Re: web Browsers (Re: gcc -O broken in CURRENT)
David O'Brien wrote:
| On Sat, Mar 16, 2002 at 06:05:13AM +0100, Dag-Erling Smorgrav wrote:
| Garrett Wollman [EMAIL PROTECTED] writes:
| What problems do you have with it?
|
| Slow. Eats memory. Crashes all the time. Does not save state
| between sessions. Does not render HTML 4 properly. Does not support
| CSS properly. Does not zoom. Does not display PNG properly.
| Incorrectly ignores cache-control headers on images. The list goes
|
| What brower available on FreeBSD does do all these things?
| Mozilla 0.9.8 was a disaster. Opera 6 is such a disaster that I went
| back to 5.05. [linux-]Netscape6 was marked BROKEN for a long time.
| konquor... well requires a lot of KDE bits to be installed.
Mozilla in all the variants I have tried is incapable of even
reliably downloading a file -- sometimes it works and sometimes
it turns it into complete junk, usually 20 Mbytes bigger than
the original. Useless.
Many of the secure sites I need to use (banks, universities,
etc.) refuse to allow access from any release 6 browser.
Linux-netscape-4.7{6,9} handles everything I want, sometimes
with minor glitches, but well enough. And I can leave it
running for weeks at a time without problems. That's good
enough for me. The alternatives that I have tried either don't
build or don't work and that means they are worse.
Greg
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: web Browsers (Re: gcc -O broken in CURRENT)
On Sun, 2002-03-17 at 09:56, Greg Black wrote: Joerg Wunsch wrote: | David O'Brien [EMAIL PROTECTED] wrote: | | Slow. Eats memory. Crashes all the time. Does not save state | between sessions. Does not render HTML 4 properly. Does not support | CSS properly. Does not zoom. Does not display PNG properly. | Incorrectly ignores cache-control headers on images. The list goes | | What brower available on FreeBSD does do all these things? | | Galeon. Yeah right. Galeon wouldn't even build on the last FreeBSD box I tried it on when somebody told me to try it. It compiles/works here like a charm, however, if you do have problems with it please send a problem report to maintainers ([EMAIL PROTECTED]) and we will try to help you. -Maxim signature.asc Description: This is a digitally signed message part
Re: web Browsers (Re: gcc -O broken in CURRENT)
Maxim Sobolev wrote: | On Sun, 2002-03-17 at 09:56, Greg Black wrote: | Yeah right. Galeon wouldn't even build on the last FreeBSD box | I tried it on when somebody told me to try it. | | It compiles/works here like a charm, however, if you do have problems | with it please send a problem report to maintainers ([EMAIL PROTECTED]) | and we will try to help you. Thank you for the offer. At present, I don't much care whether Galeon works or not -- linux-netscape works well enough for my needs. I did not submit a problem report about galeon because I had much more serious problems on the machine in question and they needed (and still need) to be resolved first. If I can ever get anybody interested in making a version of FreeBSD later than 4.3 work on my laptop, then I'll have another look at the Galeon question, as it was for the laptop that I was trying to get it going. As it stands, 4.4, 4.5 and current all have badly broken PCMCIA support which makes them unusable. Fortunately, 4.3 does not have this breakage, but it would be a waste of time to play with galeon on such an old release. Greg To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: web Browsers (Re: gcc -O broken in CURRENT)
* Greg Black ([EMAIL PROTECTED]) wrote: Joerg Wunsch wrote: Galeon. Yeah right. Galeon wouldn't even build on the last FreeBSD box I tried it on when somebody told me to try it. Tried Skipstone? Gecko based GTK browser. -- Thomas 'Freaky' Hurst - [EMAIL PROTECTED] - http://www.aagh.net/ - When in charge ponder, When in doubt mumble, When in trouble delegate. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
RE: gcc -O broken in CURRENT
okay... seems we are now out of topic... some arguments for a change some to retain the old custom (and in my opinion bootless stuff). I think later we'll need a survey for this and volunteers to do the work (if we want to do the change)... Alex are you still workin' for a patch? Jan -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Alexander Kabaev Sent: Thursday, March 14, 2002 5:26 PM To: Martin Blapp Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: Re: gcc -O broken in CURRENT Do you have a patch for this ? I do not fully understand the parts of GCC involved, so I need some time to verify my initial diagnosis and to create a patch. In other words - not yet :) -- Alexander Kabaev To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Alex are you still workin' for a patch? Yes, I am. But as I write before I am not familiar with this particular part of GCC at all, so I cannot give any estimates and even promize to produce a working patch. If some other more knowledgeable person is feeling like beating me to it, please feel free to do so. -- Alexander Kabaev To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
web Browsers (Re: gcc -O broken in CURRENT)
On Sat, Mar 16, 2002 at 06:05:13AM +0100, Dag-Erling Smorgrav wrote: Garrett Wollman [EMAIL PROTECTED] writes: What problems do you have with it? Slow. Eats memory. Crashes all the time. Does not save state between sessions. Does not render HTML 4 properly. Does not support CSS properly. Does not zoom. Does not display PNG properly. Incorrectly ignores cache-control headers on images. The list goes What brower available on FreeBSD does do all these things? Mozilla 0.9.8 was a disaster. Opera 6 is such a disaster that I went back to 5.05. [linux-]Netscape6 was marked BROKEN for a long time. konquor... well requires a lot of KDE bits to be installed. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: web Browsers (Re: gcc -O broken in CURRENT)
David O'Brien [EMAIL PROTECTED] writes: I use Opera 6. [...] Please try http://www.techiegold.com/ with Opera 6. No problem: http://www.ofug.org/~des/techiegold.png DES -- Dag-Erling Smorgrav - [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: web Browsers (Re: gcc -O broken in CURRENT)
Rich Wilson [EMAIL PROTECTED] writes: What about http://www.dice.com/jobsearch/index.html http://www.ofug.org/~des/dice.png (the error at the top is because my proxy blocks doubleclick) DES -- Dag-Erling Smorgrav - [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: web Browsers (Re: gcc -O broken in CURRENT)
David O'Brien [EMAIL PROTECTED] wrote: Slow. Eats memory. Crashes all the time. Does not save state between sessions. Does not render HTML 4 properly. Does not support CSS properly. Does not zoom. Does not display PNG properly. Incorrectly ignores cache-control headers on images. The list goes What brower available on FreeBSD does do all these things? Galeon. -- cheers, Jorg .-.-. --... ...-- -.. . DL8DTL http://www.sax.de/~joerg/NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On 2002-03-15 22:11, Kenneth Culver wrote:
#include rehash.h, see the thread we had on this a few weeks back on
-chat.
