[lnkForumImage]
TotalShareware - Download Free Software

Confronta i prezzi di migliaia di prodotti.
Asp Forum
 Home | Login | Register | Search 


 

Forums >

comp.lang.ruby

Compile 1.8.6-p110 under MinGW - test-all problems

Luis Lavena

9/27/2007 3:04:00 PM

Hello everybody,

I'm trying to bring up to date the build of Ruby for MinGW. Part of
this is that I found it performs better compiled with GCC than VC8
(Visual Studio 2005).

http://pastie.caboo.se/pa...

I don't want to start another big discussion about what compiler
performs better, but just to scratch my itch...

I searched the group and found that test-all fails on SOAP tests since
you need a previous installation of ruby to do the CGI stuff...

The thing is that even I have it installed and available on path
(after a make install) still get several failures and at least 50
errors.

Found that mix between msys don't blend too well. test-all was trying
to run /usr/local/bin/ruby.exe with D:/path/to/lib/webrick/httpservlet/
cgi_runner.rb

that path mixture don't work under MSYS.

I've narrow and didn't get working just 3 failures:
http://pastie.caboo...

/miniruby.exe ../ruby-1.8.6-p110/runruby.rb --extout=.ext -- "../
ruby-1.8.6-p110/test/runner.rb" --basedir="../ruby-1.8.6-p110/test" --
runner=console
Loaded suite .
Started
[lots of stripped dots]
Finished in 81.438 seconds.

1) Failure:
test_del_trailing_separator_25(TestPathname) [../ruby-1.8.6-p110/test/
pathname/test_pathname.rb:27]:
del_trailing_separator("\225\\\\").
<"\225\\"> expected but was
<"\225">.

2) Failure:
test_should_propagate_signaled(TestBeginEndBlock) [../ruby-1.8.6-p110/
test/ruby/test_beginendblock.rb:83]:
<nil> expected but was
<3>.

3) Failure:
test_cgi(TestWEBrickCGI)
[../ruby-1.8.6-p110/test/webrick/test_cgi.rb:27:in `test_cgi'
D:/Users/Developer/code/msys-src/ruby-1.8.6-p110/lib/net/http.rb:
1050:in `request'
D:/Users/Developer/code/msys-src/ruby-1.8.6-p110/lib/net/http.rb:
2133:in `reading_body'
D:/Users/Developer/code/msys-src/ruby-1.8.6-p110/lib/net/http.rb:
1049:in `request'
D:/Users/Developer/code/msys-src/ruby-1.8.6-p110/lib/net/http.rb:
1034:in `request'
D:/Users/Developer/code/msys-src/ruby-1.8.6-p110/lib/net/http.rb:
543:in `start'
D:/Users/Developer/code/msys-src/ruby-1.8.6-p110/lib/net/http.rb:
1032:in `request'
../ruby-1.8.6-p110/test/webrick/test_cgi.rb:27:in `test_cgi'
../ruby-1.8.6-p110/test/webrick/utils.rb:26:in `call'
../ruby-1.8.6-p110/test/webrick/utils.rb:26:in `start_server'
../ruby-1.8.6-p110/test/webrick/utils.rb:34:in `start_httpserver'
../ruby-1.8.6-p110/test/webrick/test_cgi.rb:24:in `test_cgi']:
<"/webrick.cgi"> expected but was
<"<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.0//EN\">\n<HTML>\n
<HEAD><TITLE>Internal Server Error</TITLE></HEAD>\n <BODY>\n
<H1>Internal Server Error</H1>\n Premature end of script headers:
D:/Users/Developer/code/msys-src/ruby-1.8.6-p110/test/webrick/
webrick.cgi\n <HR>\n <ADDRESS>\n WEBrick/1.3.1 (Ruby/
1.8.6/2007-09-23) at\n 127.0.0.1:4190\n </ADDRESS>\n </BODY>
\n</HTML>\n">.

1467 tests, 13738 assertions, 3 failures, 0 errors

===

Is this right?, anyone can confirm/deny, point me in the right
direction?

I'm trying to solve these issues to jump to the next step: fulfill the
extensions dependencies.

Thanks anyone for their time.

Regards,

Luis
--
Posted via http://www.ruby-....

8 Answers

Nobuyoshi Nakada

9/28/2007 6:43:00 AM

0

Hi,

At Fri, 28 Sep 2007 00:04:12 +0900,
Luis Lavena wrote in [ruby-talk:271285]:
> 1) Failure:
> test_del_trailing_separator_25(TestPathname) [../ruby-1.8.6-p110/test/
> pathname/test_pathname.rb:27]:
> del_trailing_separator("\225\\\\").
> <"\225\\"> expected but was
> <"\225">.

