Heck, I just compiled up Quakeforge last night to try it out on my little laptop here (Athlon XP-M 1600+ (1400Mhz), 352MB, 32MB shared video ram, S3 Savage video chipset) and was amused to see I was getting between 40 and 75 fps at whatever the heck the default resolution was. Well inside the 'Playable' range.
I'm disappointed that Id refuses to offer a "data files for open-sourced games" (i.e. Doom 1, Doom 2, Quake 1, Quake 2) disk or download at a reasonable price - I asked about it and got back a shocked "Oh, no, we'd NEVER do something like THAT!" as though I'd asked them if they'd consider auctioning off their genitals on E-bay or something...Since they're charging "new Quake I box" rates last time I checked ($30+) for a simple download of the original Quake disk alone, I just can't quite justify the cost at the moment. But I'd pay that much for permission to use (no support, even no media [i.e. download disk image only]) the quake 1, quake 2, and Doom 1,2 datasets to go with Quakeforge and the various DOOM ports...
That being the case - anybody know any good free "total conversion" datasets that don't require the original ID datasets? OpenQuartz appeared to be working on one, but they haven't had an update in over 2 years...
I find it's especially useful for any application that involves "talking to a service over the network" or dealing with text-like (including html/xml) files and/or relatively simple GD-compatible graphics file manipulation (generating graphs, maps, etc.), whether it's a database server, a webserver, an ftp server, a socket, etc. etc.
My own current personal project involves figuring out the "TIGER/Line" geographic data from the US Census bureau so that I can render the features at will over other existing map graphics and turn them into maps for GPSDrive. I've already got PHP scripts generating "USGS Topographic maps or TIGER/Line [1998 data rendered from the census bureau site] features rendered atop USGS Terraserver aerial photos" for GPSDrive...
I've also considered doing a Kismet client class for PHP. Not real useful for old-fashioned web use, but could be handy for CLI (or PHP-GTK or whatever).
What? And settle for UN-OPTIMIZED binaries like some plebe? Forget that!
Besides, I had to grab the CVS from the DRI branch to get S3/SavageDDR (It's a laptop, I didn't have much choice...) direct-rendering support. Why not compile it up with "-O2 -march=athlon-xp -mmmx -m3dnow -msse -mfpmath=sse,387" while I'm at it?...
Yes, I'm also running on bleeding-edge CVS compile of KDE 3.3-post-alpha as well (working fine for me, incidentally...)
Think of it as the civilized nerd equivalent of peeing on the computer to mark it as yours. Except that you don't ruin the computer and/or get painfully shocked in the process (uh, you know, or so I've HEARD....)
There is a lot of talk lately where I'm employed about the "need" for "Electronic Document Management"(tm), but nobody seems to know EXACTLY what we need or want from it, and nobody's really put any though thus far into working on some user requirements specs for it so that we can evaluate what we really DO need.
Thus far, proprietary EDM sites seem to be filled to the brim with low-content fluff but little REAL information about exactly what you get from them. Even presuming they DO have truly worth-the-price features, it seems difficult to really pin down what those features really ARE...
I think that's part of the reason they can get away with such egregiously high license fees.
Wonder if this one would make a good "Ask Slashdot"?...
I've been wondering lately myself about this. Just how inexpensively can one put together a reasonably-useful yet reasonably-standard computer off-the-shelf? (As opposed to "go on E-bay and see if you can find X" or "if you can find THIS version of THIS hardware, you can hack it to put Linux in it" and so on...)
It'd be nifty to have some good sources for very inexpensive new low-power general-purpose computer hardware. I know I'm not the only person here who's considered "Nerding Out" my car with a small multi-purpose computer...
However, it is effectively consolidating the Unix market into more or less a single front, which makes it a more formidable opponent to Windows in the long run.
The way I look at it, it's not so much a "single front" as a "single 'style'" (I'll stop myself JUST short of saying "standard").
Linux, *BSD, MacOSX, Sun, AIX, etc aren't QUITE a "single front"...but they all seem to be moving in a direction of interoperating with, and being more portable to, each other than MS Windows does with anything else (heck, they all probably work with WINDOWS better than Windows works with anything else...)
