I've found that the Linux kernel is pretty easy to grok and modify.
Agreed for small stuff (as I never tried some big changes). I once implemented a small driver for reading the sensors on my motherboard (before I found lm-sensors). It took me about a day to figure out the doc for the chip, then how to make the kernel ask the chip and report the values in/proc, and finally adjusting the values of resistors to match the BIOS values. Then I found lm-sensors, and I never applied that small patch again.
... as this sort of modification would probably be really hard to do with Windows:)
Maybe the optimization that is causing problem is simply absent there... (don't know, don't have a Windows box)
OpenBSD has a number of features to do the "transparent" routing which you desire - basically, you can plug in two interfaces and route from one to the other without changing the packets at all.
That's not exactly what the guy wants. In your option, you have:
<machine>
12
<outer network>--
--<inner network>
Which means that the OpenBSD acts as a repeater (a hub) without any IP addresses on the 2 NICs, and without touching the packets it lets go through.
Have I correctly understood what you said OpenBSD can do?
The submitter wants this:
<machine>
1
2
<switching equipment>
His goal is to send packets on NIC 1, through his switching equipment, then trough NIC 2, and check for errors.
And what prevents him from doing this is that the Linux kernel sees that the destination IP address is in the same computer, and bypasses the 2 NICs and the switching equipment to be tested.
So the question is thus: how to bypass that? I don't think that any virtual machine (UML, VMWare) will work here, as the network traffic they generate will probably pass by the TCP/IP stack of the host, and won't work. Your best bet may be to take an older kernel version (2.2.21 has been released yesterday), which doesn't implement that enhancement.
Would it be possible to test at another layer? Raw IP, IPX, or even raw 802.3 frames? It will take a bit more of coding, but the switch should be able to speak those languages. Of course, it doesn't test the TCP/IP capability of the switch, but I don't think it really matters as the switch sees MAC addresses, not IP addresses. OTOH, if you don't test it, it may break silently...
Building a radio emitter is not very difficult (I remember having an electronic kit when I was young with 60+ different circuits you could build, and one of them was a radio emitter). Ideally you'd choose it not to interfere with your local channels, or some neighbours could become upset if your power is too high.
Then, on the receiving end, a small walkman is all you need. Plug your headphones or earphones, and there you go!
Of course, the quality of the transmission will vary depending on the quality of the hardware and which frequency you choose (near or far from some other channel).
s/2 million dollars per year/2 billion dollars per year/?
I know there's been inflation since then, but 2 million wasn't that big back then... it certainly is not anymore.
I second your proposition for M. Gates. Now, who will try to convince him?:)
Is it a symptom of something that this story has more comments than the SW review? Or is it just that some people didn't have the chance to see it yet, and prefer not to read a review of it before?
Of course, with all the PK (player killers) in online FPS, you'd have a very tough ride.
Let's say a football game: the quarterback chooses the play (is he alone to choose, or the rest of the team can have an input too?), everybody lines up... and before the play is started, that moron on the left gets another penalty for encroachment, doubled by a major fault on the QB. Sure you can kick him out, but the damage's still done.
Also, seems very fun to play as an offensive guard...
Another example, featuring "real" football (soccer): the PK gets to control a very good defenseman. He tackles by behind a few times (maybe injuring an opposing player in the process), and the AI referee finally redcards the player. PK gone, but he leaves his team 10 to 11!
Again, very fun it is to control a player not part of the action... Who to be goalie wants and the game watch from 100 yards away? (bad attempt at Yoda speak)
At least in real sports, you have some practices between the games, so there's no incentive for a moron to act funny in games, as he has to act correctly in practices. Or if he does act funny in games, at least you can kick him (for real)...
Don't forget it's the same country where a law firm can send a trademark infringement letter to somebody on behalf of somebody else (without their nowing), and get paid for it in the process (see Samba, etc.)
So even if your post was meant to be funny, don't forget that there can be some things you like and some others you dislike in everything.
No need to weld anything. Check the Yeong Yang cube case (02 serie), or the Chenbro server line. Plenty of room in any of those (up to 12 5 1/4 bays in the SR101).