OK, I'll look, but I disagree... Mozilla runs flawlessly for me, and
renders much faster than netscape, however it loads really slow. Opera
runs nicely too, although it's linux only.
Netscape is still the only viable alternative for computers with less than
128 MB of physical memory, in places where other browsers can't be
installed for various reasons though. I've always meant to give galeon a
try, but I don't want to install half of Gnome to avoid downloading
Netscape's 15 MB binary. Downloading more than 15 MB of Gnome libs can
hardly be called 'better'.
Giorgos Keramidas FreeBSD Documentation Project
keramida@{freebsd.org,ceid.upatras.gr} http://www.FreeBSD.org/docproj/
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: web Browsers (Re: gcc -O broken in CURRENT)
Joerg Wunsch wrote: | David O'Brien [EMAIL PROTECTED] wrote: | | Slow. Eats memory. Crashes all the time. Does not save state | between sessions. Does not render HTML 4 properly. Does not support | CSS properly. Does not zoom. Does not display PNG properly. | Incorrectly ignores cache-control headers on images. The list goes | | What brower available on FreeBSD does do all these things? | | Galeon. Yeah right. Galeon wouldn't even build on the last FreeBSD box I tried it on when somebody told me to try it. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
RE: gcc -O broken in CURRENT
2) Bug is in os delivered gcc but not in port gcc. a) port has more or less patches / os gcc has been modified -- Didn't someone told they are the same? b) other options were set at compile time -- Why dont change to the same in the port? Leads it to a broken world? If the only difference is the lost of binary compatibility, i would say, ok... do it now and we'll need to compile or ports... SOme bugs are related to the FreeBSD use of setjmp/longjmp to do exception unwinding rather than using the DWARF primitives. When you change the toolchain, you change the exception unwinding code when you use the ports version. You also introduce incompatabilities with the installed libstdc++ library, which uses the setjmp/longjmp exception unwinding, which will be in conflict with any exception throwing/handling code compiled with the ports compiler that uses the DWARF2 version. The tests that show it working with the ports version do not test anything other than bare-bones operation, without testing code interoperability eith vendor libraries. Does that clear things up for you? A little bit... most of you argumenting about binary incompatibility for -stable. OK... no chance to do it there, its my opinion too. But why not doing it for current and using that most common dwarf unwinding now (for a later ia64 port it should be faster than setjump i think). Okay everything needs a recompile but this -current is current and not a production os. You're right that we need a patch for -stable. But if we take the approach for -current maybe we leave these problems behind us and following the path of the rank and file (using dwarf2) and making profit of their experience versus doing this ourself and creating patches. -Original Message- From: David O'Brien [mailto:[EMAIL PROTECTED]] Sent: Thursday, March 14, 2002 7:16 PM To: Jan Stocker Cc: Alexander Kabaev; Martin Blapp; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: Re: gcc -O broken in CURRENT On Thu, Mar 14, 2002 at 06:36:05PM +0100, Jan Stocker wrote: 2) Bug is in os delivered gcc but not in port gcc. a) port has more or less patches / os gcc has been modified -- Didn't someone told they are the same? Port has less patches. If you look at /usr/src/contrib/gcc/contrib/freebsd.h and /usr/src/contrib/gcc/contrib/i386/freebsd.h you will see how much things have to be modified because we support dual ELF/a.out [still]. This may be changed too for 5.0 shouldnt it? To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Jan Stocker wrote: [ ... DWARF vs. setjmp/longjmp ... ] A little bit... most of you argumenting about binary incompatibility for -stable. OK... no chance to do it there, its my opinion too. But why not doing it for current and using that most common dwarf unwinding now (for a later ia64 port it should be faster than setjump i think). Okay everything needs a recompile but this -current is current and not a production os. You're right that we need a patch for -stable. But if we take the approach for -current maybe we leave these problems behind us and following the path of the rank and file (using dwarf2) and making profit of their experience versus doing this ourself and creating patches. I guess it's possible to change over entirely. That would mean we would loase a.out support because the GNU tools are becoming incapable of supporting a.out (all machines we run on are Linux machines syndrome). If we really wanted to avoid problems like this in the future, we'd just scrap FreeBSD entirely, and go to Linux, a bit at a time, starting with ELF, then DWARF2 exceptions, and then the Linux ABI instead of the FreeBSD ABI, and then all of Linux, a piece at a time. PS: If I sound annoyed, it's because it's sometimes annoying to have your toolchain controlled by someone with an interest in a product that competes with yours; that works for people competing with Microsoft products on Microsoft platforms with a need to use Microsoft tools, and it applies to Cygnus being owned by RedHat and them controlling the FreeBSD tools. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Fri, Mar 15, 2002 at 01:37:39PM +0100, Jan Stocker wrote: A little bit... most of you argumenting about binary incompatibility for -stable. OK... no chance to do it there, its my opinion too. But why not doing it for current and using that most common dwarf unwinding now (for a There is no need to cause developers to go thru several ABI changes such that they cannot get their other FreeBSD development done. With GCC 3.1 a number of ABI changes will happen. Port has less patches. If you look at /usr/src/contrib/gcc/contrib/freebsd.h and /usr/src/contrib/gcc/contrib/i386/freebsd.h you will see how much things have to be modified because we support dual ELF/a.out [still]. This may be changed too for 5.0 shouldnt it? Why? I don't see how you justfied removing the functionality and I don't see how it is causing you any problems being there. -- -- David ([EMAIL PROTECTED]) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
