason.gmu.edu/~smetz3 >>>>> >>>>> ________________________________________ >>>>> From: IBM Mainframe Discussion List [[email protected]] on behalf >>>>> of David Crayford [[email protected]] >>>>> Sent: Tuesday, June 22, 2021 7:25 PM >>>>> To: [email protected] >>>>> Subject: Re: Coding for the future >>>>> >>>>> On 22/06/2021 10:19 pm, Seymour J Metz wrote: >>>>>> It's not a question of what environment it can run in; it's a question >>>>>> of what facilities it supports in those environments. >>>>> Indeed. Lua supports TSO, ISPF, CICS, IMS, the full MVS file system >>>>> including VSAM. DB2 SQL + a SQLite API. A date/time module that supports >>>>> dates that are greater than Mon Jan 01 1000000 BCE 00:00:00 and less >>>>> than Mon Jan 01 1000001 00:00:00 etc, etc. >>>>> We presented at SHARE where we demonstrated how to use Lua integration >>>>> to convert a 3270 CICS BMS application to a mobile web app >>>>> https://secure-web.cisco.com/1LTRemd3KBmC5tUUKSeo9EC5h9xyRF7Bz3PhdCSMXFnXeU9xUYZAMVe5chb0ZaeK3DMtRGm6fXidrWX-DdewRuiBFuunmJgylrBSK2m0dxTkBwteMqPfIErfpIkx2t0yKE_X4Vvjo2z5bwL3x5Gzol57ueWGXt0mRa5LkbagnGSXpBSeh0a0sqCHTv7dZ_yP7bvWRZEdnEhIKqF6yOijhnDFn7l75N3TqnoUwxHNnsOFWy_ZGihMqSRgOpFgNGT29c3jEMR9BWE6SwalS5Fq9qCm5mI3QT3QGXbhUjO6A4bL_zu9qSlFAFZM7PJeBTr73ngdhubfDFqh0fRWaFg5cyBhzcMGUTVbEV5QWTr76UNB_zRihnfVeyioclr2jou2HVmP2A7PlvHHZDFG0y1o-fR5mEjE0DxE6Jrzclf2Oy-W2TISt9m1rarhndnanZuca/https%3A%2F%2Fshare.confex.com%2Fshare%2F123%2Fwebprogram%2FHandout%2FSession15892%2F15892%2520Mobile%2520CICS%2520Application%2520Using%2520Lua.pdf. >>>>> >>>>> I spent a lot of time making sure Lua integrated well into the TSO/MVS >>>>> environment. The irony is I spend very little time in TSO/ISPF these >>>>> days. I spend most of my time in a shell and switch the ISPF when I want >>>>> to use SDSF. I write new scripting code in Python which is enhanced >>>>> ASCII so only >>>>> runs in z/OS UNIX. Rocket and IBM don't see any value in integrating >>>>> ported tools with TSO as it's not strategic. The main focus is on >>>>> containers. >>>>> >>>>> BTW, did you know that the photo sharing app Instagram is written in >>>>> Python? It's staggering to imagine that what we would consider a >>>>> scripting language is serving over 1 billion monthly users. >>>>> >>>>> >>>>>> -- >>>>>> Shmuel (Seymour J.) Metz >>>>>> http://mason.gmu.edu/~smetz3 >>>>>> >>>>>> >>>>>> ________________________________________ >>>>>> From: IBM Mainframe Discussion List <[email protected]> on behalf >>>>>> of David Crayford <[email protected]> >>>>>> Sent: Tuesday, June 22, 2021 9:57 AM >>>>>> To: [email protected] >>>>>> Subject: Re: Coding for the future >>>>>> >>>>>> On 22/06/2021 8:55 pm, Seymour J Metz wrote: >>>>>>> What distinguishes REXX is not syntactic sugar but the plumbing that >>>>>>> enables close coupling of scripts with applications. Lua is missing >>>>>>> that. >>>>>> Maybe. But I can't think of an environment that REXX runs in that Lua >>>>>> can't. I could port Python to run in TSO if I had the time but it's a >>>>>> heavy lift. IBM have ported golang now which is compiled. We've had a >>>>>> look at it and it's interesting. >>>>>> >>>>>> ---------------------------------------------------------------------- >>>>>> For IBM-MAIN subscribe / signoff / archive access instructions, >>>>>> send email to [email protected] with the message: INFO IBM-MAIN >>>>>> >>>>>> ---------------------------------------------------------------------- >>>>>> For IBM-MAIN subscribe / signoff / archive access instructions, >>>>>> send email to [email protected] with the message: INFO IBM-MAIN >>>>> ---------------------------------------------------------------------- >>>>> For IBM-MAIN subscribe / signoff / archive access instructions, >>>>> send email to [email protected] with the message: INFO IBM-MAIN >>>>> >>>>> ---------------------------------------------------------------------- >>>>> For IBM-MAIN subscribe / signoff / archive access instructions, >>>>> send email to [email protected] with the message: INFO IBM-MAIN >>>> ---------------------------------------------------------------------- >>>> For IBM-MAIN subscribe / signoff / archive access instructions, >>>> send email to [email protected] with the message: INFO IBM-MAIN >>>> >>>> ---------------------------------------------------------------------- >>>> For IBM-MAIN subscribe / signoff / archive access instructions, >>>> send email to [email protected] with the message: INFO IBM-MAIN >>> ---------------------------------------------------------------------- >>> For IBM-MAIN subscribe / signoff / archive access instructions, >>> send email to [email protected] with the message: INFO IBM-MAIN >>> >>> ---------------------------------------------------------------------- >>> For IBM-MAIN subscribe / signoff / archive access instructions, >>> send email to [email protected] with the message: INFO IBM-MAIN >> ---------------------------------------------------------------------- >> For IBM-MAIN subscribe / signoff / archive access instructions, >> send email to [email protected] with the message: INFO IBM-MAIN >> >> ---------------------------------------------------------------------- >> For IBM-MAIN subscribe / signoff / archive access instructions, >> send email to [email protected] with the message: INFO IBM-MAIN > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
