James Moe | 2 Jun 20:09 2008

Patch level for libc063-cds3 incorrect


Hello,
  The patch level for libc063-csd3 is incorrect. It proclaims that libc is
0.6.2 while everything else says 0.6.3.
  Here is what I found for libc063-csd3 using "gcc --verbose test.c":
__KLIBC__ = 0
__KLIBC_MINOR__ = 6
__KLIBC_PATCHLEVEL__ = 2
__KLIBC_VERSION__ = 0x00060002

  See <.../usr/lib/gcc-lib/i386-pc-os2-emx/3.3.5/specs>, the "*predefines"
section.

--
jimoe (at) sohnen-moe (dot) com
Peter Weilbacher | 20 Jun 10:14 2008

Re: Hg/Mercurial

On 20.05.2008 17:02, Andy Willis wrote:
> Dave Yeo wrote:
>> On Tue, 20 May 2008 07:12:59 -0600, Andy Willis wrote:
>>> I finally was able to build a getopt and then build xmlto but when
>>> xmlto is called to create a man file from the xml file it fails.  I
>>> finally just did a build without documentation.  Seems to have built,
>>> though I am not sure how to use Hg yet to be able to really test it.
>>
>> Did you try make tests?
> I did but:
> e:\code\mercurial-1.0[0]make tests
> cd tests && python run-tests.py
> WARNING: Did not find prerequisite tool: gunzip.exe
> Unknown command "E:"
> 
> My gunzip says that it can't run in an OS/2 session and I can't figure 
> out where E: is coming from.

Did you make any more progress with Hg on OS/2?
   Peter.
Andy Willis | 20 Jun 22:04 2008
Picon

Re: Hg/Mercurial

Peter Weilbacher wrote:
> On 20.05.2008 17:02, Andy Willis wrote:
>> Dave Yeo wrote:
>>> On Tue, 20 May 2008 07:12:59 -0600, Andy Willis wrote:
>>>> I finally was able to build a getopt and then build xmlto but when
>>>> xmlto is called to create a man file from the xml file it fails.  I
>>>> finally just did a build without documentation.  Seems to have built,
>>>> though I am not sure how to use Hg yet to be able to really test it.
>>> Did you try make tests?
>> I did but:
>> e:\code\mercurial-1.0[0]make tests
>> cd tests&&  python run-tests.py
>> WARNING: Did not find prerequisite tool: gunzip.exe
>> Unknown command "E:"
>>
>> My gunzip says that it can't run in an OS/2 session and I can't figure
>> out where E: is coming from.
>
> Did you make any more progress with Hg on OS/2?
>     Peter.
I've gotten distracted with other things and haven't worked on it 
since then... will pick it up soon.
Andy
Dave Yeo | 23 Jun 01:29 2008
Picon

Re: Re: Hg/Mercurial

On Fri, 20 Jun 2008 10:14:17 +0200, Peter Weilbacher wrote:

>
>On 20.05.2008 17:02, Andy Willis wrote:
>> Dave Yeo wrote:
>>> On Tue, 20 May 2008 07:12:59 -0600, Andy Willis wrote:
>>>> I finally was able to build a getopt and then build xmlto but when
>>>> xmlto is called to create a man file from the xml file it fails.  I
>>>> finally just did a build without documentation.  Seems to have built,
>>>> though I am not sure how to use Hg yet to be able to really test it.
>>>
>>> Did you try make tests?
>> I did but:
>> e:\code\mercurial-1.0[0]make tests
>> cd tests && python run-tests.py
>> WARNING: Did not find prerequisite tool: gunzip.exe
>> Unknown command "E:"
>> 
>> My gunzip says that it can't run in an OS/2 session and I can't figure 
>> out where E: is coming from.
>
>Did you make any more progress with Hg on OS/2?
>   Peter.
>

I also got distracted. Returned to my attempt to build it on EMX with
gcc303 and the problem is that I don't know Python :)
mercurial/util.py needs an OS2 section added in the Platform specific
variants section. Right now it has one if section for Windows and an
else for everything else. Not sure how to do an if else between the
(Continue reading)