I guess it's possible to change over entirely. That would mean we would loase a.out support because the GNU tools are becoming incapable of supporting a.out (all machines we run on are Linux machines syndrome). If we really wanted to avoid problems like this in the future, we'd just scrap FreeBSD entirely, and go to Linux, a bit at a time, starting with ELF, then DWARF2 exceptions, and then the Linux ABI instead of the FreeBSD ABI, and then all of Linux, a piece at a time. At the risk of being yelled at, I have a question: Why do we still need to support a.out? I know that a lot of people MIGHT still have some a.out binaries lying around, but FreeBSD's default binary format has been ELF for 3 or 4 years (Since 3.0-3.1 I believe). I'm not saying that we should entirely switch over to the regular gnu toolchain, but is it really necessary to keep supporting a.out? Just my $0.02 Ken PS: If I sound annoyed, it's because it's sometimes annoying to have your toolchain controlled by someone with an interest in a product that competes with yours; that works for people competing with Microsoft products on Microsoft platforms with a need to use Microsoft tools, and it applies to Cygnus being owned by RedHat and them controlling the FreeBSD tools. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Fri, Mar 15, 2002 at 04:54:59PM -0500, Kenneth Culver wrote: At the risk of being yelled at, I have a question: Why do we still need to support a.out? I know that a lot of people MIGHT still have some a.out binaries lying around, but FreeBSD's default binary format has been ELF for 3 or 4 years (Since 3.0-3.1 I believe). I'm not saying that we should entirely switch over to the regular gnu toolchain, but is it really necessary to keep supporting a.out? Just my $0.02 Rather than offer $0.02, send the patch. -- -- David ([EMAIL PROTECTED]) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
At the risk of being yelled at, I have a question: Why do we still need to support a.out? I know that a lot of people MIGHT still have some a.out binaries lying around, but FreeBSD's default binary format has been ELF for 3 or 4 years (Since 3.0-3.1 I believe). I'm not saying that we should entirely switch over to the regular gnu toolchain, but is it really necessary to keep supporting a.out? Just my $0.02 Rather than offer $0.02, send the patch. Well, I was just asking if it is necessary, I'd make a patch if there was interest. My mail was asking if there is interest. Ken To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Fri, Mar 15, 2002 at 05:26:37PM -0500, Kenneth Culver wrote: At the risk of being yelled at, I have a question: Why do we still need to support a.out? I know that a lot of people MIGHT still have some a.out ... Rather than offer $0.02, send the patch. Well, I was just asking if it is necessary, I'd make a patch if there was interest. My mail was asking if there is interest. We aren't changing this for GCC 2.95 in 5-CURRENT. PEROID. There is zero reason for subjecting users to this ABI change for what would be gained. If you want to do something productive, submit patches that Bmake GCC 3.1 (which move us to Dwarf2 unwinding as a product). -- -- David ([EMAIL PROTECTED]) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Kenneth Culver wrote: At the risk of being yelled at, I have a question: Why do we still need to support a.out? I know that a lot of people MIGHT still have some a.out binaries lying around, but FreeBSD's default binary format has been ELF for 3 or 4 years (Since 3.0-3.1 I believe). I'm not saying that we should entirely switch over to the regular gnu toolchain, but is it really necessary to keep supporting a.out? Just my $0.02 The switchover is not trivial. You're asking someone to do work for something that's not really valuable to them. There are certain boot code features that require the use of a.out kernels; this is less an issue than it was, but there were a number of things lost when we went to the new loader that are important for embedded environments. Cross-building for older platforms (not as big an issue, IMO). Other reasons I haven't even thought of yet 8-). -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
We aren't changing this for GCC 2.95 in 5-CURRENT. PEROID. There is zero reason for subjecting users to this ABI change for what would be gained. If you want to do something productive, submit patches that Bmake GCC 3.1 (which move us to Dwarf2 unwinding as a product). Oh ok, that's another story altogether... If nobody has gotten to it by the May timeframe I'll do it. I've been looking for a way to contribute to the FreeBSD project anyway. Right now I'm working nearly 40 hrs a week and going to college full-time, so I don't really have time to do anything else. Ken To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
At the risk of being yelled at, I have a question: Why do we still need to support a.out? I know that a lot of people MIGHT still have some a.out binaries lying around, but FreeBSD's default binary format has been ELF for 3 or 4 years (Since 3.0-3.1 I believe). I'm not saying that we should entirely switch over to the regular gnu toolchain, but is it really necessary to keep supporting a.out? Just my $0.02 The switchover is not trivial. You're asking someone to do work for something that's not really valuable to them. There are certain boot code features that require the use of a.out kernels; this is less an issue than it was, but there were a number of things lost when we went to the new loader that are important for embedded environments. Cross-building for older platforms (not as big an issue, IMO). Other reasons I haven't even thought of yet 8-). Yeah, I was just wondering if there were issues making us keep a.out stuff in FreeBSD aside from the I wanna run 2.2.x programs issue. Ken To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Kenneth Culver wrote: Other reasons I haven't even thought of yet 8-). Yeah, I was just wondering if there were issues making us keep a.out stuff in FreeBSD aside from the I wanna run 2.2.x programs issue. Linking with third party a.out libraries. Other reasons I haven't even thought of yet 8-). I can probably add one new reason per email indefinitely, if you want to insist... -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
[ Trim the CC's a bit ]
On Fri, Mar 15, 2002 at 04:00:08PM -0800 I heard the voice of
Terry Lambert, and lo! it spake thus:
Kenneth Culver wrote:
Other reasons I haven't even thought of yet 8-).
Yeah, I was just wondering if there were issues making us keep a.out stuff
in FreeBSD aside from the I wanna run 2.2.x programs issue.
Linking with third party a.out libraries.
Other reasons I haven't even thought of yet 8-).
(ttypa):{1078}% file /usr/local/lib/netscape/communicator-4.7.us.bin
/usr/local/lib/netscape/communicator-4.7.us.bin: FreeBSD/i386 compact
demand paged dynamically linked executable
Now, if you'd like to talk Netscape into building a version intended for
a version of FreeBSD newer than, say, 3 years, 3.5 months (approximately)
old...
--
Matthew Fuller (MF4839) |[EMAIL PROTECTED]
Unix Systems Administrator |[EMAIL PROTECTED]
Specializing in FreeBSD |http://www.over-yonder.net/
The only reason I'm burning my candle at both ends, is because I
haven't figured out how to light the middle yet
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
[Cc's trimmed]
Kenneth Culver wrote:
| (ttypa):{1078}% file /usr/local/lib/netscape/communicator-4.7.us.bin
| /usr/local/lib/netscape/communicator-4.7.us.bin: FreeBSD/i386 compact
|demand paged dynamically linked executable
|
| Now, if you'd like to talk Netscape into building a version intended for
| a version of FreeBSD newer than, say, 3 years, 3.5 months (approximately)
| old...
|
| I didn't realize anyone still used netscape 4.x. It's so disgustingly
| unstable and slow.