But I think the original poster asked for a case with bays not only facing in front, but also either up or on the side (so his cables would go in different directions and be less cluttered).
For that, one of the small P4 cases recently featured on Slashdot might be useful, as a secondary case, or any minitower for that matter.
First, schools:
High-school: Only XTs. No network. No login. Only bootdisks.
College: Student number. The email was the same.
University:
Department is Initial+Lastname (eg, jdoe). The duplicates are labeled jdoe, jdoe1, etc.
Faculty is 3FirstLettersOfLastName+Initial+Number, as in doej01.
Lastly, the University introduced a campus-wide login. I think it involves the year in which you began to attend classes here, along with a variation of your name and a sequential number (along jdoe9901).
There's also a campus-wide email system, different from the previous, where the username is your student number, but you can choose an alias which is a variation of your name: jd1, johndoe, jdoe, doej, john.doe and maybe others.
Work places:
The first one was the same thing as my faculty (jdoe01).
The second one had the employee number to login, but you also had an alias for email based on your name. The translation from name -> alias wasn't constant, though, so you had to lookup in the employee list (~50000) to know the email address of somebody.
Lastly, another one was mostly only the firstname. The company wasn't very big (~250), and it wasn't uniform at all. I heard that it changed since I left, with emails being firstname.lastname, but I don't know about the usernames.
And of course, my own systems:
There's my normal user (firstname), and root. Although I'll probably change root for something meaner.
Those are my experiences with usernames. Hope it can help somebody find their best choice.
Actually, the SRPM for 1.0RC1 is a very good starting point if you want to do it yourself. The difference between the spec from 0.9.9 to 1.0RC1 are very minimal (mostly resulting from the renaming of an secondary source file), and I didn't find any in the other source files. Get the 1.0RC1 SRPM (if you don't already have it), replace the tar.gz by the new one, change the spec file accordingly, and fire it up.
When (if?) the SRPM shows up on mozilla.org, you can compare the 2 spec files (along with the rest of the source files) and if your setup is correct, you'll already have it.
And remember, if it ever breaks, you get to keep all the pieces...
Yes, some people are still using them in their designs. Last summer, I worked in a fiber-optics card company, and they had one on each card, doing some small work I think wrt control communication between the cards, but I was in manufacturing, not in a desing/test departement, so I might be wrong.
Not so long before, I had chosen a PIC for a project of mine (controlling a small DC motor wrt a couple inputs, both A&D), and I was actually a bit surprised to see some in a production environment. Although the Microchip site doesn't hint to only "tinkerers"...
Oh, another use: it seems there's a lot of satellite boxen which use a certain PIC (I think it's a PIC16F84) for authentication (either in the box, or on a ???-card). I found quite some software programmers for that chip, but only 2 for the PIC16C71 that I chose.
With RH-7.3, you have the choice of postfix or sendmail as MTAs. There's even a tool to help you switch (ala changedesktop). See the new features announcement.
I must say that I haven't installed it yet (this weekend), but you do can install RH without sendmail.
Play games: almost finished JK II with Wine, I just need a bit more free time (those black Jedis are quite hard to defeat). After that, I'll probably add it to my usual round of native LAN games: UT, Q3A, Wolf...
Word processing: LaTeX (I'm doing a masters), OpenOffice. Didn't tried AbiWord lately.
Download music: Don't know, as I only listen to radio, and it's a news channel. And besides, if you don't want to view DVDs under Linux for claims of "semilegality", what are you doing downloading music?
Print: Don't know either, as I don't own any printer, and at the Uni there's only Windows computers. So it's not "reboot", it's "get up, walk a few minutes, print and come back". But it's not because of the OS: it's because if I ever buy a printer, I want a bigass departmental color laser printer. Which is kinda too pricey for me right now.
DVD: Began watching the Slap Shot 2 DVD this morning, but didn't have enough time to finish it. Will be tonight, since the Canadiens don't play before tomorrow.
Not if you've been cooling your juice in the fridge;-)
That was the point.
