Reposted on the correct list.
Jason Tishler wrote:
By the way, is it reliable to use objdump -x to find the base
(ImageBase) ? For cygssl-0.9.8.dll it is 1000, but I thought
the base was 0x6300
I use a command line like the following:
$ objdump -p /usr/bin/cygssl-0.9.7.dll |
- Original Message -
From: Gerrit P. Haase
To: cygwin-apps@cygwin.com
Sent: Tuesday, July 12, 2005 3:00 AM
Subject: Re: Observation for ALL maintainers who provide dlls (was Re:
question for perl maintainer)
Reposted on the correct list.
Jason Tishler wrote:
By the way, is it
On Fri, Jul 08, 2005 at 02:04:17PM -0400, Christopher Faylor wrote:
On Fri, Jul 08, 2005 at 01:42:34PM -0400, Pierre A. Humblet wrote:
From: Christopher Faylor
Do we need to coordinate this among all package maintainers, maybe?
Maybe we could publish a list of all of the dlls in the system
- Original Message -
From: Jason Tishler [EMAIL PROTECTED]
To: cygwin-apps@cygwin.com
Sent: Monday, July 11, 2005 9:06 AM
Subject: Re: Observation for ALL maintainers who provide dlls (was Re:
question for perl maintainer)
On Fri, Jul 08, 2005 at 02:04:17PM -0400, Christopher Faylor
Pierre,
On Mon, Jul 11, 2005 at 09:34:27AM -0400, Pierre A. Humblet wrote:
- Original Message -
From: Jason Tishler [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Thanks for the above. :,(
Unfortunately, I have found that the DLLs need a gap between them to
guarantee that fork() won't
Jason Tishler wrote:
By the way, is it reliable to use objdump -x to find the base
(ImageBase) ? For cygssl-0.9.8.dll it is 1000, but I thought the
base was 0x6300
I use a command line like the following:
$ objdump -p /usr/bin/cygssl-0.9.7.dll | fgrep ImageBase
ImageBase
Gerrit,
On Mon, Jul 11, 2005 at 04:58:47PM +0200, Gerrit P. Haase wrote:
Jason Tishler wrote:
By the way, is it reliable to use objdump -x to find the base
(ImageBase) ? For cygssl-0.9.8.dll it is 1000, but I thought
the base was 0x6300
I use a command line like the following:
--- Jason Tishler wrote:
I meant --enable-auto-image-base will only help, but not completely
solve the problem.
nor the centralized database will work...
I'm concerned a global, centralized database will run out of address
space sooner rather than later.
Hm, making rebase a part
The attached script takes the name of a .exe or .dll,
uses cygcheck to find the dll dependence and checks for conflicts.
This will allow you to check your favorite applications or dlls,
seeing if --enable-auto-image-base works for you.
Here is an example output:
~: check_bases
On Mon, Jul 11, 2005 at 12:11:32PM -0700, James R. Phillips wrote:
--- Jason Tishler wrote:
I meant --enable-auto-image-base will only help, but not completely
solve the problem.
nor the centralized database will work...
I'm concerned a global, centralized database will run out of
[snip]
The 0x6000-0x680 range is suggested by MSDN but I've
also seen 0x6400-0x680 suggested. That's still 64MB
of address space
...and 64MB should be more than enough for anybody. ;-)
--
Gary R. Van Sickle
Igor Pechtchanski wrote:
On Fri, 8 Jul 2005, James R. Phillips wrote:
--- Pierre A. Humblet wrote:
The more I think about this, the more I believe that we shouldn't have
to continually tell users to run rebaseall. Setting the base address
is something that should be done once, by the
--- Gerrit P. Haase wrote:
Igor Pechtchanski wrote:
Isn't it a configure option? You can change the conf() function to pass
it in. For linker options, set MY_LDFLAGS in the beginning of the script.
Igor
Doesn't libtool always defines --image-base when building a DLL, or
James R. Phillips wrote:
--- Gerrit P. Haase wrote:
Igor Pechtchanski wrote:
Isn't it a configure option? You can change the conf() function to pass
it in. For linker options, set MY_LDFLAGS in the beginning of the script.