It's less slow and much more reliable than mozilla and remains
the only available browser that can access most of the sites I
need to access.
Greg
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
It's less slow and much more reliable than mozilla and remains the only available browser that can access most of the sites I need to access. That's odd, I've never had any mozilla problems. All I know is that it doesn't crash on sites that Netscape crashes on (anything java) and for me it runs much faster than netscape. It loads slower, but renders pages much faster, and I tend to load my browser once per day, and just leave it on. Ken To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
That's odd, I've never had any mozilla problems. All I know is that it doesn't crash on sites that Netscape crashes on (anything java) and for me it runs much faster than netscape. It loads slower, but renders pages much faster, and I tend to load my browser once per day, and just leave it on. Anyway, this is way OT, so that was my last message about it. Ken To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Sat, 16 Mar 2002, Greg Black wrote:
[Cc's trimmed]
Kenneth Culver wrote:
| (ttypa):{1078}% file /usr/local/lib/netscape/communicator-4.7.us.bin
| /usr/local/lib/netscape/communicator-4.7.us.bin: FreeBSD/i386 compact
|demand paged dynamically linked executable
|
| Now, if you'd like to talk Netscape into building a version intended for
| a version of FreeBSD newer than, say, 3 years, 3.5 months (approximately)
| old...
|
| I didn't realize anyone still used netscape 4.x. It's so disgustingly
| unstable and slow.
It's less slow and much more reliable than mozilla and remains
the only available browser that can access most of the sites I
need to access.
and I use it's mail reader a lot..
Greg
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-hackers in the body of the message
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Friday 15 March 2002 08:53 pm, Kenneth Culver wrote:
| (ttypa):{1078}% file /usr/local/lib/netscape/communicator-4.7.us.bin
| /usr/local/lib/netscape/communicator-4.7.us.bin: FreeBSD/i386 compact
|demand paged dynamically linked executable
|
| Now, if you'd like to talk Netscape into building a version intended for
| a version of FreeBSD newer than, say, 3 years, 3.5 months (approximately)
| old...
|
| I didn't realize anyone still used netscape 4.x. It's so disgustingly
| unstable and slow.
Well, the linux-netscape 4 is the only browser I know that can handle Java
pages on FreeBSD.
Are there others?
If you mean the FreeBSD-native netscape 4.x; yes, it's perfectly silly to run
*that*.
|
| Ken
|
|
| To Unsubscribe: send mail to [EMAIL PROTECTED]
| with unsubscribe freebsd-hackers in the body of the message
--
Brian T. Schellenberger . . . . . . . [EMAIL PROTECTED] (work)
Brian, the man from Babble-On . . . . [EMAIL PROTECTED] (personal)
ME -- http://www.babbleon.org
http://www.eff.org -- GOOD GUYS -- http://www.programming-freedom.org
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
RE: gcc -O broken in CURRENT
Err--the linux netscape 6 runs fine. It's also quite slow to load, but
so far appears to be rather robust.
Cheers,
Ben
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]] On Behalf Of Brian
T.Schellenberger
Sent: Friday, March 15, 2002 10:41 PM
To: Kenneth Culver; Matthew D. Fuller
Cc: Terry Lambert; [EMAIL PROTECTED]; [EMAIL PROTECTED];
[EMAIL PROTECTED]
Subject: Re: gcc -O broken in CURRENT
On Friday 15 March 2002 08:53 pm, Kenneth Culver wrote:
| (ttypa):{1078}% file
/usr/local/lib/netscape/communicator-4.7.us.bin
| /usr/local/lib/netscape/communicator-4.7.us.bin:
FreeBSD/i386 compact
|demand paged dynamically linked executable
|
| Now, if you'd like to talk Netscape into building a
version intended for
| a version of FreeBSD newer than, say, 3 years, 3.5 months
(approximately)
| old...
|
| I didn't realize anyone still used netscape 4.x. It's so
disgustingly
| unstable and slow.
Well, the linux-netscape 4 is the only browser I know that
can handle Java
pages on FreeBSD.
Are there others?
If you mean the FreeBSD-native netscape 4.x; yes, it's
perfectly silly to run
*that*.
|
| Ken
|
|
| To Unsubscribe: send mail to [EMAIL PROTECTED]
| with unsubscribe freebsd-hackers in the body of the message
--
Brian T. Schellenberger . . . . . . . [EMAIL PROTECTED] (work)
Brian, the man from Babble-On . . . . [EMAIL PROTECTED] (personal)
ME -- http://www.babbleon.org
http://www.eff.org -- GOOD GUYS --
http://www.programming-freedom.org
To Unsubscribe: send
mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
[Unnecessary carbon copies trimmed.] On Fri, 15 Mar 2002 22:41:26 -0500, Brian T.Schellenberger [EMAIL PROTECTED] said: If you mean the FreeBSD-native netscape 4.x; yes, it's perfectly silly to run *that*. I don't see anything silly about it. It works with all the Web sites I care about (which is more than I can say for either mozilla or konqueror). It has a sensible (i.e., non-Windows-oriented) user interface. It has a few annoying bugs, but none of them are sufficiently problematic to keep me from getting my work done. What problems do you have with it? -GAWollman To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Brian T.Schellenberger wrote: Well, the linux-netscape 4 is the only browser I know that can handle Java pages on FreeBSD. Are there others? If you mean the FreeBSD-native netscape 4.x; yes, it's perfectly silly to run *that*. 4.7 does this just fine, if you don't move the mouse until it's done loading. That restriction only exists for image mapped interfaces, where the Java GIF loader is used, and then only if the image loading is not serialized by the page design. Note that only Solaris, Windows, and Linux, all of which assume (incorrectly) that a threaded process that is preempted involuntarily will resume executin in the thread that was runningat preemption time, handle the concurrent image loading correctly, if you move the mouse or otherwise cause input to the browser before the loads are complete. Basically, it's bad threading assumptions, and it's fixed in a more recent version, if you can find one. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Garrett Wollman [EMAIL PROTECTED] writes: What problems do you have with it? Slow. Eats memory. Crashes all the time. Does not save state between sessions. Does not render HTML 4 properly. Does not support CSS properly. Does not zoom. Does not display PNG properly. Incorrectly ignores cache-control headers on images. The list goes on... DES -- Dag-Erling Smorgrav - [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Dag-Erling Smorgrav wrote: Garrett Wollman [EMAIL PROTECTED] writes: What problems do you have with it? Slow. Eats memory. Crashes all the time. Does not save state between sessions. Does not render HTML 4 properly. Does not support CSS properly. Does not zoom. Does not display PNG properly. Incorrectly ignores cache-control headers on images. The list goes on... If you use a real network socket instead of the shared memory extension, it won't eat memory. THis lost memory due to it instancing regions for which it loses the reference counts; it's arguably a resource tracking bug in the X server, actually, since they are associated with windows that go away. I admit, this is an annoying one... -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Per thread exception stacks? THat's where I'd look... Hmm, good point. The programms that crashed were all threaded ... Martin To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
