[9fans] cryptsetup
Hello everyone, I have problems in a new installed system with cryptsetup and cwfs64. checktag pc=8694 cw "/dev/fs/fscache"w'/dev/fs/fsworm" (17cc8) tag /path=Tfile/49361; expected Tdir/49358 This is a fresh installation (not strange move-copy-dump-whatever) so this tags inconsistencies are worrying me, because that can happen in the future again with important data. Anyone has have similar problems with cryptsetup and cwfs64? Can I trust cryptsetup? Is fossil more stable than cwfs64? I have read a lot in 9fans archive but I can't make any conclusion... Also, I don't find anything about cryptsetup outside 9front. Is it possible to use it in the original Plan9? Thanks in advance, trebol.
Re: [9fans] cryptsetup
> I have problems in a new installed system with cryptsetup and cwfs64. > > checktag pc=8694 cw "/dev/fs/fscache"w'/dev/fs/fsworm" (17cc8) tag > /path=Tfile/49361; expected Tdir/49358 > > This is a fresh installation (not strange move-copy-dump-whatever) so this > tags inconsistencies are worrying me, > because that can happen in the future again with important data. > > Anyone has have similar problems with cryptsetup and cwfs64? > Can I trust cryptsetup? > Is fossil more stable than cwfs64? I have read a lot in 9fans archive but I > can't make any conclusion... > > Also, I don't find anything about cryptsetup outside 9front. Is it possible > to use it in the original Plan9? did you do "ream main" first? i have no idea what cryptsetup is. - erik
[9fans] p9p win(1) body file
Hello, Suppose $winid is the id of an acme window running win(1). An echo ls > /mnt/acme/$winid/body will execute ls, i.e. the input to body will be treated as if we had just typed ls in $winid. P9p on the other hand, will show ls in $winid but not execute it. Code reads: win.c:372: case 'E':/* write to body or tag; can't affect us */ Is there a reason for the p9p behavior? iru
[9fans] cryptsetup
> did you do "ream main" first? i have no idea what cryptsetup is. cryptsetup is a command for manage encrypted partitions: http://man.aiju.de/8/cryptsetup I have repeated the installation, hoping I made a mistake adding my user to cwfs or something. I configured the system again, recovered my data from other computer again, and when I was so happy starting to work in acme organizing my dirs the whole window is filled with those errors of Tfile... expected Tdir... Before installing Plan9 I tested the disk with linux's badblocks (write-mode test), although I think in case of bad sectors the output would be something about I-O errors. I don't know what to do. I want a stable environment to work (I'm very happy with the idea of plan9 with worm, dump and mirror) and I am used to encrypt my partitions in openbsd and linux. I don't like the idea of have my data exposed. Regards, trebol.
[9fans] ape ip6, warnings, etc.
to support jeff sickel doing a new port of the python tip to plan 9, using the native ssl and ssh implementations, it looked like a good idea to do two things 1. modernize socket handling, 2. turn on warnings. (a late addition, based on chasing some bugs that should have been noticed by the linker.) for 1, the noticable changes are, - added struct addrinfo, getaddrinfo, freeaddrinfo, getnameinfo, gai_strerror, inet_ntop, inet_pton. - definition of EALREADY and ECONNRESET - sockets support ip6. for 2, the noticable changes are, - inttypes.h has moved from /sys/include/ape to /$M/include/ape so uintptr_t can be properly sized on both 32- and 64-bit machines - _C99_SNPRINTF_EXTENSION created two type signatures for snprintf, so it had to go. since c99 is now 10 years old, the c99 extension is assumed. - segattach, segbrk, rendezvous, _SEGATTACH, _SEGBRK, and _RENDEZVOUS were convered to use void* and not int. on 64-bit machines, crashes had been observed for addresses above 0x8000 due to truncation and sign extension. - now declares the prototype as per posix extern void (*signal(int, void (*)(int)))(int); - the following functions were also declared int gethostname(char *, size_t); extern voidrerrstr(char*, unsigned int); /* gross */ extern int getfields(char*, char**, int, int, char*); /* gross */ the source is at /n/atom/plan9. if you have an atom install, you can pull as glenda and get the new stuff. a new iso will be out shortly. this may be of use if you are porting other ape stuff. note that the 9atom ape currently has a bit of skew with the distribution, which should be fixed soon. 9atom ... - declares ssize_t pread(int, void*, size_t, off_t), and pwrite(). - Runes are 21-bits so that awk plays nicely with others. i don't think this should cause any trouble, i certainly would appreciate any reports of how this works for folks, especially with contrib packages. as the intention is to continue to hone this for inclusion in the distribution so we can all have python imported from the tip. enjoy - erik