It's valid only on Japanese version system.

> 2) Failure:
> test_should_propagate_signaled(TestBeginEndBlock) [../ruby-1.8.6-p110/
> test/ruby/test_beginendblock.rb:83]:
> <nil> expected but was
> <3>.

Windows doesn't have signal mechanism, so the test should be
skipped.

--
Nobu Nakada

Luis Lavena

9/28/2007 6:50:00 AM

0

Nobuyoshi Nakada wrote:
> [...]
>
> It's valid only on Japanese version system.
>
>> 2) Failure:
>> test_should_propagate_signaled(TestBeginEndBlock) [../ruby-1.8.6-p110/
>> test/ruby/test_beginendblock.rb:83]:
>> <nil> expected but was
>> <3>.
>
> Windows doesn't have signal mechanism, so the test should be
> skipped.

So I must ignore these failures and take p110 with 0 failures?.

That's great news, now I could invest some time on getting the
dependencies and extensions builds.

Thank you Nobu for your time.

Luis
--
Posted via http://www.ruby-....

Nobuyoshi Nakada

9/28/2007 8:04:00 AM

0

Hi,

At Fri, 28 Sep 2007 15:49:55 +0900,
Luis Lavena wrote in [ruby-talk:271411]:
> So I must ignore these failures and take p110 with 0 failures?.

It's OK.

--
Nobu Nakada

Roger Pack

10/19/2007 5:40:00 AM

0

Luis Lavena wrote:
> Hello everybody,
>
> I'm trying to bring up to date the build of Ruby for MinGW. Part of
> this is that I found it performs better compiled with GCC than VC8
> (Visual Studio 2005).
>
Detailed descriptions on how it ended up working would be nice (how to
compile it). Thanks!
--
Posted via http://www.ruby-....

Roger Pack

10/19/2007 10:54:00 AM

0

My question is with minGW compilation--how do you create the zlib .so
file, especially after you've already compiled everything else (and I do
have the zlib libraries also compiled with mingw)?
there is this ext/zlib/extconf.rb file, but it doesn't seem to do
anything. Any thoughts? Trying to put together a minGW release, for
fun :)
Take care.
-Roger
--
Posted via http://www.ruby-....

Luis Lavena

10/19/2007 7:40:00 PM

0

Roger Pack wrote:
> My question is with minGW compilation--how do you create the zlib .so
> file, especially after you've already compiled everything else (and I do
> have the zlib libraries also compiled with mingw)?
> there is this ext/zlib/extconf.rb file, but it doesn't seem to do
> anything. Any thoughts? Trying to put together a minGW release, for
> fun :)
> Take care.
> -Roger

Roger, I used MinGW candidate using the Automated MinGW Installer:

http://sourceforge.net/project/showfiles.php?group_id=2435&package...

Only installed the C compiler (do not install make since will clash and
do weird things).

Installed MSYS-1.0.10 and MSYS DTK.

Then I followed the steps to install MinGWports and installed as ports
readline and zlib

(Zlib is official, readline is a patch in the tracket at sourceforge).

I unzipped the ruby package and run inside a MSYS console:

/configure --with-winsock2
make

At this point you should check the output for 'compiling readline'
'compiling zlib' etc... and see if everything worked.

make check (to run tests)
make install

That will create bin/lib/man/share in the root of your drive (thats
because generated make file isn't compatible with MSYS).

You should have there a working implementation.

zlib.dll will be located at you MinGW bin folder, which is where
MinGWport install it.

The same for readline.

Write this mail from memory took me a few minutes, but compile with VC8
(replying your other mail) is more complex.

I could give you a full checkout of my own repo where I was hacking
this.

Contact me off the list if you're interested.

Regards and hope this help,

Luis
--
Posted via http://www.ruby-....

