From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5) Gecko/20031022 Firebird/0.7 Description of problem: http://www.redhat.com/archives/fedora-devel-list/2003-November/msg00328.html ftwall-1.07 built fine in RH9, but fails build in FC1 due to changes in the contents of /usr/include/linux/tcp.h. Read the above post for details. http://www.redhat.com/archives/fedora-devel-list/2003-November/msg00338.html Roland McGrath at redhat.com believes the below tcp.h changes to be broken. --- tcp.h-rh9 2003-11-06 21:45:40.000000000 -1000 +++ tcp.h-fc1 2003-11-06 21:46:08.000000000 -1000 @@ -102,16 +102,16 @@ #define tcp_flag_word(tp) ( ((union tcp_word_hdr *)(tp))->words [3]) enum { - TCP_FLAG_CWR = __constant_htonl(0x00800000), - TCP_FLAG_ECE = __constant_htonl(0x00400000), - TCP_FLAG_URG = __constant_htonl(0x00200000), - TCP_FLAG_ACK = __constant_htonl(0x00100000), - TCP_FLAG_PSH = __constant_htonl(0x00080000), - TCP_FLAG_RST = __constant_htonl(0x00040000), - TCP_FLAG_SYN = __constant_htonl(0x00020000), - TCP_FLAG_FIN = __constant_htonl(0x00010000), - TCP_RESERVED_BITS = __constant_htonl(0x0FC00000), - TCP_DATA_OFFSET = __constant_htonl(0xF0000000) + TCP_FLAG_CWR = htonl(0x00800000), + TCP_FLAG_ECE = htonl(0x00400000), + TCP_FLAG_URG = htonl(0x00200000), + TCP_FLAG_ACK = htonl(0x00100000), + TCP_FLAG_PSH = htonl(0x00080000), + TCP_FLAG_RST = htonl(0x00040000), + TCP_FLAG_SYN = htonl(0x00020000), + TCP_FLAG_FIN = htonl(0x00010000), + TCP_RESERVED_BITS = htonl(0x0FC00000), + TCP_DATA_OFFSET = htonl(0xF0000000) }; /* TCP socket options */ Version-Release number of selected component (if applicable): glibc-kernheaders-2.4-8.36 How reproducible: Always Steps to Reproduce: 1. adfaf 2. 3. Additional info:
I wanted to compile shfs. I ran into the "headers are hosed in glibc-kernheaders" bug, so I went looking for an upgrade. None were in any of the yum repositories, but I found glibc-kernheaders-2.4-9.1.84.i386.rpm at a shadow in Czechoslovakia, (yes, the signature checked) and, of course, it will still not compile, but for other reasons. Why is this not considered important enought to release a fix for? Or even to have an update in updates-testing? If the places in yum.conf or their shadows are wrong, where should I be looking?
shfs is a kernel module; it cannot use the headers from glibc-kernheaders anyway!
This seems fixed in FC3.