Any news over Ruby 2.3 for windows?...

[Adrián Avila]

Any news of Ruby 2.3 for Windows? is it planned?

[Justin Baker]

Yeah, sorry. I've really been busy since mid-December. When I get time I work on it. But it has been really hard to find time.

Justin

[Sean Fausett]

+1 ;)

[Dan Rossiter]

+1

[Keith Smith]

What can I do to help?

Keith

[Hanlyu Sarang]

I remember almost two years ago, someone posted the url to directions for building the latest version of Ruby on Windows with the RubyInstaller code (for those of us who were impatient). I followed the directions and built my own local install of Ruby, and everything worked OK. There were a lot of steps, but everything worked as per the instructions. Note that I did not build the RubyInstaller executable/installer (which requires special tools), but rather a local install of Ruby (that worked with the RubyInstaller modifications). Unfortunately, I can't find the link.

I'm not planning on upgrading to Ruby 2.3 at this time, but I would like to build my own local version of Ruby 2.2.4, which has some security fixes.

If anyone recognizes the instructions I'm speaking of, could you please respond with a link. It might be possible to also build Ruby 2.3 in the same way?

Thanks.

On Sat, Jan 30, 2016 at 10:44 AM, Keith Smith <nzke...@gmail.com> wrote:
>
> What can I do to help?
>
> Keith
>

> --
> You received this message because you are subscribed to the Google Groups "RubyInstaller" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to rubyinstalle...@googlegroups.com.
> To post to this group, send email to rubyin...@googlegroups.com.
> Visit this group at https://groups.google.com/group/rubyinstaller.
> For more options, visit https://groups.google.com/d/optout.

[Justin Baker]

On Tuesday, January 19, 2016 at 4:27:54 PM UTC-6, Justin Baker wrote:

Yeah, sorry. I've really been busy since mid-December. When I get time I work on it. But it has been really hard to find time.

Justin

So, Just a heads up to everyone that I know has been anxiously waiting. I have literally sat down at my personal computer three times in the last week. My job no longer allows me to work with Ruby on Windows or use RubyInstaller and any time I have outside of that has been spent on personal matters. Things happens and I really want to get this out but the other side is I have more important things to take care of.

Most of the hard parts are done for the updates for the 2.0-2.2 versions, but there's a lot of manual things I haven't had time for. I want to automate those things, but that also takes time. So I'm trying I promise. =/

Justin

[Justin Baker]

On Saturday, January 30, 2016 at 11:40:58 AM UTC-6, Keith Smith wrote:

What can I do to help?

Keith

Hey Keith! That's an awesome question. I need to get release notes written and update the site for the Ruby 2.0-2.2 series.

Then I need to build Ruby 2.3, test it, write new release notes and update the site.

If you wanted to you could checkout this branch from windwiny on GitHub:

https://github.com/windwiny/rubyinstaller/tree/ruby23

Then you should be able to run `rake ruby23 DKVER=mingw64-32-4.7.2` and make sure it build Ruby 2.3.0 properly. I had a weird error earlier that I think was from another background task I was having to run, but I haven't been able to investigate.

Justin

[Justin Baker]

On Saturday, January 30, 2016 at 8:46:12 PM UTC-6, Hanlyu Sarang wrote:

I remember almost two years ago, someone posted the url to directions for building the latest version of Ruby on Windows with the RubyInstaller code (for those of us who were impatient). I followed the directions and built my own local install of Ruby, and everything worked OK. There were a lot of steps, but everything worked as per the instructions. Note that I did not build the RubyInstaller executable/installer (which requires special tools), but rather a local install of Ruby (that worked with the RubyInstaller modifications). Unfortunately, I can't find the link.

I'm not planning on upgrading to Ruby 2.3 at this time, but I would like to build my own local version of Ruby 2.2.4, which has some security fixes.

If anyone recognizes the instructions I'm speaking of, could you please respond with a link. It might be possible to also build Ruby 2.3 in the same way?

