![]() |
Mr. Arlington Hewes wrote: >>>>> On Wed, 3 Jun 1998, "MA" == Matthias Apitz wrote: MA> Mr. Arlington Hewes wrote: +> >> > > On Wed, 3 Jun 1998, "MA" == Matthias Apitz wrote: MA> The answer is simple: MA> Matthias, that is not simple . . . it's simple to you. The distinction is MA> worth bearing in mind. MA> You truncated my original e-mail to mutch. I was saying that the reason MA> is a failing chroot(2) or a failing chdir(2) sys call (and this *is* MA> simple to be grep'ed from the source). I'm sorry, but again you illustrate my point. Care to ponder the number of people who might be running a unix box and _NOT_ consider searching the source code for a string contained in the error? It's high dude. I know many people who would not think they have a hope in hell of understanding the gobbledeygook which might result from such a search. It is worth bearing in mind that, although it may not be a "good thing"(tm) there are those for whom: tar -xvzf tarball.tar.gz cd tarball ./configure make install is fairly challenging. Telling these people that the answer is simple if one just greps hfaxd.c++ for xxxx and concluding a failed chroot(2) or a failing chdir(2) is not necessarily going to make someone's day. correct, but this wasn't my message; someone who can't understand the output of the grep (together with the helping hand in my mail) will perhaps also not figure out the reaason for the failing sys calls (which perhaps has its reason in a misconfigured Linux distribution and/or environment for HylaFAX; remember this old discussion about uid of "fax"/"uucp"); my message should be: the first step (grep) is simple but the second (e.g. the bits between chroot(2) and the correct euid etc.) is much more difficult -- all is relativ; the bottom line is also: we will never find out the reason for this problem (showing up from time to time) if there is noone out there who wants to follow through this problem or give us more information or root access to the box if all other does not help; MA> The problem for the failing of both calls must be checked in the file MA> system of the original poster. Okay, now you've lost me. Since I'm in over my head, I will now bow out of this conversation, and assume that the original poster understands your reply, and what to do to debug the problem. My point was, is, and will continue to be, if we EXPECT people to try to figure out what's wrong with their setup by grepping source code, we're expecting too much. having the source is a *big* chance to figure out a problem; matthias