X-Git-Url: http://www.git.cypherpunks.ru/?a=blobdiff_plain;f=doc%2Fnews.texi;h=f0e71f9380879a3f3a8acd358fafaea671ebb3cf;hb=a1443e916b45dca888ba0931c710d5560dbb6a9b;hp=e737a52fe347649399554d30830619f237ca1707;hpb=54b0f8a0e20847d666dd445bd92c282fd9ab5dec;p=nncp.git diff --git a/doc/news.texi b/doc/news.texi index e737a52..f0e71f9 100644 --- a/doc/news.texi +++ b/doc/news.texi @@ -3,6 +3,82 @@ See also this page @ref{Новости, on russian}. +@node Release 8_0_1 +@section Release 8.0.1 +@itemize + +@item +Fixed incorrect @code{freq.chunked} value calculation. If it missed, +then anyway chunked transfer mode was forcefully turned on. + +@item +Fixed incorrect full file's size calculation in @file{.nncp.meta}. + +@end itemize + +@node Release 8_0_0 +@section Release 8.0.0 +@itemize + +@item +@strong{Incompatible} encrypted packet format change: payload and pad +sizes are sent in-bound in the encrypted stream. That gives ability to +streamingly create encrypted packets, without knowing sizes in advance, +without creating temporary file or buffer data in memory. + +@item +Proper encrypted packet padding verification is done now. This is not +critical issue, but previously neither padding value, nor its size were +authenticated, giving ability to iteratively strip trailing bytes and +determine payload's size by observing the reaction of the encrypted +packet processing. + +@item +@command{nncp-exec} loses its @option{-use-tmp} option, because of +streaming-compatible encrypted packets format. + +@item +@command{nncp-file} and @command{nncp-exec} commands have +@option{-maxsize} option, limiting maximal resulting encrypted packet's +maximal size (returning error if it is exceeded). Could be useful, +because no payload size could be known in advance. + +@end itemize + +@node Release 7_7_0 +@section Release 7.7.0 +@itemize + +@item +Experimental @code{kqueue} and @code{inotify} based notifications +support about spool directory changes, for reducing their often reading +overhead. + +@item +@file{.seen} and @file{.hdr} files moved to @file{seen/} and @file{hdr/} +subdirectories, for faster scanning of spool directories. +Current files migration required: + +@example +$ find $NNCPSPOOL -type f -name "*.hdr" -exec rm @{@} + + +$ find $NNCPSPOOL -type d -name rx | while read rx ; do + cd $rx + mkdir -p seen + find . -type f -name "*.seen" | while read fn ; do + mv $fn seen/$@{fn%.seen@} + done +done + +$ find $NNCPSPOOL -type d -name area | while read area ; do + find $area -type f -name "*.seen" | while read fn ; do + mv $fn $@{fn%.seen@} + done +done +@end example + +@end itemize + @node Release 7_6_0 @section Release 7.6.0 @itemize