This is a case of exception context register getting clobbered in shared library function call. GCC does not reload it when needed and this ultimately leads to semi-random word in program memory decremented by the __cp_pop_exception function. The bug is only triggered under very specific circumstances involving inline functions and nested degenerate exception handlers, that's why it existed unnoticed for quite some time. On Wed, 13 Mar 2002 22:53:48 -0800 Terry Lambert [EMAIL PROTECTED] wrote: M. Warner Losh wrote: In message: [EMAIL PROTECTED] Ed Hall [EMAIL PROTECTED] writes: : Exception-handling is broken with -O in -stable, and has been for years.: FreeBSD is one of the few systems that use setjmp/longjmp stack unwinds: to implement exceptions, so when the GCC folks broke that path, it was: never fixed. There are supposedly patches floating around that fix the: problem, but they either didn't work as advertised or the ball got dropped. H, C++ exceptions work in -stable with -O and have for at least a year. At least they are working for us in our environment. What's busted? Per thread exception stacks? THat's where I'd look... -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
This is a case of exception context register getting clobbered in shared library function call. GCC does not reload it when needed and this ultimately leads to semi-random word in program memory decremented by the __cp_pop_exception function. The bug is only triggered under very specific circumstances involving inline functions and nested degenerate exception handlers, that's why it existed unnoticed for quite some time. On Wed, 13 Mar 2002 22:53:48 -0800 Terry Lambert [EMAIL PROTECTED] wrote: M. Warner Losh wrote: In message: [EMAIL PROTECTED] Ed Hall [EMAIL PROTECTED] writes: : Exception-handling is broken with -O in -stable, and has been for years.: FreeBSD is one of the few systems that use setjmp/longjmp stack unwinds: to implement exceptions, so when the GCC folks broke that path, it was: never fixed. There are supposedly patches floating around that fix the: problem, but they either didn't work as advertised or the ball got dropped. H, C++ exceptions work in -stable with -O and have for at least a year. At least they are working for us in our environment. What's busted? Per thread exception stacks? THat's where I'd look... -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Hi, This is a case of exception context register getting clobbered in shared library function call. GCC does not reload it when needed and this ultimately leads to semi-random word in program memory decremented by the __cp_pop_exception function. The bug is only triggered under very specific circumstances involving inline functions and nested degenerate exception handlers, that's why it existed unnoticed for quite some time. Do you have a patch for this ? Martin To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
In message: [EMAIL PROTECTED] Terry Lambert [EMAIL PROTECTED] writes: : M. Warner Losh wrote: : In message: [EMAIL PROTECTED] : Ed Hall [EMAIL PROTECTED] writes: : : Exception-handling is broken with -O in -stable, and has been for years. : : FreeBSD is one of the few systems that use setjmp/longjmp stack unwinds : : to implement exceptions, so when the GCC folks broke that path, it was : : never fixed. There are supposedly patches floating around that fix the : : problem, but they either didn't work as advertised or the ball got dropped. : : H, C++ exceptions work in -stable with -O and have for at least a : year. At least they are working for us in our environment. What's : busted? : : Per thread exception stacks? THat's where I'd look... Yes, that works. Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Do you have a patch for this ? I do not fully understand the parts of GCC involved, so I need some time to verify my initial diagnosis and to create a patch. In other words - not yet :) -- Alexander Kabaev To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
RE: gcc -O broken in CURRENT
So now i am a little bit confused... State of the art: 1) Bug is in -stable and -current -- This means possible patches only in -current arent responsible for this behaviour 2) Bug is in os delivered gcc but not in port gcc. a) port has more or less patches / os gcc has been modified -- Didn't someone told they are the same? b) other options were set at compile time -- Why dont change to the same in the port? Leads it to a broken world? If the only difference is the lost of binary compatibility, i would say, ok... do it now and we'll need to compile or ports... -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Alexander Kabaev Sent: Thursday, March 14, 2002 5:26 PM To: Martin Blapp Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: Re: gcc -O broken in CURRENT Do you have a patch for this ? I do not fully understand the parts of GCC involved, so I need some time to verify my initial diagnosis and to create a patch. In other words - not yet :) -- Alexander Kabaev To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Do you have a small, reproducible test case? Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
in a related tangential note, i recently found (out of sheer irritation) in less than an hour that several (including the latest) versions of GCC -O and -O2 failed the paranoia test in different ways, to wit: gcc -o paranoia paranoia.c [paranoia output elided] The number of DEFECTs discovered = 1. The number of FLAWs discovered = 1. gcc -O2 -o paranoia paranoia.c [paranoia output elided] The number of FAILUREs encountered = 4. The number of SERIOUS DEFECTs discovered = 4. The number of DEFECTs discovered = 2. The number of FLAWs discovered = 2. i assume everyone knows about kahan and paranoia. if not see netlib. oz --- a technology is indistinguishable from | electric: [EMAIL PROTECTED] its implementation. -- Marshall Rose | or 905 415 2878 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Thu, Mar 14, 2002 at 06:36:05PM +0100, Jan Stocker wrote: 2) Bug is in os delivered gcc but not in port gcc. a) port has more or less patches / os gcc has been modified -- Didn't someone told they are the same? Port has less patches. If you look at /usr/src/contrib/gcc/contrib/freebsd.h and /usr/src/contrib/gcc/contrib/i386/freebsd.h you will see how much things have to be modified because we support dual ELF/a.out [still]. b) other options were set at compile time -- Why dont change to the same in the port? I am willing to -- the gcc295 port isn't used very much now AFAIK. However, it will probably be once 5-CURRENT moves to a newer version. The FSF GCC people had settings in the i386/freebsd.h file I did not agree with, but it would have been too much pain to change them in the FSF 2.95 release branch. I am willing (and may have to anyway), make the port more agree with the FreeBSD system compiler. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