Increasingly, it seems like the IT world is moving towards portability and compatibility...EXCEPT for MS Windows. Windows seems to be ever more tightly imploding into itself in an attempt to exclude competitors. A previous poster mentioned that MS wants to position MS SQL as "the" database server for Microsoft Windows. Much as Internet Explorer is "the" browser for MS Windows, and MS Office 200x is "the" office suite for MS Windows, and "Outlook" is "the" mail client for MS Windows, and "SharePoint" is "the" WebDAV-(sorta)-based document sharing platform for MS Windows...and so on and so forth. They mess with the standards (I notice that SharePoint(tm) uses a non-standard mechanism for doing searches, so it won't be compatible with anyone else's compliant WebDAV client. Imagine that.) Unfortunate for Microsoft (but fortunate for everyone else - including "consumers" I think) the "everyone else" category is now large and influential enough that they don't have to capitulate to MS to stay in business.
I think EVENTUALLY even MS will have to adjust their business model and developement to play nicer with others (at least from a technical perspective), but I get the feeling it'll be ALMOST too late for them by the time they get around to doing so.
So where's Xiph's committment to the expectations they created with that announcement?
In the same place the committment to Theora (Now nearly a year behind the original "final release" date, and not even in Beta yet, and only being very sporadically worked on - though there's more going on with it now than there was for the 'dead' time that went on for so long after alpha1...the "Monty's the only developer and he's busy with Tremor [Integer-only vorbis for hardware players, if you didn't already know] for the forseeable future" period)
I'm beginning to think that Xiph, collectively, just doesn't have an attention span. NOW their big project is apparently connecting things to the Helix player. They're apparently "committed" to incorporating Theora support into Helix, too, which will be nice if they can finalize it (and Theora itself) before they get distracted by the NEXT project, whatever it may be...
As far as I can tell (as a "lurker" rather than a directly involved developer - YMMV) they are still debating what the specification for the OGG file format should look like in its final form, or so the only recent activity in the Theora mailing list seems to imply (revolving - as I amateurishly interpret it - around complaints that Ogg is hard to map to Microsoft(r) DirectShow(r)...)
I agree, though - it's FREE software, so even the incomplete stuff we get from them is still better than the nothing we'd get without them, so consider all of this a "lament" rather than a "complaint" from me...
I tell people not to think of it as the "command line" - I think it's more like "keyboard shortcuts" for the core system...
And by extension, a system with no CLI is like a program with no keyboard shortcuts: You shouldn't be FORCED to use them, but some things are always quicker and easier that way.
Echoing some of the other posts here: It depends what you want to get out of Linux.
If you "just want to use it" (i.e. you just want a nice desktop system that isn't proprietary, or just want to try out some Linux programs) I'd recommend (roughly in order of preference):
Knoppix - don't even need to 'install' to try it out
SuSe (Or is it SuSE? Or Suse? I can never keep it straight). They've always seemed like a reasonably decent distribution, and now with Novell's apparently-earnest-and-sincere backing, I tend to think of them as "the new Red Hat" (in that they'll be a 'default' distribution in the near future and thus useful to be familiar with).
Mandrake - I personally didn't much care for what I saw of Mandrake, but many people do, and this one is often suggested for "newbies".
If, on the other hand, you actually want to learn Linux:
Slackware - my personal favorite. It seems to be the most "generic-Unix-like" in general style. I think that although it's fairly "manual" in terms of setup, it's not difficult. But that's my opinion. It's probably the easiest-to-install-and-configure "nerd" distribution.
Gentoo - takes forever to install initially, because it actually compiles (most) of itself directly from source. You can configure it for the best compiler optimizations for your hardware and have, theoretically, the fastest-possible system. Maybe a bit of a steep learning curve to start with though - although if you're up for it, it may be worth it.
Debian - Possibly one of the most obtuse installation procedures (at least, last time I tried it), but once installed it's amazingly easy to maintain. ("apt-get" really is as cool as they say it is). Debian's one of the "big" famous distributions also, so if you find yourself working with a wide variety of Linux boxen, it's good to be at least a little familiar with it. Debian's other downside (apart from the initial install procedure) from my perspective is that by default their software versions tend to be quite old, though you can get around this by reconfiguring your system to pull updates from the "unstable" and "testing" branches.
It would, of course, be tragic if the University kicked him out for having an OpenOffice version of the file available...
Seriously - the department might short-sightedly require Microsoft(tm) PowerPoint(tm) files to be given to them, but I find it difficult to believe that actually developing or even presenting the slides on his own computer with OpenOffice would be forbidden...
...except for the ones who have hit their term limits. They'll do whatever they want. Typically, I would imagine, to curry favor with people who might give them high-paying jobs later.
And, similarly, just how "proprietary" and/or patent-encumbered is this ""Material eXchange Format that Dirac plans to use as a primary container?