When you pay x joules to a fridge for him to keep your juice cold, he extracts some more heat from that juice before relinquishing (x + some) joules in the surroundings. The x joules are what is used to cool it, the balance is only displaced from inside the fridge to outside, and doesn't really participate in global warming since if I take the juice out of the fridge, that same heat can bring it back to the temperature it was at before I put it in the fridge. Still following?
Secondly, the heat needed to be removed from the juice is less than that needed to boil a cup of water or of coffee (the temperature difference is way smaller).
Therefore, it uses less energy to cool it (so less participation in global warming), and since when I finish it it's a bit warmer than when I started to drink it, but still colder than before I put it in the fridge, it gives back more energy (when in the fridge) than it absorbs (when I drink it).
So it's better to drink juice or a pop than coffee.
I'm not sure about the average 50W. Let's take an Athlon XP 1800+:
According to it's datasheet, it's typical power consumption is 59.2W. Add to this the rest of the cards in the system, RAM, chipset, HDD, CD, etc. Lets say the total is 80W, which is conservative. You don't need a fan on your NB or your GPU for nothing, although when you're not doing 3D the latter shouldn't be hot. Then you have the power supply, the efficiency of which is usually 70% at full load, and less than that if it's not fully loaded. Since the vast majority of PSU shipped nowadays is 300W, and 80W is way lower (but you still need 300W for peaks, like when you boot or game, or if you have quite some cards or HDD), the efficiency is probably around 50%. The rest of the calculation seems correct, so the total for 100 000 computers is about 17.4 million cup of teas.
Total is, it's still lower than the power needed to boil the number of cup of tea drank in UK in a day (I would have taken coffee rather than tea for the example, as it's more common internationally), but the starting figure of 50W per computer seems low. In my room, my 2 computers quickly heat the room more than if I have a 100W light bulb on (although part of it is light, so doesn't heat the room as much).
Just to do this calculation is intersting: it shows the relative weight of some human activities. I'd also really like to have an accurate view of the electrical consumption of my computers.
And since I prefer cold juice to tea or coffee, I don't take part as much as you to the global warming. My drink gives back more heat than it absorbs:)
I think what they want to do is, given the state in 1950 and what we know about the inputs (use of automobile, etc.), which models predicts correctly what we see in 2000, so that those models can then be used, along with some new inputs, to forecast what 2050 will be. Either prospective inputs, to get a glimpse of our possible futures, or actual inputs, to further validate the model or get a sharper view of the future climate.
That being said, 8 months is way too long to get something useful. I know a couple friends who reinstall their OS (and apps) in shorter terms than that, and don't really bother with bringing all data along, just some backup on CDR "in case I really want it again". I think they could at least chop it in periods of a few years, so that if you finish a "unit", somebody else can then pick up where you left. I'd like to see the completion efficiency of whole units in a few months.
I see you have gcc-2.96 for C, java, C++, FORTRAN, objective-C and chill (although I don't know that last language). I also see you have egcs-1.1.2 for compiling C, C++, FORTRAN and obective-C code for a RH-6.2 target. You also have the libs to run those. If you check the 6.2 release, you can probably find some compat packages for targetting RH-5.2. And you can even install them if you want to target such an old release. You can probably even find some compat-libc5 rpms in RH-5.2. But in day to day use, or for compiling something for a RH-7.3 system, I don't see why you'd use egcs-1.1.2.
Back to kgcc... I don't see a kgcc in your rpm -a. By checking on a RH mirror, the only distro I can find with kgcc-1.1.2-40 is RH-7.0. After that, they always shipped only gcc-2.96-something (and it has been used extensively for kernels), except in 7.2 where gcc3 was also included. Unless you call kgcc something packaged as compat-egcs, there's no kgcc in 7.1, 7.2 nor 7.3.
Regarding the presence of gcc3 in RH-7.2, wasn't it meant as a technology preview only? Which explains both it's absence of 7.3 and the fact that it was a patched CVS checkout...
Of course, the other distributors can put whatever they want in their distros. The same way that some prefer CUPS and others LPRng. As long as they do the same thing...
A lot of the GCC 3 is broke with regards to the C++, that's a crock.