I can't say for sure I know of a link. But if you open `config/rubyinstaller.rb` and change the version number and download links then run, `rake ruby22 DKVER=mingw64-32-4.7.2` it will build Ruby for you from scratch. =)

Justin

[Hanlyu Sarang]

Thank you very much. I'll try it tomorrow.

--

[Keith Smith]

On Sunday, 31 January 2016 16:46:30 UTC+13, Justin Baker wrote:

... I need to get release notes written and update the site for the Ruby 2.0-2.2 series.

Then I need to build Ruby 2.3, test it, write new release notes and update the site.

If you wanted to you could checkout this branch from windwiny on GitHub:

https://github.com/windwiny/rubyinstaller/tree/ruby23

Then you should be able to run `rake ruby23 DKVER=mingw64-32-4.7.2` and make sure it build Ruby 2.3.0 properly. I had a weird error earlier that I think was from another background task I was having to run, but I haven't been able to investigate.

Justin, I've built Ruby 2.3.0. The only obvious error was: 

file2lastrev.rb: last revision not found
make: [.revision.time] Error 1 (ignored)
../ruby_2_3/revision.h unchanged

... which also appears in a comparable build of Ruby 2.2 so seems mostly harmless.

The output from 'rake check23 DKVER=mingw64-32-4.7.2' is less encouraging. A bunch of tests apparently succeed, but then rake aborts with:

make: *** [yes-test-almost] Error 5

That said I don't know what constitutes a good result from check23, because a comparable check22 against my Ruby 2.2 build also errors out.

How can I help from here? I'd offer to help with the release of the Ruby 2.0-2.2 series but I see that you've already done that. Are there tests other than 'check23' that need to be done? Release notes for 2.3? Let me know.

Regards

Keith

[Keith Smith]

Justin, attached are the outputs from my build and check of Ruby 2.3.

[Justin Baker]

Awesome! Thanks Keith!

More comments inline.

On Sunday, February 7, 2016 at 12:34:05 PM UTC-6, Keith Smith wrote:

On Sunday, 31 January 2016 16:46:30 UTC+13, Justin Baker wrote:

... I need to get release notes written and update the site for the Ruby 2.0-2.2 series.

Then I need to build Ruby 2.3, test it, write new release notes and update the site.

If you wanted to you could checkout this branch from windwiny on GitHub:

https://github.com/windwiny/rubyinstaller/tree/ruby23

Then you should be able to run `rake ruby23 DKVER=mingw64-32-4.7.2` and make sure it build Ruby 2.3.0 properly. I had a weird error earlier that I think was from another background task I was having to run, but I haven't been able to investigate.

Justin, I've built Ruby 2.3.0. The only obvious error was: 

file2lastrev.rb: last revision not found
make: [.revision.time] Error 1 (ignored)
../ruby_2_3/revision.h unchanged

... which also appears in a comparable build of Ruby 2.2 so seems mostly harmless.

Honestly most of these bugs should probably have tickets on the Ruby bug tracker, but I haven't filed anything because I am lazy. =/

The output from 'rake check23 DKVER=mingw64-32-4.7.2' is less encouraging. A bunch of tests apparently succeed, but then rake aborts with:

make: *** [yes-test-almost] Error 5

That said I don't know what constitutes a good result from check23, because a comparable check22 against my Ruby 2.2 build also errors out.

How can I help from here? I'd offer to help with the release of the Ruby 2.0-2.2 series but I see that you've already done that. Are there tests other than 'check23' that need to be done? Release notes for 2.3? Let me know.

The process I actually use to test is `rake ruby23:sh` which drops into a `msys shell`.

Then `cd sandbox/ruby23_build; make test-all`

Which gives output on the tests that are failing.

More importantly I've got something weird going on with my build environment that is intermittent and I need to figure out what.