It appears that the 'Official documentation' is only accesible to 'members' at the "Pro-MPEG Forum". Not a particularly "Open Source"-friendly situation...
Or better still, ffmpeg (the "libavcodec" and "libavformat" libraries that are components of ffmpeg are also used by mplayer/mencoder, so adding them there will 'automatically' add them to MPlayer, as well as other projects, at the same time.)
I really wish the Theora project would quit grinding to a near halt after every reluctant release...they're now, what, nearly 2 years behind their original schedule to go "beta"? (The Theora mailing list has gone 'dead' again...)
The "handout" linked at the BBC site says that they intend to use the "MXF" file format (for streaming applications), and that MPlayer has "some support" for MXF, though of course without the codec supported, that doesn't mean much.
I wonder how Theora/VP3 and Dirac compare in quality? Anyone know if Dirac will be relatively easy to incorporate into Ogg files/streams (and conversely Theora into MXF streams?)
I notice Dirac is based on wavelets - is it related to Tarkin (Xiph's even-deader-than-Theora, purely experimental wavelet-based video codec)?
The idea of traditional open source is that anybody has access to the source code. Not just educational institutions and/or somebody specifically registered with certain institutions such as the National Corn Growers Association.
Agreed. How about anybody with a web browser? In true open-source fashion, you may need to compile the genome yourself from its components, but with over 2,000,000 nucleotide sequences, over 7,500 peptide sequences, the complete genome of Zea mays' chloroplasts, etc. to work with it should be possible...
You may be on to something there. I genuinely keep finding myself wondering: If we presume that a substantial portion of "Global Warming" is caused by human activity, how much of that warming is REALLY caused by "greenhouse gasses" and how much is the result of land-use changes (e.g. the "Urban heat island" effect, and other changes brought about by land use)?
The whole "Greenhouse Gas" issue is vague and not realistic to 'solve' (we can 'cut back' on our production...if we can get enough of the world to agree on it...and if we don't just switch to different greenhouse gasses e.g. more water vapor...) and in any case proposed solutions would presumably take a long time to make any difference, while requiring huge changes in ways of life and political policies. If land uses are a major factor, though, that would imply to me that observable mitigation on a relatively short timescale might be possible, maybe even on a 'local' basis.)
I'm cynical enough to wonder if people just don't WANT Global Warming to become a 'readily solvable' problem hence what seems like an unusual effort to me to keep the focus on 'greenhouse gases'. There seems to be an "if it doesn't require radical changes, we must not give it credibility". That might explain why G. Tyler Miller Jr's current "Environmental Science" text entirely avoids mentioning BioDiesel (despite being 'renewable', biodiesel wouldn't require a radical change in automotive technology to use...Miller doesn't even mention it once, despite the fact that supposedly this is a very popular "Environmental Science" textbook.)
Getting off the topic, I know, but I've come to the conclusion that "Environmental Science" is a science only in the same way that "Political Science" is....
I THINK, like the Linux kernel, the odd-numbered 'minor' releases (e.g. 2.9.x, 3.1.x, 3.3.x, etc.) are the 'development' branches, and the 'even' numbered ones are the 'release' ones...meaning that the binary compatibility changes actually took place in 2.9.x, 3.1.x, and (most importantly for me) 3.3.x.
Yeah, I know, replying to myself...
A post above contradicts this, so I may be wrong about this...but I DO think I was remembering the binary incompatibility occurring in the 3.3 series correctly in this case. (My impression is that 3.4 doesn't have too many 'new features' beyond 3.3, but had more of a focus on optimizing the compile speed of 3.3.)
I think you're in the same boat I'm in. Care for a bailing bucket?
Nope, no need it turns out....
I THINK, like the Linux kernel, the odd-numbered 'minor' releases (e.g. 2.9.x, 3.1.x, 3.3.x, etc.) are the 'development' branches, and the 'even' numbered ones are the 'release' ones...meaning that the binary compatibility changes actually took place in 2.9.x, 3.1.x, and (most importantly for me) 3.3.x.
In any case, I just dropped out of KDE (QT/KDE are C++...), updated to my newly-compiled 3.4.0 GCC/GCC-G++/GCC-JAVA packages, and restarted. No problems so far running off of the new C++ library it appears. (And I DO remember, I think, having to go through the 'recompile QT/KDE because of binary incompatibility' when I updated to the 3.3.x series...)
So, it looks like anyone who's been running off of the 3.3 series GCC compilers SHOULD be okay to upgrade.