2) Bug is in os delivered gcc but not in port gcc. a) port has more or less patches / os gcc has been modified -- Didn't someone told they are the same? GCC from ports uses DWARF2 exception unwinding while GCC in src tree uses sjlj exceptions. The exception handling code generated by these two compilers is very different as a result. b) other options were set at compile time -- Why dont change to the same in the port? Leads it to a broken world? If the only difference is the lost of binary compatibility, i would say, ok... do it now and we'll need to compile or ports... Pretty much each and every C++ binary and shared library will have to be recompiled. Massive binary compatibility breakage is not an option for -STABLE, one can hope. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Thu, Mar 14, 2002 at 12:59:31PM -0500, ozan s. yigit wrote: in a related tangential note, i recently found (out of sheer irritation) in less than an hour that several (including the latest) versions of GCC -O and -O2 failed the paranoia test in different ways, to wit: gcc -o paranoia paranoia.c The number of DEFECTs discovered = 1. The number of FLAWs discovered = 1. gcc -O2 -o paranoia paranoia.c The number of FAILUREs encountered = 4. The number of SERIOUS DEFECTs discovered = 4. The number of DEFECTs discovered = 2. The number of FLAWs discovered = 2. i assume everyone knows about kahan and paranoia. if not see netlib. Add the -ffloat-store flag to your compilation flags (or add -msoft-float). No failures, defects nor flaws have been discovered. Rounding appears to conform to the proposed IEEE standard P754, except for possibly Double Rounding during Gradual Underflow. The arithmetic diagnosed appears to be Excellent! END OF TEST. -- Steve To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Thu, Mar 14, 2002 at 01:20:51PM -0500, Alexander Kabaev wrote: b) other options were set at compile time -- Why dont change to the same in the port? Leads it to a broken world? If the only difference is the lost of binary compatibility, i would say, ok... do it now and we'll need to compile or ports... Pretty much each and every C++ binary and shared library will have to be recompiled. Massive binary compatibility breakage is not an option for -STABLE, one can hope. No it is not an option for -STABLE. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Add the -ffloat-store flag to your compilation flags (or add -msoft-float). that really means for this compiler on certain platforms, you can have slow and correct or fast and incorrect, but NOT fast and correct. oz --- freedom has a mental cost. -- peter roosen-runge To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
ozan s. yigit writes: Add the -ffloat-store flag to your compilation flags (or add -msoft-float). that really means for this compiler on certain platforms, you can have slow and correct or fast and incorrect, but NOT fast and correct. Actually, if -ffloat-store is the solution, the problem arises because you have fast and *too* correct. -- Raymond WikerMail: [EMAIL PROTECTED] Senior Software Engineer Web: http://www.fast.no/ Fast Search Transfer ASA Phone: +47 23 01 11 60 P.O. Box 1677 Vika Fax: +47 35 54 87 99 NO-0120 Oslo, NORWAY Mob: +47 48 01 11 60 Try FAST Search: http://alltheweb.com/ To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Wed, Mar 13, 2002 at 10:24:20PM +0100, Martin Blapp wrote: We are using a set of patches that were part of gcc 2.95.3_test3. Do you have a sample program in which exceptions are still broken on FreeBSD 4.5? cd /usr/ports/devel/stlport make install cd work/STL*/test/eh add -O to gcc-freebsd.mk gmake -f gcc-freebsd.mk clean gmake -f gcc-freebsd.mk and see what happens ... This is not a small, [relatively] simple example program. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Thu, Mar 14, 2002 at 07:50:38PM +0100, Raymond Wiker wrote: ozan s. yigit writes: Add the -ffloat-store flag to your compilation flags (or add -msoft-float). that really means for this compiler on certain platforms, you can have slow and correct or fast and incorrect, but NOT fast and correct. Actually, if -ffloat-store is the solution, the problem arises because you have fast and *too* correct. If the gcc manual is to be believed, then yes you are correct. If you really want to investigate FreeBSD FP/math capabilities search for UCBTEST or visit www.cs.berkeley.edu/~jhauser/arithmetic/TestFloat.html -- Steve To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
float [was Re: gcc -O broken in CURRENT]
If you really want to investigate FreeBSD FP/math capabilities search for UCBTEST or visit www.cs.berkeley.edu/~jhauser/arithmetic/TestFloat.html cool! thanks for the pointer. oz --- gag reflex is an essential part of computing. -- anon To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Jan Stocker wrote: So now i am a little bit confused... State of the art: 1) Bug is in -stable and -current -- This means possible patches only in -current arent responsible for this behaviour Unless they were MFC'ed to -STABLE. THis is why you generally should compare -RELEASE versions, not -STABLE versions, since -STABLE versions are moving targets and -RELEASE versions are not. 2) Bug is in os delivered gcc but not in port gcc. a) port has more or less patches / os gcc has been modified -- Didn't someone told they are the same? b) other options were set at compile time -- Why dont change to the same in the port? Leads it to a broken world? If the only difference is the lost of binary compatibility, i would say, ok... do it now and we'll need to compile or ports... SOme bugs are related to the FreeBSD use of setjmp/longjmp to do exception unwinding rather than using the DWARF primitives. When you change the toolchain, you change the exception unwinding code when you use the ports version. You also introduce incompatabilities with the installed libstdc++ library, which uses the setjmp/longjmp exception unwinding, which will be in conflict with any exception throwing/handling code compiled with the ports compiler that uses the DWARF2 version. The tests that show it working with the ports version do not test anything other than bare-bones operation, without testing code interoperability eith vendor libraries. Does that clear things up for you? -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Thu, 14 Mar 2002, ozan s. yigit wrote: Add the -ffloat-store flag to your compilation flags (or add -msoft-float). that really means for this compiler on certain platforms, you can have slow and correct or fast and incorrect, but NOT fast and correct. I think fast and correct is impossible on i386's. Correct requires assignments and casts to discard any extra precision, and the fastest way to implement this is probably to store to memory and reload. The -ffloat-store kludge only does a subset of the necessary conversions. Doing them all would be slower and correct, which is why gcc doesn't do them. C90 can be read as permitting this incorrectness, but C99 doesn't permit it. Bruce To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Hi, Here are my test news. The -O bug doesn't happen with gcc295 from ports ! I tried all these FLAGS, but noone of them was creating the problems we see with -O : Optimization Options -fcaller-saves -fcse-follow-jumps -fcse-skip-blocks -fdelayed-branch -felide-constructors -fexpensive-optimizations -ffast-math -ffloat-store -fforce-addr -fforce-mem -finline-functions -fkeep-inline-functions -fmemoize-lookups -fno-default-inline -fno-defer-pop -fno-function-cse -fno-inline -fno-peephole -fomit-frame-pointer -frerun-cse-after-loop -fschedule-insns -fschedule-insns2 -fstrength-reduce -fthread-jumps -funroll-all-loops So what does -O exactly ? Martin To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
STABLE is broken too, but in a different manner. I just added -O and
then this happened.
