This fixes issues where if xbps-create creates a new archive on a full
disk, it will write an incomplete archive and then later fails to be
read by xbps-rindex.
See the manual page:
XBPS-DIGEST(1) General Commands Manual XBPS-DIGEST(1)
NAME
xbps-digest - XBPS utility to generate message digests
SYNOPSIS
xbps-digest [OPTIONS] [FILE] [FILE+N]
DESCRIPTION
The xbps-digest utility generates message digests for specified FILE or
stdin if unset.
OPTIONS
-m, --mode mode
Sets the message digest mode. Supported: sha256. If unset, defaults
to sha256.
-h, --help
Show the help message.
-V, --version
Show the version information.
SEE ALSO
xbps.d(5), xbps-checkvers(1), xbps-create(1), xbps-dgraph(1),
xbps-fbulk(1), xbps-install(1), xbps-pkgdb(1), xbps-query(1),
xbps-reconfigure(1), xbps-remove(1), xbps-rindex(1), xbps-uchroot(1),
xbps-uunshare(1)
AUTHORS
Juan Romero Pardines <xtraeme@gmail.com>
BUGS
Probably, but I try to make this not happen. Use it under your own
responsibility and enjoy your life.
Report bugs at https://github.com/void-linux/xbps/issues
June 12, 2019
Signed-off-by: Juan RP <xtraeme@gmail.com>
A segfault in xbps-create(1) was found in:
https://github.com/void-linux/void-packages/pull/7524
xbps-create adds a file type string to the xentry for every file it
finds in `destdir`, files other than symlinks, regular files and
directories weren't handled and resulted in a segfault later when
when processing the xentry structures.
This commit adds checks for sockets and fifos and exits with an
appropriate error message or with a generic error message for every
other unhandled file.
Closes#66
This is the first commit in a series to make xbps-create build
reproducable packages.
xbps-create uses nftw(3) to walk through all files. As nftw doesn't
sort files it may happen that the resulting packages will have
different checksums due to differentiating orders of files on
different filesystems.
To solve this issue xbps-create uses xbps_directory_t, which is always
sorted, instead of a simple linked list.
The latter could be too expensive, and we only care for the data to be written
to the binary package because a bit later we rename() the temporary file,
which will ensure a data flush to disk.
There's no reason to make them absolute, simply store in the metadata
the target file as is. This vastly simplifies the code and makes all
test pass correctly.
