Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

from." -Chip- Original message From: Jason Martin Date: 2/16/19 13:50 (GMT-05:00) To: Enrico Sorichetti via Oorexx-devel Subject: Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766 I know, knew, just saying even standa

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

The problem is that there is no automatic way to find out to such detail Just run … cmake --system-information [file] = Dump information about this system. To see what cmake can find out about Your system E > On 16 Feb 2019, at 21:07, Gil Barmwater wrote: > > I was just suggesting that if

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

Let me try again.  When a user wants to install ooRexx on his platform, he goes to SF and looks in the Files section under the ooRexx tab for the most recent file with a name that matches his platform, downloads it and uses the appropriate installation mechanism for that platform.  I was just

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

Funny … after setting a missing CPACK variable Before ooRexx-5.0.0-11767.x86_64.deb The name had a reference to the svn revision After ooRexx-5.0.0-Linux.deb Nothing Weird ??? E ___ Oorexx-devel mailing list Oorexx-devel@lists.sourceforge.net

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

something built on a certain a platform WILL NOT run on a different one LINUX will be even more selective There are subtle differences between the different distributions And there are also common practices to deal with … xx.deb is likely a package for Debian family system x.rpm

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

That’s only the platform it was built on, not necessarily the one it’s running on. On Sat, Feb 16, 2019 at 2:14 PM Gil Barmwater wrote: > OK, so CMAKE "knows" the PLATFORM_ID which, I'm guessing, is used to > generate the name of the ooRexx installation file; e.g. >

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

I guess not … From my tests they create something . ooRexx-5.0.0-Linux.sh ooRexx-5.0.0-Linux.tar.Z ooRexx-5.0.0-Linux.tar.gz I did it on Fedora but the result is exactly the same for Debian For the other tests I run I got the same layout with Darwin, FreeBSD, OpenBSD, instead of Linux Just

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

OK, so CMAKE "knows" the PLATFORM_ID which, I'm guessing, is used to generate the name of the ooRexx installation file; e.g. ooRexx-5.0.0-11742.centos7.x86_64.rpm . Would it make

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

I know, knew, just saying even standards are not standard. ___ Oorexx-devel mailing list Oorexx-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/oorexx-devel

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

Cmake should return “sunOS” From the compiler builtin macro #elif defined(__sun) || defined(sun) # define PLATFORM_ID "SunOS" > On 16 Feb 2019, at 18:40, Jason Martin wrote: > > agrellum@openindiana:~$ uname -v > illumos-c78b1a4529 > > Not OpenIndiana and Not SunOS > > > > >

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

agrellum@openindiana:~$ uname -v illumos-c78b1a4529 Not OpenIndiana and Not SunOS ___ Oorexx-devel mailing list Oorexx-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/oorexx-devel

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

> Should maybe "PARSE SOURCE" return "UBUNTU" instead of "LINUX" to indicate > that it behaves > differently compared to other Linuxes? Parse source return the info provides by CMAKE_SYSTEM_NAME And by some tweak of configure.ac before that I just run a cmake --system-information info.txt

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

What would you suggest/advise given your huge knowledge and experiences on so many different Unix systems? Should the test be adapted to check whether Ubuntu is the host system (e.g. "if sysVersion()~caselessPos('Ubuntu')>0 then ...")? Should the function work differently on Ubuntu to match the

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

I started a bit earlier ( do You remember Slackware ) After that I went for Red Hat And got used to the red-hat and friends terminology I can agree on the Debian ways But Ubuntu has taken things a bit too far Apart being too much much windowish for my taste ;-) This morning it took me just

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

Enrico Sorichetti via Oorexx-devel wrote: Unfortunately Ubuntu is well known also for its odd way of doing things haha... The Debian / Ubuntu way makes sense, the other distros are the odd ones! ;-) Seriously... I started out with Red Hat 5.2 back in probably 1999. Yuck. SuSE, Mandrake...

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766 and testsuite at r11767

You should be more open minded and Spend more time to learn about the different Linux distributions behaviour I spent a couple of hours installing one more Linux distribution I choose Debian, much more trustworthy as a high lever developer platform … And guess what ??? Executing

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

Unfortunately Ubuntu is well known also for its odd way of doing things > On 16 Feb 2019, at 12:55, Rick McGuire wrote: > > It is Unbuntu's though. ___ Oorexx-devel mailing list Oorexx-devel@lists.sourceforge.net

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

It is Unbuntu's though. Rick On Sat, Feb 16, 2019 at 6:54 AM Enrico Sorichetti via Oorexx-devel < oorexx-devel@lists.sourceforge.net> wrote: > Well… that’ not Fedora Linux opinion > > > Interpreter:REXX-ooRexx_5.0.0(MT)_64-bit 6.05 16 Feb 2019 > OS Name:LINUX > SysVersion:

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

Well… that’ not Fedora Linux opinion Interpreter:REXX-ooRexx_5.0.0(MT)_64-bit 6.05 16 Feb 2019 OS Name:LINUX SysVersion: Linux 4.20.7-200.fc29.x86_64 Tests ran: 22641 Assertions: 377418 Failures: 1 Errors: 0 [failure]

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

It's not so simple. That test works as expected on Linux, it looks like it is only failing on Darwin. Rick On Sat, Feb 16, 2019 at 6:38 AM Enrico Sorichetti via Oorexx-devel < oorexx-devel@lists.sourceforge.net> wrote: > Certainly different from windows , > > On unix it behaves according to the

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

Certainly different from windows , On unix it behaves according to the docs deleteFile uses unlink which does not check the permissions on the file E Quoting the libc manual 14.6 Deleting Files <> <> <>You can delete a file with unlink or remove. Deletion actually deletes a file name. If

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

No, those are correct. The test is expecting that an attempt to delete a read-only file will fail. Just another one of those maddening differences between platforms. Rick On Sat, Feb 16, 2019 at 4:02 AM Enrico Sorichetti via Oorexx-devel < oorexx-devel@lists.sourceforge.net> wrote: > > Seems to

Re: [Oorexx-devel] Testing Open Object Rexx Version 5.0.0 r11766

Seems to work, But It looks like [failure] [20190216 09:49:03.638306] svn:r11734 Change date: 2019-02-08 15:41:41 -0500 Test: TEST_FILE_EXISTS Class: SysFileXXX.testgroup File: /Users/enrico/ooRexx.testSuite/.../base/rexxutil/SysFileXXX.testGroup Line: 116 Failed: