Hi,
I am using fink 0.5.3.cvs, Mac OSX 10.2.6 and Apple's GCC 3.1. I have
enabled
prebinding by following the Fink-devel which Ben Reed posted at May
16th.
Today, I did fink selfupdate-cvs, and seg_add_table in /sw/dists
disappeared.
This made the consequent `fink update-all` fail atlinking.
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
kinako wrote:
Hi,
I am using fink 0.5.3.cvs, Mac OSX 10.2.6 and Apple's GCC 3.1. I have
enabled prebinding by following the Fink-devel which Ben Reed posted
at May 16th.
Today, I did fink selfupdate-cvs, and seg_add_table in /sw/dists
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi,
First of all, thanks a lot for your great work! Fink has become an
essential part of my system for me. :)
I have one question: Do you have any plans on enabling prebinding for
qt3 by default any time soon? I have skimmed over the previous
-BEGIN PGP SIGNED MESSAGE-
Hash: RIPEMD160
On Mittwoch, April 2, 2003, at 04:36 Uhr, Karin Kosina wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi,
First of all, thanks a lot for your great work! Fink has become an
essential part of my system for me. :)
Thank you.
I have one
Hi again,
+-- David [20030402 20:08] --
| As far as I have understood from previous discussions and information
| passed on the irc channel there wont be much support for prebinding
| from our side in the near future.
I see. Thanks for the heads-up.
Please do not
Please do not get this wrong - by no means do I want to question
your decisions. But I'm interested to hear why you want to keep
prebinding disabled. I have experienced a quite noticable
improvement in application startup time after prebinding the
libraries used by that application, and I wasn't
-BEGIN PGP SIGNED MESSAGE-
Hash: RIPEMD160
On Mittwoch, April 2, 2003, at 08:54 Uhr, Benjamin Reed wrote:
Please do not get this wrong - by no means do I want to question
your decisions. But I'm interested to hear why you want to keep
prebinding disabled. I have experienced a quite
+-- Benjamin Reed [20030402 13:54] --
| As far as I understand, the only challenge is to find an address
| range that is not also used by other libraries that the library
| directly links against... -- it does not have to be a system-wide
| unique address - since
David,
+-- David [20030402 21:04] --
| Sometimes relocatable Objects that overlap in memory tend to
| be prebound wrongly
If that happens, it definitely sounds like a bug to me... in any
case, on Mac OS X, if you have an overlap, prebinding is just
disabled, so this
Karin Kosina wrote:
I still do not see how enabling prebinding can ever be bad. If I
understand the documentation correctly, the worst thing that can
happen is that prebinding cannot be used, resulting in the exact
same situation as if it hadn't been enabled in the first place...
I don't see how
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
+-- Benjamin Reed [20030402 14:45] --
| Karin Kosina wrote:
|
| I still do not see how enabling prebinding can ever be bad.
|
| I don't see how it can be bad either, but I also don't see how it can be
| *automated*. If
On Wednesday, April 2, 2003, at 11:45 AM, Benjamin Reed wrote:
Karin Kosina wrote:
I still do not see how enabling prebinding can ever be bad. If I
understand the documentation correctly, the worst thing that can
happen is that prebinding cannot be used, resulting in the exact
same situation as
On Wednesday, April 2, 2003, at 08:44 PM, Ben Hines wrote:
On Wednesday, April 2, 2003, at 11:45 AM, Benjamin Reed wrote:
Karin Kosina wrote:
I still do not see how enabling prebinding can ever be bad. If I
understand the documentation correctly, the worst thing that can
happen is that
I've had some experience with this, trying to get the oss tuxracer to
stop crashing under certain conditions.
Here are my 2 cents:
1. If one application wants to be prebound, all of its libraries and
dependencies must be compiled prebound. That means X11, qt3, and
everything elseif you're
I finally tracked down some useful information which seems related to
that got slid prebinding warning message.
Apparently in 10.2, redo_prebinding can *slide* the seg1addr of a
dynamiclib to another location. A master lookup file is required which
would contain a list of all the library
On Tuesday, February 11, 2003, at 10:46 AM, Carsten Klapp wrote:
What I see this to mean is fink might be able to build any dylib
prebound with a single static address (arbitrarily chosen for
convenience), then run redo_prebinding on it at the end of the compile
to specify our own
On Tuesday, February 11, 2003, at 06:46 pm, Carsten Klapp wrote:
I imagine Apple has such a seg_addr_table list internally for use with
all it's frameworks and dylibs. Now we need to get our hands on one to
find out what the format of this file is.
Or just pester some cctools engineers. We
On Tuesday, February 11, 2003, at 07:02 pm, Ben Hines wrote:
Just do man seg_addr_table to get all the info on the table.
No manual entry for seg_addr_table
Sure about that one?
From cctools src:
/*
* parse_seg_addr_table() opens the file_name passed to it and parses
it as a
* segment
On Tuesday, February 11, 2003, at 11:12 AM, Finlay Dobbie wrote:
On Tuesday, February 11, 2003, at 07:02 pm, Ben Hines wrote:
Just do man seg_addr_table to get all the info on the table.
No manual entry for seg_addr_table
Sure about that one?
Aha, i must have moved it into
Hehe, well at least with redo_prebinding the entire disk isn't scanned
like update_prebinding does. (BTW I think in 10.2 the installer's
rebinding is a little quicker, maybe it only calls redo_prebinding
now.) Anyway...
A little more investigation reveals:
% strings /usr/bin/redo_prebinding
On Tuesday, February 11, 2003, at 02:41 PM, Carsten Klapp wrote:
Hehe, well at least with redo_prebinding the entire disk isn't scanned
like update_prebinding does. (BTW I think in 10.2 the installer's
rebinding is a little quicker, maybe it only calls redo_prebinding
now.) Anyway...
It
-BEGIN PGP SIGNED MESSAGE-
Hash: RIPEMD160
On Samstag, Februar 8, 2003, at 02:56 Uhr, Ben Hines wrote:
On Friday, February 7, 2003, at 10:36 AM, David wrote:
Furthermore it is not as simple as simply adding --prebind. Many
pieces of software are not yet fixed to fully support
It is true that to build prebound libraries an address must be
specified, but not so for executables.
So on the Mac simply adding -prebind flags to all compiles and links
will not cause any problems for executables or libraries. If the
prebind fails when building an executable (because one or
On Saturday, February 8, 2003, at 11:49 AM, Carsten wrote:
For fink to build prebound libraries, there is a tool someone here
made called gcc-prebind 0.1 (search the mailing list archives around
last November). It might be useful to get started in the process but
ultimately I think each shlib
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Apple's prebinding release notes
(/Developer/Documentation/ReleaseNotes/Prebinding.html) say that in
order to build a library/binary prebound, all you need to do is pass
- -prebind to ld or define the LD_PREBIND environment variable. I'm
wondering
-BEGIN PGP SIGNED MESSAGE-
Hash: RIPEMD160
On Freitag, Februar 7, 2003, at 07:11 Uhr, John Davidorff Pell wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Apple's prebinding release notes
(/Developer/Documentation/ReleaseNotes/Prebinding.html) say that in
order to
Hi,
There's a lot of work involved to build all of fink's dylib libraries
as prebound, it probably won't happen any time soon.
However, any program which depends only on Apple-supplied dylibs can
easily be built prebound using the -prebind flag in CFLAGS and LDFLAGS.
Some packages already do
Would adding (either manually, or automatically) -prebind/LD_PREBIND to packages break them if it doesn't work? I don't think so, so why not start adding this automatically so that if the package can be prebound, it is? Please tell me if I'm missing something big.
Thanx,
JP
On Friday,
No it doesn't break anything when prebinding fails for some reason,
just outputs a compiler warning and continues building. I patched my
personal copy of fink a few months ago to include these flags by
default and have had no problems due to it at all.
-Carsten
On Friday, February 7, 2003, at
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
- -BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Would it be possible to add prebinding to the core fink files, so that
other packages can start adding it?
JP
On Friday, February 7, 2003, at 10:36 AM, David wrote:
- - --snip--
- -BEGIN PGP
On Friday, February 7, 2003, at 10:36 AM, David wrote:
Furthermore it is not as simple as simply adding --prebind. Many
pieces of software are not yet fixed to fully support prebinding, much
less proper relocation nor 'reentrancy'.
This is wrong.
-Ben
31 matches
Mail list logo
