Commits


portable: release 0.111


bump version number


CHANGES for 0.111


fix gotsysd regress 'make vm' in case bsd.rd is looking for an upcoming release


ensure that gotsysctl gets installed into /usr/local/sbin


sync dist-file list


switch to UID 600 for _gotsysd to match the ports tree


remove a duplicate imsgbuf_clear() call


remove another gotsysd todo item that is done


make gotsys-apply-conf use gotd reload instead of stop + restarting


set close-on-exec flag on more files used during gotd reload


In parse_ref_file, release lock when fstat fails It's released in case of other errors. This seems to be an oversight. ok stsp@


Explicitly test S_ISREG in parse_ref_file parse_ref_file previously detected directories in the refs directory by trying to read and checking for EISDIR. This doesn't work on NetBSD, and the S_ISREG should exclude other kinds of non-regular files. ok stsp@


portable: add a release script This script automates the creation of the dist tarball, including verifying the tarball by unpacking and compiling it. Note that, the CHANGELOG should have been filled out already prior to running this script, so that it's included in the tarball; but that commit will need amending with the updated changes this script creates. Also assumes ed(1) is installed which is slightly more portable (and available) across different *nix systems, unlike sed(1)'s -i flag.


portable: gotwebd: handle sha{1,2}.h portably


portable: handle SOCK_CLOEXEC portably


portable: fixup after merge Fix the recent gotwebd changes to use common libexec helpers.


got-fetch-http: improve handling of HTTP chunked responses Use a cursor inside the buffer to parse the reply and handle the chunking delimiters instead of memmove()'ing around memory all the time. Spotted after a report from Colin Percival on IRC (thank you!) about got-fetch-http hogging the CPU while cloning the FreeBSD src.git repository. While here also simplify the chunked/non-chunked distinction and teach http_read() how to directly copy the data to a FILE. tested also by Colin, ok stsp@


Plug some memory leaks in got-{send,fetch}-pack my_capabilities was leaked in both got-send-pack and got-fetch-pack and needed freeing it in both. ok op@


revert part of previous commit which wasn't actually needed for memleak checks


fix all remaining gotwebd leaks the regress memleak checker was flagging


free gotweb.c iev_server on shutdown


also clear iev_gotweb->ibuf before exit to appease the memleak checker


free iev_gotweb before exiting; harmless leak but 'make regress' flags it


reduce the send_response() timeout For some reason this timeout was blocking long enough to make other waiting client connections time out. Reduce the amount of retries.