There you have it! The C ABI is pretty standard, and doesn't change between compiler releases. But the C++ side does. That's why you need a different libstdc++ for each compiler release. It's possible to upgrade the C compiler of a system easily (ICC, the Borland one, GCC). The C++ compiler is a bit more work, and RH prefered to not do it right now, but when GCC 3 will be ready for prime time (whether it's 3.0.4, 3.1, PI or 3.2 is not really relevant to the discussion).
Your suggestion of having 3 or 4 different releases of the same program is not a good idea: do you really expect users and developpers to choose between the different compilers dependant on the project they compile? "Do sdjy 0.58 needs EGCS 1.1.2, or is it safe to use 3.0.4? Oh right, it's 2.95.2." The goal of RH is to have only one: for the time being it's their own release, tagged 2.96-110, but it should evolve to a 3.?? based compiler for 8.0.
The other thing to keep in mind is that the gcc they ship now in 7.3 is tagged 2.96-110. Which means there's been a lot of patches applied since the beginning! When people grunt against gcc-2.96, they seldomly mention the RPM release. So yes, probably the first ones had a lot of problems (which explains why they shipped a kgcc in 7.0), but the current version works quite fine for me, even for kernel compilation (which explains why kgcc disappeared).
In one of the James bond (sorry, don't recall which one), James has a watch with a "magnetic beam" which he uses to move a car, and to undress a woman:) The first time is the more difficult to pass by, as he holds his watch with only 2 fingers and is still able to move (at distance) a car.
Slashdot Mirror
How are they supposed to do 3D stuff on a PDA? Will it feel like Q3A on a P200MMX with a Voodoo2, or better/worse?
For one, I'd prefer see them continue on their WineX development. But I can understand, if they need the money, that they'd join such a venture.
No login name. Or no real name, only email address. Hence AC.
I've found that the Linux kernel is pretty easy to grok and modify. /proc, and finally adjusting the values of resistors to match the BIOS values. Then I found lm-sensors, and I never applied that small patch again.
... as this sort of modification would probably be really hard to do with Windows :)
Agreed for small stuff (as I never tried some big changes). I once implemented a small driver for reading the sensors on my motherboard (before I found lm-sensors). It took me about a day to figure out the doc for the chip, then how to make the kernel ask the chip and report the values in
Maybe the optimization that is causing problem is simply absent there... (don't know, don't have a Windows box)
OpenBSD has a number of features to do the "transparent" routing which you desire - basically, you can plug in two interfaces and route from one to the other without changing the packets at all.
That's not exactly what the guy wants. In your option, you have:
<machine>
12
<outer network>-- --<inner network>
Which means that the OpenBSD acts as a repeater (a hub) without any IP addresses on the 2 NICs, and without touching the packets it lets go through.
Have I correctly understood what you said OpenBSD can do?
The submitter wants this:
<machine>
1 2
<switching equipment>
His goal is to send packets on NIC 1, through his switching equipment, then trough NIC 2, and check for errors.
And what prevents him from doing this is that the Linux kernel sees that the destination IP address is in the same computer, and bypasses the 2 NICs and the switching equipment to be tested.
So the question is thus: how to bypass that? I don't think that any virtual machine (UML, VMWare) will work here, as the network traffic they generate will probably pass by the TCP/IP stack of the host, and won't work. Your best bet may be to take an older kernel version (2.2.21 has been released yesterday), which doesn't implement that enhancement.
Would it be possible to test at another layer? Raw IP, IPX, or even raw 802.3 frames? It will take a bit more of coding, but the switch should be able to speak those languages. Of course, it doesn't test the TCP/IP capability of the switch, but I don't think it really matters as the switch sees MAC addresses, not IP addresses. OTOH, if you don't test it, it may break silently...
Building a radio emitter is not very difficult (I remember having an electronic kit when I was young with 60+ different circuits you could build, and one of them was a radio emitter). Ideally you'd choose it not to interfere with your local channels, or some neighbours could become upset if your power is too high.
Then, on the receiving end, a small walkman is all you need. Plug your headphones or earphones, and there you go!
Of course, the quality of the transmission will vary depending on the quality of the hardware and which frequency you choose (near or far from some other channel).
s/2 million dollars per year/2 billion dollars per year/?