Re:Broken C++ ABI ... again
on
GCC 3.4.0 Released
·
· Score: 4, Interesting
Well, if you want to be technical about it, it's not "Broken" C++ ABI, but a "Finally fixed, even though that makes it no longer bug-for-bug compatible with older GCC C++ ABI's"...
As I understand it, they've been working towards a more standards compliant C++ implementation, and that's why the binary compatibility gets lost.
I am, though, hoping that there was NOT a loss of compatibility between the 3.3.3 that I'm using now and the 3.4 series. Will find out once I clean off enough disk space to finish compiling up slack packages for myself...
Jeez, keep up with the times, will you! I mean, how hard is it REALLY to keep a local copy of the CVS tree on your local hard drive and recompile every night from there? Sheesh!
(And the sad thing is that, although I'm joking, this isn't too far off of what I actually DO....)
Slashdot Mirror
Heck, I just compiled up Quakeforge last night to try it out on my little laptop here (Athlon XP-M 1600+ (1400Mhz), 352MB, 32MB shared video ram, S3 Savage video chipset) and was amused to see I was getting between 40 and 75 fps at whatever the heck the default resolution was. Well inside the 'Playable' range.
I'm disappointed that Id refuses to offer a "data files for open-sourced games" (i.e. Doom 1, Doom 2, Quake 1, Quake 2) disk or download at a reasonable price - I asked about it and got back a shocked "Oh, no, we'd NEVER do something like THAT!" as though I'd asked them if they'd consider auctioning off their genitals on E-bay or something...Since they're charging "new Quake I box" rates last time I checked ($30+) for a simple download of the original Quake disk alone, I just can't quite justify the cost at the moment. But I'd pay that much for permission to use (no support, even no media [i.e. download disk image only]) the quake 1, quake 2, and Doom 1,2 datasets to go with Quakeforge and the various DOOM ports...
That being the case - anybody know any good free "total conversion" datasets that don't require the original ID datasets? OpenQuartz appeared to be working on one, but they haven't had an update in over 2 years...
They have Ogg Vorbis as well...
Yeah, but imagine a beowulf^h^h^h^h^h^h^htraffic jam of them!
(Why DOES Slashdot forbid the "<strike>" tag, anyway?...)
I find it's especially useful for any application that involves "talking to a service over the network" or dealing with text-like (including html/xml) files and/or relatively simple GD-compatible graphics file manipulation (generating graphs, maps, etc.), whether it's a database server, a webserver, an ftp server, a socket, etc. etc.
My own current personal project involves figuring out the "TIGER/Line" geographic data from the US Census bureau so that I can render the features at will over other existing map graphics and turn them into maps for GPSDrive. I've already got PHP scripts generating "USGS Topographic maps or TIGER/Line [1998 data rendered from the census bureau site] features rendered atop USGS Terraserver aerial photos" for GPSDrive...
I've also considered doing a Kismet client class for PHP. Not real useful for old-fashioned web use, but could be handy for CLI (or PHP-GTK or whatever).
And why not? It's fun.
What? And settle for UN-OPTIMIZED binaries like some plebe? Forget that!
Besides, I had to grab the CVS from the DRI branch to get S3/SavageDDR (It's a laptop, I didn't have much choice...) direct-rendering support. Why not compile it up with "-O2 -march=athlon-xp -mmmx -m3dnow -msse -mfpmath=sse,387" while I'm at it?...
Yes, I'm also running on bleeding-edge CVS compile of KDE 3.3-post-alpha as well (working fine for me, incidentally...)
Think of it as the civilized nerd equivalent of peeing on the computer to mark it as yours. Except that you don't ruin the computer and/or get painfully shocked in the process (uh, you know, or so I've HEARD....)
There is a lot of talk lately where I'm employed about the "need" for "Electronic Document Management"(tm), but nobody seems to know EXACTLY what we need or want from it, and nobody's really put any though thus far into working on some user requirements specs for it so that we can evaluate what we really DO need.
Thus far, proprietary EDM sites seem to be filled to the brim with low-content fluff but little REAL information about exactly what you get from them. Even presuming they DO have truly worth-the-price features, it seems difficult to really pin down what those features really ARE...
I think that's part of the reason they can get away with such egregiously high license fees.
Who needs special software for that? Everything you need to know about writing a scientific paper is right here.
Wonder if this one would make a good "Ask Slashdot"?...
I've been wondering lately myself about this. Just how inexpensively can one put together a reasonably-useful yet reasonably-standard computer off-the-shelf? (As opposed to "go on E-bay and see if you can find X" or "if you can find THIS version of THIS hardware, you can hack it to put Linux in it" and so on...)
