[Please don't drop the CC against [EMAIL PROTECTED] I'm quoting your entire message so that it gets archived as part of the bug log.]
Mark Hobley <[EMAIL PROTECTED]> writes: >> From: Ben Pfaff <[EMAIL PROTECTED]> > >> Please give a specific example of a package that is adversely >> affected by this bug. > > I am not sure of the extent of this. I suspect that this is a general > problem. I am not sure how many distribution packages are affected, > but I have noticed that some packages that produce verbose output > state that they are running 686 components. This is a problem for > organizations that require all machines to be running the same code. Please give a specific example, so that I can try to figure out what is going on. I do not think that Autoconf is at fault, but I would like to find out. > There is a problem with invalid instructions in the current kernel (in > Lenny). I am not sure whether this is due to processor detection > relating to this. I compiled a vanilla kernel on a 486 and this worked > fine. However, the kernel supplied by distribution does not work, so > this may be due to detection of the distributors higher specification > machine. Is the distributor using the pentium-builder kludge? The kernel does not use Autoconf, so this is not the problem. What "distributor" are you talking about? This is the Debian bug system, so bug reports against other distributions would not ordinarily be pertinent here. > It is a big problem for packages that I compile myself, using > Autotools, because the compiler keeps detecting the architecture as > 686 and building for a 686, but I really want 486, because I am > actually building shared binaries for use mainly on traditional > Pentium (and some Cyrix based) machines. Please give a specific example. > Ideally, I need all the processor detection systems to lie and say > that the processors are 486 (albeit extremely fast 486 processors), so > that all computers on the network can behave as fast 486 machines, and > the target architecture is treated as a fast 486 machine. This would > also alleviate the need for a pentium-builder kludge. What is the pentium-builder kluge? -- Ben Pfaff http://benpfaff.org -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