:)
I know there's been inflation since then, but 2 million wasn't that big back then... it certainly is not anymore.
I second your proposition for M. Gates. Now, who will try to convince him?
Just make sure nobody tries to do inter-dimensional space travel over there. You never know what Hell will do... and I don't want my rabbit to die.
Is it a symptom of something that this story has more comments than the SW review? Or is it just that some people didn't have the chance to see it yet, and prefer not to read a review of it before?
Of course, with all the PK (player killers) in online FPS, you'd have a very tough ride.
Let's say a football game: the quarterback chooses the play (is he alone to choose, or the rest of the team can have an input too?), everybody lines up... and before the play is started, that moron on the left gets another penalty for encroachment, doubled by a major fault on the QB. Sure you can kick him out, but the damage's still done.
Also, seems very fun to play as an offensive guard...
Another example, featuring "real" football (soccer): the PK gets to control a very good defenseman. He tackles by behind a few times (maybe injuring an opposing player in the process), and the AI referee finally redcards the player. PK gone, but he leaves his team 10 to 11!
Again, very fun it is to control a player not part of the action... Who to be goalie wants and the game watch from 100 yards away? (bad attempt at Yoda speak)
At least in real sports, you have some practices between the games, so there's no incentive for a moron to act funny in games, as he has to act correctly in practices. Or if he does act funny in games, at least you can kick him (for real)...
Don't forget it's the same country where a law firm can send a trademark infringement letter to somebody on behalf of somebody else (without their nowing), and get paid for it in the process (see Samba, etc.)
So even if your post was meant to be funny, don't forget that there can be some things you like and some others you dislike in everything.
No need to weld anything. Check the Yeong Yang cube case (02 serie), or the Chenbro server line. Plenty of room in any of those (up to 12 5 1/4 bays in the SR101).
But I think the original poster asked for a case with bays not only facing in front, but also either up or on the side (so his cables would go in different directions and be less cluttered).
For that, one of the small P4 cases recently featured on Slashdot might be useful, as a secondary case, or any minitower for that matter.
First, schools:
High-school: Only XTs. No network. No login. Only bootdisks.
College: Student number. The email was the same.
University:
Department is Initial+Lastname (eg, jdoe). The duplicates are labeled jdoe, jdoe1, etc.
Faculty is 3FirstLettersOfLastName+Initial+Number, as in doej01.
Lastly, the University introduced a campus-wide login. I think it involves the year in which you began to attend classes here, along with a variation of your name and a sequential number (along jdoe9901).
There's also a campus-wide email system, different from the previous, where the username is your student number, but you can choose an alias which is a variation of your name: jd1, johndoe, jdoe, doej, john.doe and maybe others.
Work places:
The first one was the same thing as my faculty (jdoe01).
The second one had the employee number to login, but you also had an alias for email based on your name. The translation from name -> alias wasn't constant, though, so you had to lookup in the employee list (~50000) to know the email address of somebody.
Lastly, another one was mostly only the firstname. The company wasn't very big (~250), and it wasn't uniform at all. I heard that it changed since I left, with emails being firstname.lastname, but I don't know about the usernames.
And of course, my own systems:
There's my normal user (firstname), and root. Although I'll probably change root for something meaner.
Those are my experiences with usernames. Hope it can help somebody find their best choice.
Have you read the article?
:0ep :0fp '0+, v > 2eg 1- : 2ep !#v_ 4eg : fg 1- \ fp 4eg fg #v _ 3eg ! 3ep 4eg 1+ 4ep v
This is an excerpt of the Befunge part:
> 2ep 1ep
v pe30 < a
And this is an excerpt of the BrainF*ck part (a language with 8 instructions):
>+>+<<<<<-]>>,--------- ---
And keep in mind that the actual program is much longer than that.
I wish I could include more, but the lameness filter won't allow me to...
So much for nearly identical languages...
Actually, the SRPM for 1.0RC1 is a very good starting point if you want to do it yourself. The difference between the spec from 0.9.9 to 1.0RC1 are very minimal (mostly resulting from the renaming of an secondary source file), and I didn't find any in the other source files. Get the 1.0RC1 SRPM (if you don't already have it), replace the tar.gz by the new one, change the spec file accordingly, and fire it up.
