On 11/21/09, Christopher Faylor
wrote:
> On Sat, Nov 21, 2009 at 06:54:37PM +, Dave Korn wrote:
>>mike marchywka wrote:
>>>On XP, go to control panel->System->Advanced ->Envireonment Variables.
>>>These do no appear to be process specific. These are what I was
>>>talking about where I saw gar
On Sat, Nov 21, 2009 at 06:54:37PM +, Dave Korn wrote:
>mike marchywka wrote:
>>On XP, go to control panel->System->Advanced ->Envireonment Variables.
>>These do no appear to be process specific. These are what I was
>>talking about where I saw garbage,
>
>Since registry accesses are atomic an
mike marchywka wrote:
> On XP, go to control panel->System->Advanced ->Envireonment Variables.
> These do no appear to be process specific. These are what I was talking
> about where I saw garbage,
Since registry accesses are atomic and serialized, whatever you saw there
must reflect exactly t
On 11/21/09, Dave Korn wrote:
> mike marchywka wrote:
>
>> No, read my posts. I contributed an observation that suggests the
>> windoze variables don't handle concurrent updates well. Two threads
>> needn't come from same process.
>
> This is where your misunderstanding comes in. If they aren't
Dave Korn wrote:
> If they aren't from the same
> process, they cannot be updating the same copy of the environment, since each
> process has its own
http://msdn.microsoft.com/en-us/library/ms682653(VS.85).aspx
MSDN reference: hopefully this should clear up the details about how
environment blo
mike marchywka wrote:
> No, read my posts. I contributed an observation that suggests the
> windoze variables don't handle concurrent updates well. Two threads
> needn't come from same process.
This is where your misunderstanding comes in. If they aren't from the same
process, they cannot be
On 11/21/09, Christopher Faylor
wrote:
> On Sat, Nov 21, 2009 at 08:16:29AM -0500, mike marchywka wrote:
>>On 11/21/09, Corinna Vinschen wrote:
>>>No, never. If there is a concurrency problem with the envionment, it's
>>>between threads of the same process. In that case, we can make getenv,
>>>s
On Sat, Nov 21, 2009 at 08:16:29AM -0500, mike marchywka wrote:
>On 11/21/09, Corinna Vinschen wrote:
>>No, never. If there is a concurrency problem with the envionment, it's
>>between threads of the same process. In that case, we can make getenv,
>>setenv and friends thread-safe, but it's still
On Nov 21 08:16, mike marchywka wrote:
> On 11/21/09, Corinna Vinschen wrote:
> > On Nov 20 14:04, mike marchywka wrote:
> >> The OS is the only place you can when the threads are in different
> >> processes unknown to each other. Or, can two different processes share
> >> the
> >> same thread?
>
On 11/21/09, Corinna Vinschen wrote:
> On Nov 20 14:04, mike marchywka wrote:
>> On 11/20/09, Christopher Faylor wrote:
>> > On Fri, Nov 20, 2009 at 01:19:57PM -0500, mike marchywka wrote:
>> >>On 11/20/09, Corinna Vinschen wrote:
>> >>> And given POSIX, if so, it would be an application bug if t
On Nov 20 14:04, mike marchywka wrote:
> On 11/20/09, Christopher Faylor wrote:
> > On Fri, Nov 20, 2009 at 01:19:57PM -0500, mike marchywka wrote:
> >>On 11/20/09, Corinna Vinschen wrote:
> >>> And given POSIX, if so, it would be an application bug if the application
> >>> doesn't care by itself
On 11/20/09, Christopher Faylor
wrote:
> On Fri, Nov 20, 2009 at 02:04:01PM -0500, mike marchywka wrote:
>>I guess my point is that there is nothing cygwin can do if windoze
>>doesn't do it. But, that was why I asked what the OP was actually
>>complaining about. In may case, IIRC, I got garbage
On Fri, Nov 20, 2009 at 02:04:01PM -0500, mike marchywka wrote:
>I guess my point is that there is nothing cygwin can do if windoze
>doesn't do it. But, that was why I asked what the OP was actually
>complaining about. In may case, IIRC, I got garbage in the windoze
>variables rather than just so
On 11/20/09, Christopher Faylor
wrote:
> On Fri, Nov 20, 2009 at 01:19:57PM -0500, mike marchywka wrote:
>>On 11/20/09, Corinna Vinschen wrote:
>>> On Nov 20 09:32, Christopher Faylor wrote:
On Fri, Nov 20, 2009 at 12:20:52PM +0100, Corinna Vinschen wrote:
>On Nov 20 06:04, mike marchyw
On Fri, Nov 20, 2009 at 01:19:57PM -0500, mike marchywka wrote:
>On 11/20/09, Corinna Vinschen wrote:
>> On Nov 20 09:32, Christopher Faylor wrote:
>>> On Fri, Nov 20, 2009 at 12:20:52PM +0100, Corinna Vinschen wrote:
>>> >On Nov 20 06:04, mike marchywka wrote:
>>> >> On 11/19/09, Larry Hall (Cygw
On 11/20/09, Corinna Vinschen wrote:
> On Nov 20 09:32, Christopher Faylor wrote:
>> On Fri, Nov 20, 2009 at 12:20:52PM +0100, Corinna Vinschen wrote:
>> >On Nov 20 06:04, mike marchywka wrote:
>> >> On 11/19/09, Larry Hall (Cygwin)
>> >> wrote:
>> >>>Wow! I had a hunch that BLODA could be the ca
On Nov 20 09:32, Christopher Faylor wrote:
> On Fri, Nov 20, 2009 at 12:20:52PM +0100, Corinna Vinschen wrote:
> >On Nov 20 06:04, mike marchywka wrote:
> >> On 11/19/09, Larry Hall (Cygwin) wrote:
> >>>Wow! I had a hunch that BLODA could be the cause of the file related
> >>>problems but it never
On Fri, Nov 20, 2009 at 12:20:52PM +0100, Corinna Vinschen wrote:
>On Nov 20 06:04, mike marchywka wrote:
>> On 11/19/09, Larry Hall (Cygwin) wrote:
>>>Wow! I had a hunch that BLODA could be the cause of the file related
>>>problems but it never occurred to me that it could be screwing up
>>>envir
On Nov 20 06:04, mike marchywka wrote:
> On 11/19/09, Larry Hall (Cygwin) wrote:
> > Wow! I had a hunch that BLODA could be the cause of the file related
> > problems
> > but it never occurred to me that it could be screwing up environment
> > variables too.