There's some stuff going on with `rdoc_chm` which there's a quickfix patch that I'm going to use for this release but next release `rdoc_chm` needs to be updated to use rdoc 4+.

Release Notes are copy and paste more or less. Add a disclaimer about how most native gems that are distributed as fat binaries won't work until a new version is released.

But more importantly, a specific version of InnoSetup has be installed as well as Microsoft's HTMLHelp compiler. Then the `chm` files and `exe` installers have to be built and I think `rdoc` will currently throw an error parsing Ruby 2.3 because we are locked to an old version of `rdoc` for `rdoc_chm`.

There are probably other things, but I can't name them off the top of my head. The process probably could be simplified a bit but for now it works well enough. =/

Justin

[Keith Smith]

Hi again Justin. I found some more time to work on this.

The summary output of (cd sandbox/ruby23_build; make test-all):

15556 tests, 2217645 assertions, 6 failures, 3 errors, 168 skips

For now I don't know enough to say which of the failures/errors are serious, so I compared the test result with that from a ruby22 build of the latest in https://github.com/oneclick/rubyinstaller.git ...

  1) Failure: [Keith: Existing failure in ruby22]
TestDir_M17N#test_entries_compose [d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/ruby/test_dir_m17n.rb:396]:
[ruby-core:48745] [Bug #7267].
<["d\xE9tente"(Windows-1252), "x?????"(UTF-8)]> expected but was
<["d\x82tente"(IBM437), "x?????"(IBM437)]>.

  2) Failure: [Keith: New failure]
TestFile#test_file_share_delete [d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/ruby/test_file.rb:418]:
Exception raised:
<#<Errno::EACCES: Permission denied @ unlink_internal - C:/Users/keith/AppData/Local/Temp/test_file_share_delete20160213-26592-1qcjzrd/x>>.

  3) Failure: [Keith: New failure]
TestWIN32OLE_EVENT_SWbemSink#test_on_event [d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/win32ole/test_win32ole_event.rb:112]:
Expected /OnObjectReady/ to match "".

  4) Failure: [Keith: Existing failure in ruby22]
TestFile#test_utime [d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/ruby/test_file.rb:341]:
[ruby-core:44776].
<2011-05-28 08:10:39 +1200> expected but was
<2011-05-28 09:10:39 +1200>.

  5) Failure: [Keith: Existing failure in ruby22? (Assertion failure reporting is different but it appears to be the same issue.)]
TestRubyOptions#test_command_line_progname_nonascii [d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/ruby/test_rubyoptions.rb:709]:
[ruby-dev:48752] [Bug #10555].

1. [1/2] Assertion for "stdout"
   | <["\xFF.rb"]> expected but was
   | <[]>.

2. [2/2] Assertion for "stderr"
   | <[]> expected but was
   | <["d:\\Build\\rubyinstaller-ruby23\\sandbox\\ruby23_build\\ruby.exe: No such file or directory -- .rb (LoadError)"]>..

  6) Failure: [Keith: Existing failure in ruby22]
TestRDocOptions#test_parse_write_options [d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/rdoc/test_rdoc_options.rb:640]:
[SystemExit] exception expected, not
Class: <ArgumentError>
Message: <"different prefix: \"d:/\" and \"C:/Users/keith/AppData/Local/Temp/test_rdoc_options_26592\"">
---Backtrace---
d:/Build/rubyinstaller-ruby23/sandbox/ruby23_build/.ext/common/pathname.rb:520:in `relative_path_from'
d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:1157:in `block in sanitize_path'
d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:1155:in `reject'
d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:1155:in `sanitize_path'
d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:494:in `encode_with'
d:/Build/rubyinstaller-ruby23/sandbox/ruby23_build/.ext/common/psych/visitors/yaml_tree.rb:562:in `dump_coder'
d:/Build/rubyinstaller-ruby23/sandbox/ruby23_build/.ext/common/psych/visitors/yaml_tree.rb:161:in `accept'
d:/Build/rubyinstaller-ruby23/sandbox/ruby23_build/.ext/common/psych/visitors/yaml_tree.rb:127:in `push'
d:/Build/rubyinstaller-ruby23/sandbox/ruby23_build/.ext/common/psych.rb:416:in `dump'
d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:1247:in `block in write_options'
d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/open-uri.rb:37:in `open'
d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/open-uri.rb:37:in `open'
d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:1244:in `write_options'
d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:1127:in `parse'
d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/rdoc/test_rdoc_options.rb:641:in `block (2 levels) in test_parse_write_options'
---------------