Dave Heil

5/2/2011 11:31:00 AM

0

On 4/30/2011 23 17, Ray Fischer wrote:
> Shall not be infringed<hot-ham-and-cheese@hotmail.com> wrote:
>> On Apr 30, 6:34 pm, rfisc...@sonic.net (Ray Fischer) wrote:
>>> Shall not be infringed<hot-ham-and-che...@hotmail.com> wrote:
>>>> On Apr 30, 4:05 pm, rfisc...@sonic.net (Ray Fischer) wrote:
>>>>> Harry Hope<riv...@ix.netcom.com> wrote:
>>>>>> After two and a half years, Obama comes up with a forgery.
>>>
>>>>> Kook.
>>>
>>>>>> Look at the type. If this birth certificate was prepared on the
>>>>>> same typewriter,
>>>
>>>>> Who says it was, kook?
>>>
>>>>>> the partial strikes of the "O" and "K" keys
>>>>>> would be consistent throughout the document. They are not.
>>>
>>>>> Yeah, it's obviously not prepared with a modern computer since
>>>>> computers are flawless in producing test. That measns that it
>>>>> must be a forgery.
>>>
>>>>>> The
>>>>>> capital "O" is misaligned in 6b yet aligned with other letters
>>>>>> through the document. This is an impossibility since the upper
>>>>>> and lower characters are on the same striking key.
>>>
>>>>> The crazy rightard hasn't even seen a real typewriter.
>>>
>>>> I still own one. Other than breaking out the wrenches and
>>>> screwdrivers, how do I make it change alignment in mid-document?
>>>
>>> LOL! You don't own a mechanical typewriter, racist birther, or you
>>> wouldn't be asking such a stupid question. Shit, it was tough to get
>>> mechanical typewriters to have all the letters aligned reasonably well.
>>
>> The fuck I don't, Ray. It's a Royal; 100% mechanical just like in
>> 1964.
>
> Then you're just a lying troll.
>
>> An if the alignment of one of the letters if off on line #1, it is
>
> Bump the carriage, roll the paper up to check what you typed, bang on
> the keys harder, or softer. There were countless ways to alter the
> alignment on those things. Mechanical parts wear out and get loose,
> and alignment shifts.

You raise some interesting points, Wrong Ray--interesting but ill-informed.

How does one bump the carriage while sitting in front of a typewriter?
If someone bumps the carriage, how does that make a character printer
lighter or darker? It is possible to defeat the detent action of a
carriage but it must be deliberately done. How does that change whether
a character prints lighter or darker? It is not necessary to "roll the
paper up to check what you typed." What you've typed can be seen
without touching the carriage.

> Which you should know, but you're a racist little troll pretending to
> be stupid.

You've just called another a racist. I know who this one is and I don't
believe you for a moment.

When you're stupid, I don't think you're pretending, Wrong Ray.


rfischer

5/3/2011 3:48:00 AM

0

Dave Heil <k8mn@frontiernet.net> wrote:
>On 4/30/2011 23 17, Ray Fischer wrote:
>> Shall not be infringed<hot-ham-and-cheese@hotmail.com> wrote:
>>> On Apr 30, 6:34 pm, rfisc...@sonic.net (Ray Fischer) wrote:
>>>> Shall not be infringed<hot-ham-and-che...@hotmail.com> wrote:
>>>>> On Apr 30, 4:05 pm, rfisc...@sonic.net (Ray Fischer) wrote:
>>>>>> Harry Hope<riv...@ix.netcom.com> wrote:
>>>>>>> After two and a half years, Obama comes up with a forgery.
>>>>
>>>>>> Kook.
>>>>
>>>>>>> Look at the type. If this birth certificate was prepared on the
>>>>>>> same typewriter,
>>>>
>>>>>> Who says it was, kook?
>>>>
>>>>>>> the partial strikes of the "O" and "K" keys
>>>>>>> would be consistent throughout the document. They are not.
>>>>
>>>>>> Yeah, it's obviously not prepared with a modern computer since
>>>>>> computers are flawless in producing test. That measns that it
>>>>>> must be a forgery.
>>>>
>>>>>>> The
>>>>>>> capital "O" is misaligned in 6b yet aligned with other letters
>>>>>>> through the document. This is an impossibility since the upper
>>>>>>> and lower characters are on the same striking key.
>>>>
>>>>>> The crazy rightard hasn't even seen a real typewriter.
>>>>
>>>>> I still own one. Other than breaking out the wrenches and
>>>>> screwdrivers, how do I make it change alignment in mid-document?
>>>>
>>>> LOL! You don't own a mechanical typewriter, racist birther, or you
>>>> wouldn't be asking such a stupid question. Shit, it was tough to get
>>>> mechanical typewriters to have all the letters aligned reasonably well.
>>>
>>> The fuck I don't, Ray. It's a Royal; 100% mechanical just like in
>>> 1964.
>>
>> Then you're just a lying troll.
>>
>>> An if the alignment of one of the letters if off on line #1, it is
>>
>> Bump the carriage, roll the paper up to check what you typed, bang on
>> the keys harder, or softer. There were countless ways to alter the
>> alignment on those things. Mechanical parts wear out and get loose,
>> and alignment shifts.
>
>You raise some interesting points, Wrong Ray--interesting but ill-informed.

You're still a stupid liar, Sig.

>How does one bump the carriage while sitting in front of a typewriter?

Figure it out, moron.

--
Ray Fischer | Mendacracy (n.) government by lying
rfischer@sonic.net | The new GOP ideal