>
> Anyone care to determine if env u
On 11/19/09, Larry Hall (Cygwin) wrote:
> On 11/19/2009 10:49 PM, Pete Brunet wrote:
>> ---
>> Pete Brunet wrote:
>>> I've been trying to build OpenJDK for several weeks now and have never
>>> been able to get to the end of the build because of random failures.
>>> Last night I had a hunch to turn
On 11/19/2009 10:49 PM, Pete Brunet wrote:
---
Pete Brunet wrote:
I've been trying to build OpenJDK for several weeks now and have never
been able to get to the end of the build because of random failures.
Last night I had a hunch to turn off one of the cores in BIOS and after
doing that the pro
This problem went away by disabling Norton 360.
---
Pete Brunet wrote:
> I've been trying to build OpenJDK for several weeks now and have never
> been able to get to the end of the build because of random failures.
> Last night I had a hunch to turn off one of the cores in BIOS and after
> doing t
Dave, Migrating to 1.7 didn't help. I'll disable dual core for now. -Pete
---
Dave Korn wrote:
> Pete Brunet wrote:
>
>> Thanks Dave, What is the proper way to upgrade the DLL?
>>
>
> The "proper" way? Heh:
>
> #1. Read the manual.
> #2. Take back-ups!
> #3. Then do it.
>
> Howeve
2009/11/17 Pete Brunet:
>>$'\r': command not found
>
> Please ignore my post about this. That was the result of using a
> different bat file (not cygwin.bat) that I use to start an OpenJDK build
> job and it hadn't yet been modified to set SHELLOPTS=igncr
That's your sixth post in a row on this t
>$'\r': command not found
Please ignore my post about this. That was the result of using a
different bat file (not cygwin.bat) that I use to start an OpenJDK build
job and it hadn't yet been modified to set SHELLOPTS=igncr
Pete
---
Pete Brunet wrote:
> Well I guess it's not the same. I got
> $
>$'\r': command not found
Please ignore my post about this. That was the result of using a
different bat file (not cygwin.bat) that I use to start an OpenJDK build
job and it hadn't yet been modified to set SHELLOPTS=igncr
Pete
---
Pete Brunet wrote:
> Well I guess it's not the same. I got
> $
Well I guess it's not the same. I got
$'\r': command not found
when running a script
For now I'll just dos2unix that file and figure the rest out later.
Pete
---
Pete Brunet wrote:
> This may solve the second one. One selection google found suggests
> editing cygwin.bat to add
>
> set SHELLOPTS
This may solve the second one. One selection google found suggests
editing cygwin.bat to add
set SHELLOPTS=igncr
after the @echo off in cygwin.bat
Hopefully this is the same as choosing DOS line endings in the cygwin
1.5 setup program.
Pete
---
Pete Brunet wrote:
> I solved the first problem b
I solved the first problem by copying cygintl-2.dll from my 1.5 cygwin/bin.
Pete Brunet wrote:
> Thnaks Dave, I've installed 1.7. The OpenJDK build requires
> backleveling make from 3.81 to 3.80. When I do that on cygwin 1.7 I get
> the following:
>
> p...@bevo ~
> $ cd /
>
> p...@bevo /
> $ tar
Thnaks Dave, I've installed 1.7. The OpenJDK build requires
backleveling make from 3.81 to 3.80. When I do that on cygwin 1.7 I get
the following:
p...@bevo ~
$ cd /
p...@bevo /
$ tar xjf /cygdrive/c/downloads/make-3.80-1.tar.bz2
p...@bevo /
$ bin/make.exe --version
/usr/bin/make.exe: error wh
Pete Brunet wrote:
> Thanks Dave, What is the proper way to upgrade the DLL?
The "proper" way? Heh:
#1. Read the manual.
#2. Take back-ups!
#3. Then do it.
However that's just the generic proper way to do anything. You probably
wanted a short answer along the lines of "Download the 1.7-
Thanks Dave, What is the proper way to upgrade the DLL?
Dave Korn wrote:
> Pete Brunet wrote:
>
>> I've been trying to build OpenJDK for several weeks now and have never
>> been able to get to the end of the build because of random failures.
>> Last night I had a hunch to turn off one of the c
> 1.5 is known to have bugs in this area. You can give 1.7 a try because
> I've
> fixed a number of race conditions; I'm pretty sure there are still more
> to be
> tracked down, but it's a good deal better.
Ah, that's interesting. I've been seeing odd effects too (in 1.7, 4 core
machine). Some o
Pete Brunet wrote:
> I've been trying to build OpenJDK for several weeks now and have never
> been able to get to the end of the build because of random failures.
> Last night I had a hunch to turn off one of the cores in BIOS and after
> doing that the problems have gone away. My system is a Len
34 matches
Mail list logo