Dave Yeo | 23 Jun 05:03 2008
Picon

Re: Re: Hg/Mercurial

On Sun, 22 Jun 2008 16:29:41 -0700, Dave Yeo wrote:

>On Fri, 20 Jun 2008 10:14:17 +0200, Peter Weilbacher wrote:
[...]
>>Did you make any more progress with Hg on OS/2?
>>   Peter.
>>
>
>I also got distracted. Returned to my attempt to build it on EMX with
>gcc303 and the problem is that I don't know Python :)
>mercurial/util.py needs an OS2 section added in the Platform specific
>variants section. Right now it has one if section for Windows and an
>else for everything else. Not sure how to do an if else between the
>Windows section and the else:
>We need to at least make sure stdout is in binary format and define
>samestat as false (I think samestat just checks for hard links).
>Probably some other fixes can be borrowed from the Windows section and
>of course some things will never work like updating open files.  

Got a little further. I guess someone who knows how is going to have to
write a module to import a few things like setmode() and _fsetmode.
Probably others. Right now it is dieing with
AttributeError: 'module' object has no attribute 'set_binary'
Windows has a msvcrt module which imports things like setmode() and
quite a few other things
Dave
Andy Willis | 23 Jun 09:32 2008
Picon

Re: Hg/Mercurial

Dave Yeo wrote:
> On Fri, 20 Jun 2008 10:14:17 +0200, Peter Weilbacher wrote:
>
>> On 20.05.2008 17:02, Andy Willis wrote:
>>> Dave Yeo wrote:
>>>> On Tue, 20 May 2008 07:12:59 -0600, Andy Willis wrote:
>>>>> I finally was able to build a getopt and then build xmlto but when
>>>>> xmlto is called to create a man file from the xml file it fails.  I
>>>>> finally just did a build without documentation.  Seems to have built,
>>>>> though I am not sure how to use Hg yet to be able to really test it.
>>>> Did you try make tests?
>>> I did but:
>>> e:\code\mercurial-1.0[0]make tests
>>> cd tests&&  python run-tests.py
>>> WARNING: Did not find prerequisite tool: gunzip.exe
>>> Unknown command "E:"
>>>
>>> My gunzip says that it can't run in an OS/2 session and I can't figure
>>> out where E: is coming from.
>> Did you make any more progress with Hg on OS/2?
>>    Peter.
>>
>
> I also got distracted. Returned to my attempt to build it on EMX with
> gcc303 and the problem is that I don't know Python :)
> mercurial/util.py needs an OS2 section added in the Platform specific
> variants section. Right now it has one if section for Windows and an
> else for everything else. Not sure how to do an if else between the
> Windows section and the else:
> We need to at least make sure stdout is in binary format and define
(Continue reading)

Andy Willis | 23 Jun 09:59 2008
Picon

Re: Hg/Mercurial

Andy Willis wrote:
> Andy Willis wrote:
>> Peter Weilbacher wrote:
>>> On 20.05.2008 15:12, Andy Willis wrote:
>>>
>>>> I finally was able to build a getopt and then build xmlto but when
>>>> xmlto is called to create a man file from the xml file it fails. I
>>>> finally just did a build without documentation. Seems to have built,
>>>> though I am not sure how to use Hg yet to be able to really test it.
>>>
>>> Great!
>>>
>>> There are docs for Mozilla at
>>> http://developer.mozilla.org/en/docs/Mercurial_FAQ
>>> which are perhaps helpful to tell you how to use it. The first thing
>>> to test
>>> is apparently
>>> hg clone<URL> <local-dirname>
>>>
>>> Peter.
>> OK, I ran
>>> python \usr\local\scripts\hg clone
>>> http://hg.mozilla.org/mozilla-central/ mozilla
>> and it pulled the source.
>> However, when I ran
>> cd mozilla
>> python client.py checkout
>> I then get errors from python:
>>> e:\hg\mozilla[0]python client.py checkout
>>> Executing command: ['cvs', '-d',
(Continue reading)