When (if?) the SRPM shows up on mozilla.org, you can compare the 2 spec files (along with the rest of the source files) and if your setup is correct, you'll already have it.
And remember, if it ever breaks, you get to keep all the pieces...
Yes, some people are still using them in their designs. Last summer, I worked in a fiber-optics card company, and they had one on each card, doing some small work I think wrt control communication between the cards, but I was in manufacturing, not in a desing/test departement, so I might be wrong.
Not so long before, I had chosen a PIC for a project of mine (controlling a small DC motor wrt a couple inputs, both A&D), and I was actually a bit surprised to see some in a production environment. Although the Microchip site doesn't hint to only "tinkerers"...
Oh, another use: it seems there's a lot of satellite boxen which use a certain PIC (I think it's a PIC16F84) for authentication (either in the box, or on a ???-card). I found quite some software programmers for that chip, but only 2 for the PIC16C71 that I chose.
With RH-7.3, you have the choice of postfix or sendmail as MTAs. There's even a tool to help you switch (ala changedesktop). See the new features announcement.
I must say that I haven't installed it yet (this weekend), but you do can install RH without sendmail.
Nice trolling, but I'll bite.
Play games: almost finished JK II with Wine, I just need a bit more free time (those black Jedis are quite hard to defeat). After that, I'll probably add it to my usual round of native LAN games: UT, Q3A, Wolf...
Word processing: LaTeX (I'm doing a masters), OpenOffice. Didn't tried AbiWord lately.
Download music: Don't know, as I only listen to radio, and it's a news channel. And besides, if you don't want to view DVDs under Linux for claims of "semilegality", what are you doing downloading music?
Print: Don't know either, as I don't own any printer, and at the Uni there's only Windows computers. So it's not "reboot", it's "get up, walk a few minutes, print and come back". But it's not because of the OS: it's because if I ever buy a printer, I want a bigass departmental color laser printer. Which is kinda too pricey for me right now.
DVD: Began watching the Slap Shot 2 DVD this morning, but didn't have enough time to finish it. Will be tonight, since the Canadiens don't play before tomorrow.
Not if you've been cooling your juice in the fridge ;-)
That was the point.
When you pay x joules to a fridge for him to keep your juice cold, he extracts some more heat from that juice before relinquishing (x + some) joules in the surroundings. The x joules are what is used to cool it, the balance is only displaced from inside the fridge to outside, and doesn't really participate in global warming since if I take the juice out of the fridge, that same heat can bring it back to the temperature it was at before I put it in the fridge. Still following?
Secondly, the heat needed to be removed from the juice is less than that needed to boil a cup of water or of coffee (the temperature difference is way smaller).
Therefore, it uses less energy to cool it (so less participation in global warming), and since when I finish it it's a bit warmer than when I started to drink it, but still colder than before I put it in the fridge, it gives back more energy (when in the fridge) than it absorbs (when I drink it).
So it's better to drink juice or a pop than coffee.
Ouch. You have an expensive computer. My 1GHz Toshiba laptop draws about 30watts finding prime numbers. I bet your air conditioner gets a workout.
I bet what you save on electricity bills, he saved at the time of purchase.
I'm not sure about the average 50W. Let's take an Athlon XP 1800+:
:)
According to it's datasheet, it's typical power consumption is 59.2W. Add to this the rest of the cards in the system, RAM, chipset, HDD, CD, etc. Lets say the total is 80W, which is conservative. You don't need a fan on your NB or your GPU for nothing, although when you're not doing 3D the latter shouldn't be hot. Then you have the power supply, the efficiency of which is usually 70% at full load, and less than that if it's not fully loaded. Since the vast majority of PSU shipped nowadays is 300W, and 80W is way lower (but you still need 300W for peaks, like when you boot or game, or if you have quite some cards or HDD), the efficiency is probably around 50%. The rest of the calculation seems correct, so the total for 100 000 computers is about 17.4 million cup of teas.