[algo] :testing inplace_merge #1() (weak) ... eh_test in free(): warning: junk
pointer, too high to make sense
eh_test in free(): warning: junk pointer, too high to make sense
eh_test in free(): warning: junk pointer, too high to make sense
eh_test in free(): warning: junk pointer, too high to make sense
eh_test in free(): warning: junk pointer, too high to make sense
eh_test in free(): warning: junk pointer, too high to make sense
eh_test in free(): warning: junk pointer, too high to make sense
eh_test in free(): warning: junk pointer, too high to make sense
eh_test in free(): warning: junk pointer, too high to make sense
eh_test in free(): warning: junk pointer, too high to make sense
eh_test in free(): warning: junk pointer, too high to make sense
eh_test in free(): warning: junk pointer, too high to make sense
eh_test in free(): warning: junk pointer, too high to make sense
eh_test in free(): warning: junk pointer, too high to make sense
eh_test in free(): warning: modified (chunk-) pointer
# gdb eh_test eh_test.core
#0 0x806630b in void _STL::inplace_mergeSortClass * (__first=0xbfbff0ac,
__middle=0xbfbff23c, __last=0xbfbff3cc) at test_algo.cpp:58
58{
(gdb) bt
#0 0x806630b in void _STL::inplace_mergeSortClass * (__first=0xbfbff0ac,
__middle=0xbfbff23c, __last=0xbfbff3cc) at test_algo.cpp:58
#1 0x8066e8a in void WeakCheckSortBuffer, test_inplace_merge_1
(v=@0xbfbff83c, op=@0xbfbff434, max_iters=200) at test_algo.cpp:216
#2 0x804b41e in test_algo () at test_algo.cpp:248
#3 0x8049e37 in main (argc=3, argv=0xbfbffbe0) at main.cpp:275
#4 0x8049759 in _start ()
Martin
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
I removed now #undef DEFAULT_VTABLE_THUNKS and set again #define DWARF2_UNWIND_INFO 1 in the port. The -O tests still succeeded. All cpp* files are the same in the port and our system compilers. And ideas and pointers which subsystems I could test for this breakage ? Martin To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
Exception-handling is broken with -O in -stable, and has been for years.
FreeBSD is one of the few systems that use setjmp/longjmp stack unwinds
to implement exceptions, so when the GCC folks broke that path, it was
never fixed. There are supposedly patches floating around that fix the
problem, but they either didn't work as advertised or the ball got dropped.
This problem should exist in -current since I think FreeBSD finally drops
setjmp/longjmp stack unwinds and goes to dwarf 2 unwinds, which do work
(and which are used by the GCC 2.95 port, which also works but which
isn't compatible with /usr/lib/libstdc++.{a,so}).
This issue is why Yahoo! has to use its own build of GCC, and I doubt
we're the only ones...
-Ed
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Wed, Mar 13, 2002 at 12:15:34PM -0800, Ed Hall wrote: Exception-handling is broken with -O in -stable, and has been for years. FreeBSD is one of the few systems that use setjmp/longjmp stack unwinds to implement exceptions, so when the GCC folks broke that path, it was never fixed. There are supposedly patches floating around that fix the problem, but they either didn't work as advertised or the ball got dropped. We are using a set of patches that were part of gcc 2.95.3_test3. Do you have a sample program in which exceptions are still broken on FreeBSD 4.5? To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
We are using a set of patches that were part of gcc 2.95.3_test3. Do you have a sample program in which exceptions are still broken on FreeBSD 4.5? cd /usr/ports/devel/stlport make install cd work/STL*/test/eh add -O to gcc-freebsd.mk gmake -f gcc-freebsd.mk clean gmake -f gcc-freebsd.mk and see what happens ... To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Wed, Mar 13, 2002 at 02:08:55PM +0100, Martin Blapp wrote: I removed now #undef DEFAULT_VTABLE_THUNKS and set again #define DWARF2_UNWIND_INFO 1 in the port. The -O tests still succeeded. All cpp* files are the same in the port and our system compilers. And ideas and pointers which subsystems I could test for this breakage ? Did you pursue my suggestion of comparing recent patches in the port and in the source tree? Kris msg36040/pgp0.pgp Description: PGP signature
Re: gcc -O broken in CURRENT
Hi Kris, Did you pursue my suggestion of comparing recent patches in the port and in the source tree? Easy to say, hard to do. STABLE is broken as current is, and it seems that 4.4 and 4.3 are also broken for the STLport test. This is a very difficult thing to do for someone that does not know gcc internals. Impossible for me. I don't have the resources (time) and knowledge (compiler coding) to do this. I can only state that: - plain gcc without patches works - gcc295 from ports works - gcc is STABLE and CURRENT is broken - It's not the dewarf unwinding. Martin To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Wed, Mar 13, 2002 at 11:42:46PM +0100, Martin Blapp wrote: Hi Kris, Did you pursue my suggestion of comparing recent patches in the port and in the source tree? Easy to say, hard to do. STABLE is broken as current is, and it seems that 4.4 and 4.3 are also broken for the STLport test. That gives you MORE information: look for patches to the gcc directory which have been MFCed. This is a very difficult thing to do for someone that does not know gcc internals. You don't have to understand the changes, just look at the cvs logs for the past few months, and try backing out revisions to see if it fixes things, or at least identify a list of possible changes which others can test. Kris msg36042/pgp0.pgp Description: PGP signature
Re: gcc -O broken in CURRENT
Kris, fixes things, or at least identify a list of possible changes which others can test. How can I compile gcc without doing a make world ? Martin To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
cd /usr/src/gnu/usr.bin/cc make make install On Wed, 13 Mar 2002, Martin Blapp wrote: Kris, fixes things, or at least identify a list of possible changes which others can test. How can I compile gcc without doing a make world ? Martin To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Wed, Mar 13, 2002 at 11:49:52PM +0100, Martin Blapp wrote: Kris, fixes things, or at least identify a list of possible changes which others can test. How can I compile gcc without doing a make world ? cd /usr/src/gnu/usr.bin/cc make all Kris msg36045/pgp0.pgp Description: PGP signature