It'd be nifty to have some good sources for very inexpensive new low-power general-purpose computer hardware. I know I'm not the only person here who's considered "Nerding Out" my car with a small multi-purpose computer...
The way I look at it, it's not so much a "single front" as a "single 'style'" (I'll stop myself JUST short of saying "standard").
Linux, *BSD, MacOSX, Sun, AIX, etc aren't QUITE a "single front"...but they all seem to be moving in a direction of interoperating with, and being more portable to, each other than MS Windows does with anything else (heck, they all probably work with WINDOWS better than Windows works with anything else...)
Increasingly, it seems like the IT world is moving towards portability and compatibility...EXCEPT for MS Windows. Windows seems to be ever more tightly imploding into itself in an attempt to exclude competitors. A previous poster mentioned that MS wants to position MS SQL as "the" database server for Microsoft Windows. Much as Internet Explorer is "the" browser for MS Windows, and MS Office 200x is "the" office suite for MS Windows, and "Outlook" is "the" mail client for MS Windows, and "SharePoint" is "the" WebDAV-(sorta)-based document sharing platform for MS Windows...and so on and so forth. They mess with the standards (I notice that SharePoint(tm) uses a non-standard mechanism for doing searches, so it won't be compatible with anyone else's compliant WebDAV client. Imagine that.) Unfortunate for Microsoft (but fortunate for everyone else - including "consumers" I think) the "everyone else" category is now large and influential enough that they don't have to capitulate to MS to stay in business.
I think EVENTUALLY even MS will have to adjust their business model and developement to play nicer with others (at least from a technical perspective), but I get the feeling it'll be ALMOST too late for them by the time they get around to doing so.
In the same place the committment to Theora (Now nearly a year behind the original "final release" date, and not even in Beta yet, and only being very sporadically worked on - though there's more going on with it now than there was for the 'dead' time that went on for so long after alpha1...the "Monty's the only developer and he's busy with Tremor [Integer-only vorbis for hardware players, if you didn't already know] for the forseeable future" period)
I'm beginning to think that Xiph, collectively, just doesn't have an attention span. NOW their big project is apparently connecting things to the Helix player. They're apparently "committed" to incorporating Theora support into Helix, too, which will be nice if they can finalize it (and Theora itself) before they get distracted by the NEXT project, whatever it may be...
As far as I can tell (as a "lurker" rather than a directly involved developer - YMMV) they are still debating what the specification for the OGG file format should look like in its final form, or so the only recent activity in the Theora mailing list seems to imply (revolving - as I amateurishly interpret it - around complaints that Ogg is hard to map to Microsoft(r) DirectShow(r)...)
I agree, though - it's FREE software, so even the incomplete stuff we get from them is still better than the nothing we'd get without them, so consider all of this a "lament" rather than a "complaint" from me...
I tell people not to think of it as the "command line" - I think it's more like "keyboard shortcuts" for the core system...
And by extension, a system with no CLI is like a program with no keyboard shortcuts: You shouldn't be FORCED to use them, but some things are always quicker and easier that way.
Echoing some of the other posts here: It depends what you want to get out of Linux.
If you "just want to use it" (i.e. you just want a nice desktop system that isn't proprietary, or just want to try out some Linux programs) I'd recommend (roughly in order of preference):
If, on the other hand, you actually want to learn Linux:
Those would be my suggestions, anyway.
You forgot "Obnoxio the Clown".
(Egad, did I just admit in public that I knew of that character?...)
It would, of course, be tragic if the University kicked him out for having an OpenOffice version of the file available...
Seriously - the department might short-sightedly require Microsoft(tm) PowerPoint(tm) files to be given to them, but I find it difficult to believe that actually developing or even presenting the slides on his own computer with OpenOffice would be forbidden...
...except for the ones who have hit their term limits. They'll do whatever they want. Typically, I would imagine, to curry favor with people who might give them high-paying jobs later.
"I just saved paying $20 for a retail CD by downloading the one song I wanted without permission from the copyright holder..."
How difficult, really, is it going to be for lawyers to argue that this is 'private financial gain'?
And, similarly, just how "proprietary" and/or patent-encumbered is this ""Material eXchange Format that Dirac plans to use as a primary container?
It appears that the 'Official documentation' is only accesible to 'members' at the "Pro-MPEG Forum". Not a particularly "Open Source"-friendly situation...