Igor
Doesn't libtool always defines --image-base
Gerrit P. Haase wrote:
Doesn't libtool always defines --image-base when building a DLL, or is
this only with ancient libtool versions?
only ancient versions did this. Versions of libtool that are merely
immensely old defined '--enable-auto-image-base'. No version in the
lasts three years
On Sat, Jul 09, 2005 at 02:28:32PM -0400, Charles Wilson wrote:
Gerrit P. Haase wrote:
Doesn't libtool always defines --image-base when building a DLL, or is
this only with ancient libtool versions?
only ancient versions did this. Versions of libtool that are merely
immensely old defined
Christopher Faylor wrote:
I vaguely remember somebody more knowledgeable (Danny? cgf? Mumit?)
than I suggesting that --e-a-i-b was a bad default choice -- but I don't
remember who or why, and I can't find it in the archives.
I vaguely recall this, too. I'm probably being terribly
Charles Wilson wrote:
Still looking for the issue previously discussed.
Robert's meesage that I just posted was originally sent to the list in
June 2001. The following thread is three months later, from September
2001. However, it seems to pretty thoroughly rehash the issue in
question,
Charles Wilson wrote:
I'll try to find the *original* discussion of this issue, which should
have occured BEFORE June 2001...
I think the previous discussion that Robert mentioned was actually only
a few days earlier in June 2001, on the cygwin-developers list:
dll base address
Wouldn't this patch to ld/pe.em solve the e-a-i-b problem?
*** pe.em.orig Sun Jul 10 12:33:54 2005
--- pe.em Sun Jul 10 12:33:29 2005
*** static unsigned long
*** 666,672
compute_dll_image_base (const char *ofile)
{
unsigned long hash = strhash (ofile);
! return
On Sun, Jul 10, 2005 at 12:42:12PM +1200, Danny Smith wrote:
Wouldn't this patch to ld/pe.em solve the e-a-i-b problem?
*** pe.em.orig Sun Jul 10 12:33:54 2005
--- pe.em Sun Jul 10 12:33:29 2005
*** static unsigned long
*** 666,672
compute_dll_image_base (const char *ofile)
{
On Sat, Jul 09, 2005 at 07:53:48PM -0400, Charles Wilson wrote:
Charles Wilson wrote:
I'll try to find the *original* discussion of this issue, which should
have occured BEFORE June 2001...
I think the previous discussion that Robert mentioned was actually only
a few days earlier in June 2001,
cgf wrote
Am I reading this right, though? auto-image-base puts everything
above 0x6000. That doesn't seem quite right.
I apologise for breaking thread but I'm on someone else's machine and have to
use the mail clinet available.
As I understand it, memory address space is usually
[redirecting to cygwin-apps]
On Fri, Jul 08, 2005 at 07:27:55PM +0200, Corinna Vinschen wrote:
On Jul 8 17:20, Gerrit P. Haase wrote:
Christopher Faylor wrote:
I don't think so but I don't think it will use cygwin's address anyway.
Ok. Maybe Corinna should do the same for openssl?
What?
On Jul 8 13:32, Christopher Faylor wrote:
[redirecting to cygwin-apps]
On Fri, Jul 08, 2005 at 07:27:55PM +0200, Corinna Vinschen wrote:
On Jul 8 17:20, Gerrit P. Haase wrote:
Christopher Faylor wrote:
I don't think so but I don't think it will use cygwin's address anyway.
Ok. Maybe
- Original Message -
From: Christopher Faylor [EMAIL PROTECTED]
To: cygwin-apps@cygwin.com
Cc: cygwin-apps@cygwin.com
Sent: Friday, July 08, 2005 1:32 PM
Subject: Observation for ALL maintainers who provide dlls (was Re: question
for perl maintainer)
[redirecting to cygwin-apps]
On
On Fri, Jul 08, 2005 at 01:42:34PM -0400, Pierre A. Humblet wrote:
From: Christopher Faylor
Do we need to coordinate this among all package maintainers, maybe?
Maybe we could publish a list of all of the dlls in the system along
with standard base addresses for each and ask that maintainers make
--- Pierre A. Humblet wrote:
The more I think about this, the more I believe that we shouldn't have
to continually tell users to run rebaseall. Setting the base address
is something that should be done once, by the maintainer, not every
time a person installs a package.
Amen, but
On Fri, 8 Jul 2005, James R. Phillips wrote:
--- Pierre A. Humblet wrote:
The more I think about this, the more I believe that we shouldn't have
to continually tell users to run rebaseall. Setting the base address
is something that should be done once, by the maintainer, not every
29 matches
Mail list logo