Re: gcc -O broken in CURRENT
I wrote: : This problem should exist in -current since I think FreeBSD finally drops ^^ That should be shouldn't. I shouldn't post in a hurry (like I'm doing now). -Ed To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
In message: [EMAIL PROTECTED] Ed Hall [EMAIL PROTECTED] writes: : Exception-handling is broken with -O in -stable, and has been for years. : FreeBSD is one of the few systems that use setjmp/longjmp stack unwinds : to implement exceptions, so when the GCC folks broke that path, it was : never fixed. There are supposedly patches floating around that fix the : problem, but they either didn't work as advertised or the ball got dropped. H, C++ exceptions work in -stable with -O and have for at least a year. At least they are working for us in our environment. What's busted? Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
M. Warner Losh wrote: In message: [EMAIL PROTECTED] Ed Hall [EMAIL PROTECTED] writes: : Exception-handling is broken with -O in -stable, and has been for years. : FreeBSD is one of the few systems that use setjmp/longjmp stack unwinds : to implement exceptions, so when the GCC folks broke that path, it was : never fixed. There are supposedly patches floating around that fix the : problem, but they either didn't work as advertised or the ball got dropped. H, C++ exceptions work in -stable with -O and have for at least a year. At least they are working for us in our environment. What's busted? Per thread exception stacks? THat's where I'd look... -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: gcc -O broken in CURRENT
On Tue, Mar 12, 2002 at 01:49:02AM +0100, Martin Blapp wrote: Hi all, Here are my test news. The -O bug doesn't happen with gcc295 from ports ! Since this problem was apparently introduced recently, can you check the commits against the gcc code in -current with the patches to the port? Presumably one or more commits have not yet appeared as patches, and one of those is likely to be the cause. Kris msg35991/pgp0.pgp Description: PGP signature
RE: gcc -O broken in CURRENT
Hello... isn't anyone workin on this prob? Or do i have a leak in my mailbox? Jan -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Jan Stocker Sent: Wednesday, March 06, 2002 9:15 AM To: Martin Blapp; [EMAIL PROTECTED] Subject: RE: gcc -O broken in CURRENT This can be reproduced on my system and it may be a big problem. I installed everthing from source with -O set in make.conf and some proggys running fine under 4.4 wont unter -current. Recompiling the proggys without -O dont fix it but maybe it is caused by some libs... Is anyone examining this problem? Jan -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Martin Blapp Sent: Monday, March 04, 2002 9:02 PM To: [EMAIL PROTECTED] Subject: gcc -O broken in CURRENT Hi all, Unfortunatly I have a example from the ports, needed by OpenOffice port (work in progress) cd /usr/ports/devel/stlport/ make install cd /usr/ports/devel/stlport/work/STLport-4.5.3/test/eh gmake -f gcc-freebsd.mak [vector] :testing n-size constructor (const) ... 95 try successful [vector] :testing pointer range constructor (const) ... Bus error - core dumped Then remove the three -O from gcc-freebsd.mak and run it again. You will see that all tests are successfully done. [...] [hash_multiset] :testing default constructor (const) ... 2 try successful [hash_multiset] :testing iterator range n-size constructor (const) ... 127 try successful [hash_multiset] :testing copy constructor (const) ... 54 try successful [hash_multiset] :testing assignment operator (weak) ... 53 try successful [...] [string] :testing copy constructor (const) ... 2 try successful [string] :testing assignment operator (weak) ... 1 try successful EH test : Done Martin Martin Blapp, [EMAIL PROTECTED] [EMAIL PROTECTED] -- ImproWare AG, UNIXSP ISP, Zurlindenstrasse 29, 4133 Pratteln, CH Phone: +41 061 826 93 00: +41 61 826 93 01 PGP Fingerprint: B434 53FC C87C FE7B 0A18 B84C 8686 EF22 D300 551E -- To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
RE: gcc -O broken in CURRENT
Hi all, Here are my test news. The -O bug doesn't happen with gcc295 from ports ! Previously I had stated before, the gcc295 from ports did not work too. but it seems that that was a user error :-) /usr/ports/devel/stlport (and the tests test/eh) can be succesfully be made. My staroffice build has survived the segfaulting part now in saxparser. Martin Martin Blapp, [EMAIL PROTECTED] [EMAIL PROTECTED] -- ImproWare AG, UNIXSP ISP, Zurlindenstrasse 29, 4133 Pratteln, CH Phone: +41 061 826 93 00: +41 61 826 93 01 PGP Fingerprint: B434 53FC C87C FE7B 0A18 B84C 8686 EF22 D300 551E -- To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
gcc -O broken in CURRENT
Hi all, Unfortunatly I have a example from the ports, needed by OpenOffice port (work in progress) cd /usr/ports/devel/stlport/ make install cd /usr/ports/devel/stlport/work/STLport-4.5.3/test/eh gmake -f gcc-freebsd.mak [vector] :testing n-size constructor (const) ... 95 try successful [vector] :testing pointer range constructor (const) ... Bus error - core dumped Then remove the three -O from gcc-freebsd.mak and run it again. You will see that all tests are successfully done. [...] [hash_multiset] :testing default constructor (const) ... 2 try successful [hash_multiset] :testing iterator range n-size constructor (const) ... 127 try successful [hash_multiset] :testing copy constructor (const) ... 54 try successful [hash_multiset] :testing assignment operator (weak) ... 53 try successful [...] [string] :testing copy constructor (const) ... 2 try successful [string] :testing assignment operator (weak) ... 1 try successful EH test : Done Martin Martin Blapp, [EMAIL PROTECTED] [EMAIL PROTECTED] -- ImproWare AG, UNIXSP ISP, Zurlindenstrasse 29, 4133 Pratteln, CH Phone: +41 061 826 93 00: +41 61 826 93 01 PGP Fingerprint: B434 53FC C87C FE7B 0A18 B84C 8686 EF22 D300 551E -- To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