Or better still, ffmpeg (the "libavcodec" and "libavformat" libraries that are components of ffmpeg are also used by mplayer/mencoder, so adding them there will 'automatically' add them to MPlayer, as well as other projects, at the same time.)
I really wish the Theora project would quit grinding to a near halt after every reluctant release...they're now, what, nearly 2 years behind their original schedule to go "beta"? (The Theora mailing list has gone 'dead' again...)
The "handout" linked at the BBC site says that they intend to use the "MXF" file format (for streaming applications), and that MPlayer has "some support" for MXF, though of course without the codec supported, that doesn't mean much.
I wonder how Theora/VP3 and Dirac compare in quality? Anyone know if Dirac will be relatively easy to incorporate into Ogg files/streams (and conversely Theora into MXF streams?)
I notice Dirac is based on wavelets - is it related to Tarkin (Xiph's even-deader-than-Theora, purely experimental wavelet-based video codec)?
If you had well-designed directional antennae, could you bounce wifi network signals off of these?....
Agreed. How about anybody with a web browser? In true open-source fashion, you may need to compile the genome yourself from its components, but with over 2,000,000 nucleotide sequences, over 7,500 peptide sequences, the complete genome of Zea mays' chloroplasts, etc. to work with it should be possible...
You may be on to something there. I genuinely keep finding myself wondering: If we presume that a substantial portion of "Global Warming" is caused by human activity, how much of that warming is REALLY caused by "greenhouse gasses" and how much is the result of land-use changes (e.g. the "Urban heat island" effect, and other changes brought about by land use)?
The whole "Greenhouse Gas" issue is vague and not realistic to 'solve' (we can 'cut back' on our production...if we can get enough of the world to agree on it...and if we don't just switch to different greenhouse gasses e.g. more water vapor...) and in any case proposed solutions would presumably take a long time to make any difference, while requiring huge changes in ways of life and political policies. If land uses are a major factor, though, that would imply to me that observable mitigation on a relatively short timescale might be possible, maybe even on a 'local' basis.)
I'm cynical enough to wonder if people just don't WANT Global Warming to become a 'readily solvable' problem hence what seems like an unusual effort to me to keep the focus on 'greenhouse gases'. There seems to be an "if it doesn't require radical changes, we must not give it credibility". That might explain why G. Tyler Miller Jr's current "Environmental Science" text entirely avoids mentioning BioDiesel (despite being 'renewable', biodiesel wouldn't require a radical change in automotive technology to use...Miller doesn't even mention it once, despite the fact that supposedly this is a very popular "Environmental Science" textbook.)
Getting off the topic, I know, but I've come to the conclusion that "Environmental Science" is a science only in the same way that "Political Science" is....
Yeah, I know, replying to myself...
A post above contradicts this, so I may be wrong about this...but I DO think I was remembering the binary incompatibility occurring in the 3.3 series correctly in this case. (My impression is that 3.4 doesn't have too many 'new features' beyond 3.3, but had more of a focus on optimizing the compile speed of 3.3.)
Nope, no need it turns out....
I THINK, like the Linux kernel, the odd-numbered 'minor' releases (e.g. 2.9.x, 3.1.x, 3.3.x, etc.) are the 'development' branches, and the 'even' numbered ones are the 'release' ones...meaning that the binary compatibility changes actually took place in 2.9.x, 3.1.x, and (most importantly for me) 3.3.x.
In any case, I just dropped out of KDE (QT/KDE are C++...), updated to my newly-compiled 3.4.0 GCC/GCC-G++/GCC-JAVA packages, and restarted. No problems so far running off of the new C++ library it appears. (And I DO remember, I think, having to go through the 'recompile QT/KDE because of binary incompatibility' when I updated to the 3.3.x series...)
So, it looks like anyone who's been running off of the 3.3 series GCC compilers SHOULD be okay to upgrade.
Well, if you want to be technical about it, it's not "Broken" C++ ABI, but a "Finally fixed, even though that makes it no longer bug-for-bug compatible with older GCC C++ ABI's"...
As I understand it, they've been working towards a more standards compliant C++ implementation, and that's why the binary compatibility gets lost.
I am, though, hoping that there was NOT a loss of compatibility between the 3.3.3 that I'm using now and the 3.4 series. Will find out once I clean off enough disk space to finish compiling up slack packages for myself...
Jeez, keep up with the times, will you! I mean, how hard is it REALLY to keep a local copy of the CVS tree on your local hard drive and recompile every night from there? Sheesh!
(And the sad thing is that, although I'm joking, this isn't too far off of what I actually DO....)