Andrew MacIntyre | 23 Jun 13:00 2008
Picon
Picon

Re: Re: Hg/Mercurial

Dave Yeo wrote:
> On Fri, 20 Jun 2008 10:14:17 +0200, Peter Weilbacher wrote:
> 
>> On 20.05.2008 17:02, Andy Willis wrote:
>>> Dave Yeo wrote:
>>>> On Tue, 20 May 2008 07:12:59 -0600, Andy Willis wrote:
>>>>> I finally was able to build a getopt and then build xmlto but when
>>>>> xmlto is called to create a man file from the xml file it fails.  I
>>>>> finally just did a build without documentation.  Seems to have built,
>>>>> though I am not sure how to use Hg yet to be able to really test it.
>>>> Did you try make tests?
>>> I did but:
>>> e:\code\mercurial-1.0[0]make tests
>>> cd tests && python run-tests.py
>>> WARNING: Did not find prerequisite tool: gunzip.exe
>>> Unknown command "E:"
>>>
>>> My gunzip says that it can't run in an OS/2 session and I can't figure 
>>> out where E: is coming from.
>> Did you make any more progress with Hg on OS/2?
>>   Peter.
>>
> 
> I also got distracted. Returned to my attempt to build it on EMX with
> gcc303 and the problem is that I don't know Python :)
> mercurial/util.py needs an OS2 section added in the Platform specific
> variants section. Right now it has one if section for Windows and an
> else for everything else. Not sure how to do an if else between the
> Windows section and the else:
> We need to at least make sure stdout is in binary format and define
(Continue reading)

Andrew MacIntyre | 23 Jun 13:12 2008
Picon
Picon

Re: Re: Hg/Mercurial

Andy Willis wrote:
> Andy Willis wrote:
>> Andy Willis wrote:
>>> Peter Weilbacher wrote:
>>>> On 20.05.2008 15:12, Andy Willis wrote:
>>>>
>>>>> I finally was able to build a getopt and then build xmlto but when
>>>>> xmlto is called to create a man file from the xml file it fails. I
>>>>> finally just did a build without documentation. Seems to have built,
>>>>> though I am not sure how to use Hg yet to be able to really test it.
>>>>
>>>> Great!
>>>>
>>>> There are docs for Mozilla at
>>>> http://developer.mozilla.org/en/docs/Mercurial_FAQ
>>>> which are perhaps helpful to tell you how to use it. The first thing
>>>> to test
>>>> is apparently
>>>> hg clone<URL> <local-dirname>
>>>>
>>>> Peter.
>>> OK, I ran
>>>> python \usr\local\scripts\hg clone
>>>> http://hg.mozilla.org/mozilla-central/ mozilla
>>> and it pulled the source.
>>> However, when I ran
>>> cd mozilla
>>> python client.py checkout
>>> I then get errors from python:
>>>> e:\hg\mozilla[0]python client.py checkout
(Continue reading)

Dave Yeo | 23 Jun 17:11 2008
Picon

Re: Re: Hg/Mercurial

On Mon, 23 Jun 2008 22:00:16 +1100, Andrew MacIntyre wrote:

>> I also attempted to build Python with klibc but have had problems with
>> the leading underline in def files. It is hard to see how Python builds
>> a def file.
>> Andy and/or Paul did you adapt the EMX makefile or did you go with
>> configure? I went with the EMX makefile as it should build DLLs.
>
>I'll try and have a look at the mercurial code in the next few days.
>
>I did have intentions of adding support for the libc toolchain to the
>EMX makefile, but the tuits haven't been available.

It doesn't really need much adapting for libc. I generated a new
pyconfig.h and merged with the included one in PC/os2emx. Added
underlines to python25.def. Changed the makefile a bit to use -Zomf
(omf builds support fork), removed the rule for EXEOPT (libc has pretty
decent defaults for stack) as emxbind did not support -a. This just
left adding support for the autogenerated DEF files to have the
underline prefix for symbols.
I did have a problem with pthreads.h being included which I just hacked
around.
This should create a perfectly usable Python. Then thought can be given
to removing some of the EMX hacks like symlinks.

>
>Did you try Paul's build of Python 2.5.2.  I downloaded it but haven't
>had the time to check it out.

I tried Paul's build, it never created a makefile for Hg/Mercurial
(Continue reading)


Gmane