I was also active on the project in the early , and wrote a lot of the early code with Werber Almesberger (mostly known for lilo) .
Although I was working on the goefox - which I bought because I thought
it would be better for linux that the psion 5 . There was a lot of discussion back then (around 1998 for those who want to check the archives) about trying to persuade psion to give help.
So it isn't excactly wrong to say psion weren't helpful -they didn't really seem to grok the linux / open source philposhies.
OTOH Maybe we (As in the linux-7k project) managed to alert the psion management to linux in early enough that they can now see the value of helping us.
Note that the AT&T version of this code is also probably old enough to appear in the Lyons book.
A quick check in my 2nd ed of k & R show something very similiar
in there too.
This isn't surprising. We're discussing a simple allocator after
all - a pretty fundemental algorithm in all of comp sci.
Indeed you might even/copy/ this code if you had a version (license allowing) lying around as that one is already debugged, and even if you don't your debugged version is going to turn out within the level of differences discussed here.
hell this code prolly appears in many compsci students lecture notes as well
I wasn't able to get the card to work with XFree86 v4.2.0, but it works in 2D with 4.3.0
That's not surprising . Use the web interface to check the commit
comment in the ati drivers sources.
I did sometime ago looking for support for my radeon U1 and found that
someone with a @ati.com mail address had submitted source patches to xfree. These were 2d only patches.
<speculation mode=rampant>
I guess ati may only be supporting the drivers throught the xfree project now. That's a good thing in my book.
Firebird is a common term, searching for "Firebird" isn't going to turn up their database without "SQL" attached, so who cares?
You didn't actualy try this in goggle did you?
Also, where would you expect www.firebird-conference.com to be about ?
I agree with the previous poster that the story is not about the legal side, but about the politiness side. It seems mozilla should have contacted the firebird
core team in advance, and they didn't seem to on any of the firebirdsql lists i'm subscribed to.
However I feel the behaviour and calls to mail bomb do rather let down the firebirdsql project.
Without that at least the firebirdsql project could claim the moral high ground, but that been ceeded now.
Using clashing names will cause a real problem in some other places to, I don't what to have to think
too hard about what
apt-get source firebird
is going to do. Whats worse is I plan to deploy elements from both projects so this spat over names really does make my life harder, as well as the packaging names issue , there the confusion between developers after all saying "it appears to be a bug in firebird" to co-developers doesn't narrow stuff down much anymore...
IIRC the 'backdoor' was rarely changed default system password. This combined with bootstrap caused some interesting behaviour forming a backdoor.
In Interbase a seperate database schema is used to store database username and password pairs, unfortunely you can't access that until you are authenticated, so a backdoor was added so get round this problem.
BTW, this is all from memory so check the archives
before taking what i've said as gospel.
the "check number" is essentially the electrical meter on your house. meaning that, at least in some way, the number can be changed, at least by you moving elsewhere. furthermore, there is no reason why anyone else would have that number,
theoretically. it's a semi-constant
Admittely ODBC is/was gennerally improved AFAICT by microsoft, it is essentially still an X/Open standard.
It is availble on many platforms, Mac/VMS/Unix/Windows and probably others too. It is a relative striaghtforward C API for database access. Ok, native access could be quicker but I think you'd find difficulty building a thinner layer for all databse engines.
Lets be honest about this two in many cases I reckon you will find OLE DB implement on top of the ODBC drivers . Not that I've ever used OLE DB being a crossplatform developer.
Yes but in the specialist envoironment of scientific instruments and data capture you could
provide a custom case as well. Which although not an ideal solution is a lot better than nothing.
And there is always the option of using the slots that the orignal case have that are beyond the
'end' of the motherboard. Given the low speed of the ISA bus this is strikes me at first thought as an attractive soln.
Eg, mount a card and alongside the edge of the motherbaoard lining up with the slots which can't
mate with the motherboard. Put a PCIISA adaptor on a PCI card. Then ribbon cable between the PCI can with all the birdge electronics on it and the ISA extended card.
It can't be that difficult, I'm pretty sure We'd be interesting in taking the job on as consultancy if no one else was.
And quite frankly I don't believe you will fine it that difficult, you will however find the design costs, very costly the first time round - but you only have to do those once.
It probably isn't economic in the mass market sense though.
linux hasn't got strlcpy. but it has got strncpy,and so have windows and Irix being two other OS's I can check easily.
Unfortunately a quick google for strlcpy() gave me
this
Which shows it to better in a number of instances than strncpy()
To be portable almost all my apps have this anyway which solves similiar issues to strlcpy.
True. Although to be fair we don't know that
KMail didn't set the file +x .