Total is, it's still lower than the power needed to boil the number of cup of tea drank in UK in a day (I would have taken coffee rather than tea for the example, as it's more common internationally), but the starting figure of 50W per computer seems low. In my room, my 2 computers quickly heat the room more than if I have a 100W light bulb on (although part of it is light, so doesn't heat the room as much).
Just to do this calculation is intersting: it shows the relative weight of some human activities. I'd also really like to have an accurate view of the electrical consumption of my computers.
And since I prefer cold juice to tea or coffee, I don't take part as much as you to the global warming. My drink gives back more heat than it absorbs
I think what they want to do is, given the state in 1950 and what we know about the inputs (use of automobile, etc.), which models predicts correctly what we see in 2000, so that those models can then be used, along with some new inputs, to forecast what 2050 will be. Either prospective inputs, to get a glimpse of our possible futures, or actual inputs, to further validate the model or get a sharper view of the future climate.
That being said, 8 months is way too long to get something useful. I know a couple friends who reinstall their OS (and apps) in shorter terms than that, and don't really bother with bringing all data along, just some backup on CDR "in case I really want it again". I think they could at least chop it in periods of a few years, so that if you finish a "unit", somebody else can then pick up where you left. I'd like to see the completion efficiency of whole units in a few months.
I see you have gcc-2.96 for C, java, C++, FORTRAN, objective-C and chill (although I don't know that last language). I also see you have egcs-1.1.2 for compiling C, C++, FORTRAN and obective-C code for a RH-6.2 target. You also have the libs to run those. If you check the 6.2 release, you can probably find some compat packages for targetting RH-5.2. And you can even install them if you want to target such an old release. You can probably even find some compat-libc5 rpms in RH-5.2. But in day to day use, or for compiling something for a RH-7.3 system, I don't see why you'd use egcs-1.1.2.
Back to kgcc... I don't see a kgcc in your rpm -a. By checking on a RH mirror, the only distro I can find with kgcc-1.1.2-40 is RH-7.0. After that, they always shipped only gcc-2.96-something (and it has been used extensively for kernels), except in 7.2 where gcc3 was also included. Unless you call kgcc something packaged as compat-egcs, there's no kgcc in 7.1, 7.2 nor 7.3.
Regarding the presence of gcc3 in RH-7.2, wasn't it meant as a technology preview only? Which explains both it's absence of 7.3 and the fact that it was a patched CVS checkout...
Of course, the other distributors can put whatever they want in their distros. The same way that some prefer CUPS and others LPRng. As long as they do the same thing...
A lot of the GCC 3 is broke with regards to the C++, that's a crock.
There you have it! The C ABI is pretty standard, and doesn't change between compiler releases. But the C++ side does. That's why you need a different libstdc++ for each compiler release. It's possible to upgrade the C compiler of a system easily (ICC, the Borland one, GCC). The C++ compiler is a bit more work, and RH prefered to not do it right now, but when GCC 3 will be ready for prime time (whether it's 3.0.4, 3.1, PI or 3.2 is not really relevant to the discussion).
Your suggestion of having 3 or 4 different releases of the same program is not a good idea: do you really expect users and developpers to choose between the different compilers dependant on the project they compile? "Do sdjy 0.58 needs EGCS 1.1.2, or is it safe to use 3.0.4? Oh right, it's 2.95.2." The goal of RH is to have only one: for the time being it's their own release, tagged 2.96-110, but it should evolve to a 3.?? based compiler for 8.0.
The other thing to keep in mind is that the gcc they ship now in 7.3 is tagged 2.96-110. Which means there's been a lot of patches applied since the beginning! When people grunt against gcc-2.96, they seldomly mention the RPM release. So yes, probably the first ones had a lot of problems (which explains why they shipped a kgcc in 7.0), but the current version works quite fine for me, even for kernel compilation (which explains why kgcc disappeared).
In one of the James bond (sorry, don't recall which one), James has a watch with a "magnetic beam" which he uses to move a car, and to undress a woman :) The first time is the more difficult to pass by, as he holds his watch with only 2 fingers and is still able to move (at distance) a car.
Can somebody enlighten me?
What is his copyrighted work for which he sued AOL?