  7) Error: [Keith: Existing error in ruby22]
TestRDocOptions#test_encode_with_trim_paths:
ArgumentError: different prefix: "d:/" and "C:/Users/keith/AppData/Local/Temp/d20160213-26592-ra5v7c/project"
    d:/Build/rubyinstaller-ruby23/sandbox/ruby23_build/.ext/common/pathname.rb:520:in `relative_path_from'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:1157:in `block in sanitize_path'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:1155:in `reject'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:1155:in `sanitize_path'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:494:in `encode_with'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/rdoc/test_rdoc_options.rb:115:in `block (2 levels) in test_encode_with_trim_paths'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/rdoc/test_rdoc_options.rb:101:in `chdir'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/rdoc/test_rdoc_options.rb:101:in `block in test_encode_with_trim_paths'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/test_case.rb:184:in `block (2 levels) in temp_dir'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/test_case.rb:183:in `chdir'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/test_case.rb:183:in `block in temp_dir'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/tmpdir.rb:89:in `mktmpdir'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/test_case.rb:182:in `temp_dir'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/rdoc/test_rdoc_options.rb:96:in `test_encode_with_trim_paths'

  8) Error: [Keith: Existing error in ruby22]
TestRDocOptions#test_parse_page_dir:
ArgumentError: different prefix: "C:/" and "d:/Build/rubyinstaller-ruby23/sandbox/ruby23_build"
    d:/Build/rubyinstaller-ruby23/sandbox/ruby23_build/.ext/common/pathname.rb:520:in `relative_path_from'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:555:in `finish_page_dir'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:524:in `finish'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/rdoc/options.rb:1124:in `parse'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/rdoc/test_rdoc_options.rb:496:in `block in test_parse_page_dir'

  9) Error: [Keith: New error]
TestFile#test_realpath_encoding:
Errno::EACCES: Permission denied @ rb_file_s_symlink - (A, C:/Users/keith/AppData/Local/Temp/rubytest-realpath20160213-26592-eejxxc/x)
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/ruby/test_file.rb:271:in `symlink'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/ruby/test_file.rb:271:in `block in test_realpath_encoding'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/lib/tmpdir.rb:89:in `mktmpdir'
    d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/ruby/test_file.rb:267:in `test_realpath_encoding'

I'm assuming that the ruby23 test failures that existed in ruby22 are either not serious or particular to my machine, and either way won't impede releasing RubyInstaller. Based on that I'll look into failures 2, 3 and 9. Any insights you can offer on how to approach these would be appreciated.

Cheers

Keith

[Dušan D. Majkić]

> 2) Failure: [Keith: New failure] TestFile#test_file_share_delete
> [d:/Build/rubyinstaller-ruby23/sandbox/ruby_2_3/test/ruby/test_file.rb:418]:

This one looks reported, and fixed 15 days ago - probably it will be
there in nex patch level.

https://bugs.ruby-lang.org/issues/12171

The bug was 32bit related - constant was defined beyond 32bit range.
Patch is:
https://github.com/ruby/ruby/commit/e13db770f4bacf05ad99a90203385fe04b11978d

> 9) Error: [Keith: New error]
> TestFile#test_realpath_encoding:
> Errno::EACCES: Permission denied @ rb_file_s_symlink -

Looks like this one is related to 2), and should work with fix above.