What we can take away from this though is that
it would be a good idea to implement at least
on option (if not make it mandatory) such that
files without +x are not run by interpreters.
If your feeling nice I suppose you might let
file without +x run in a sandbox - but I think not at all is better.
Then programs such as KMail can indicate the the these files are not safe for execution by not setting them +x
Slashdot Mirror
Although I was working on the goefox - which I bought because I thought it would be better for linux that the psion 5 . There was a lot of discussion back then (around 1998 for those who want to check the archives) about trying to persuade psion to give help.
So it isn't excactly wrong to say psion weren't helpful -they didn't really seem to grok the linux / open source philposhies. OTOH Maybe we (As in the linux-7k project) managed to alert the psion management to linux in early enough that they can now see the value of helping us.
A quick check in my 2nd ed of k & R show something very similiar in there too.
This isn't surprising. We're discussing a simple allocator after all - a pretty fundemental algorithm in all of comp sci.
Indeed you might even /copy/ this code if you had a version (license allowing) lying around as that one is already debugged, and even if you don't your debugged version is going to turn out within the level of differences discussed here.
hell this code prolly appears in many compsci students lecture notes as well
That's not surprising . Use the web interface to check the commit comment in the ati drivers sources.
I did sometime ago looking for support for my radeon U1 and found that someone with a @ati.com mail address had submitted source patches to xfree. These were 2d only patches.
I guess ati may only be supporting the drivers throught the xfree project now. That's a good thing in my book. Go buy more ati is my conclusion from that.Absoulutely , It would have given us eutopeand software authors a 17 year advantage (until the patent expired) over you americans.
Maybe our beancounters would have seen computers as useful by then.
The top hit I got was: Firebird - Relational Database for the New Millennium
Also, where would you expect www.firebird-conference.com to be about ?
I agree with the previous poster that the story is not about the legal side, but about the politiness side. It seems mozilla should have contacted the firebird core team in advance, and they didn't seem to on any of the firebirdsql lists i'm subscribed to. However I feel the behaviour and calls to mail bomb do rather let down the firebirdsql project.
Without that at least the firebirdsql project could claim the moral high ground, but that been ceeded now.
Using clashing names will cause a real problem in some other places to, I don't what to have to think too hard about what
is going to do. Whats worse is I plan to deploy elements from both projects so this spat over names really does make my life harder, as well as the packaging names issue , there the confusion between developers after all saying "it appears to be a bug in firebird" to co-developers doesn't narrow stuff down much anymore...I guess you are referring to Interbase here.
IIRC the 'backdoor' was rarely changed default system password. This combined with bootstrap caused some interesting behaviour forming a backdoor.
In Interbase a seperate database schema is used to store database username and password pairs, unfortunely you can't access that until you are authenticated, so a backdoor was added so get round this problem.
BTW, this is all from memory so check the archives before taking what i've said as gospel.
Er, except for social engineering...
Yeah lets screw open standards.
Admittely ODBC is/was gennerally improved AFAICT by microsoft, it is essentially still an X/Open standard.
It is availble on many platforms, Mac/VMS/Unix/Windows and probably others too. It is a relative striaghtforward C API for database access. Ok, native access could be quicker but I think you'd find difficulty building a thinner layer for all databse engines.
Lets be honest about this two in many cases I reckon you will find OLE DB implement on top of the ODBC drivers . Not that I've ever used OLE DB being a crossplatform developer.
Yes but in the specialist envoironment of scientific instruments and data capture you could provide a custom case as well. Which although not an ideal solution is a lot better than nothing. And there is always the option of using the slots that the orignal case have that are beyond the 'end' of the motherboard. Given the low speed of the ISA bus this is strikes me at first thought as an attractive soln.
Eg, mount a card and alongside the edge of the motherbaoard lining up with the slots which can't mate with the motherboard. Put a PCIISA adaptor on a PCI card. Then ribbon cable between the PCI can with all the birdge electronics on it and the ISA extended card.
It can't be that difficult, I'm pretty sure We'd be interesting in taking the job on as consultancy if no one else was.
And quite frankly I don't believe you will fine it that difficult, you will however find the design costs, very costly the first time round - but you only have to do those once. It probably isn't economic in the mass market sense though.
Unfortunately a quick google for strlcpy() gave me this
Which shows it to better in a number of instances than strncpy() To be portable almost all my apps have this anyway which solves similiar issues to strlcpy.
True. Although to be fair we don't know that KMail didn't set the file +x . What we can take away from this though is that it would be a good idea to implement at least on option (if not make it mandatory) such that files without +x are not run by interpreters. If your feeling nice I suppose you might let file without +x run in a sandbox - but I think not at all is better. Then programs such as KMail can indicate the the these files are not safe for execution by not setting them +x