aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2024-06-03fix `fastcgi off' handlingOmar Polo
When a matching location has a `fastcgi off' directive, we should honour that and stop searching for further location which may have a `fastcgi' directive. Bug reported by Alex // nytpu, thanks!
2024-05-29regress: add test_ipv6_serverOmar Polo
2024-05-29regress: add a knob to disable test_ipv6_addrOmar Polo
at least on the CI is failing with "can't connect to ::1:10965: Address not available" which suggests IPv6 is broken there.
2024-05-29typoOmar Polo
2024-05-29regress: rename ipv4 test and add another with ipv6Omar Polo
2024-05-29iri: add support for raw IPv6 addressesOmar Polo
2024-05-29gg: unbreak -nOmar Polo
2024-05-29regress: run test_ip_addr with host=127.0.0.1Omar Polo
2024-05-29explain why we disable runtime tests on macosOmar Polo
2024-05-29regress: use the new gg -q to reduce the blabberingOmar Polo
2024-05-29gg: add -q to avoid printing "Server says"Omar Polo
2024-05-29pretty-print the socket address at configuration parsing timeOmar Polo
saves a getnameinfo(NI_NUMERICHOST) at runtime, even if it's pretty cheap.
2024-05-29add regress that hit gmid via a raw IPv4 addressOmar Polo
2024-05-29relax the SNI requirementOmar Polo
There are legitimate cases where SNI can't be used, for example when connecting via an IPv6 address, so don't rejects those requests. Instead, fill the requested domain with the address (literal) of the socket they're connected to and attempt to match on it. This possibly still incur in a "won't proxy" error if the client then requests a different hostname. See the github issue https://github.com/omar-polo/gmid/issues/25
2024-05-27github: add workflow to build images for ghcr.ioOmar Polo
2024-05-25s/MIN/MINIMUM/gOmar Polo
2024-04-27contrib/gmid.service: remove User and GroupOmar Polo
May cause weird errors (status=216/GROUP) on some distros, and running as root is already the default, so remove the two lines. Reported by and debugged together with leandro del Flug, thanks!
2024-04-27contrib/gmid.service: start as root by defaultOmar Polo
Various techniques used by gmid are effective only when the daemon is started as root. Strongly suggest to do so by switching the sample configuration. This way, provided that a local user is created as well, the chroot configuration will work out-of-the-box and the TLS certificates can be readable only by root.
2024-04-27improve the description for -fOmar Polo
2024-04-11typo: semicolors -> semicolonsOmar Polo
2024-04-04set next versionOmar Polo
2024-04-04prepare release 2.0.12.0.2Omar Polo
2024-04-04changelog for 2.0.2Omar Polo
2024-04-04add a test for the config dumpingOmar Polo
2024-04-04fix config dumping (-nn) handlingOmar Polo
with the privsep rework the config dumping was unadvertitely broken, it prints the content of the key itself.
2024-04-03use -Werror=implicit-function-declaration for function detectionOmar Polo
the previous -Werror triggers too easily: on NixOS for example the FORTIFY_SOURCE #warning about a missing optimization level breaks all the checks when using -O0 (which is the default for non-release builds).
2024-04-03fix landlock testOmar Polo
include stddef.h for size_t
2024-03-12remove dead codeOmar Polo
2024-03-06update mac CI target to sonomaOmar Polo
2024-03-03fix `log access path' with chrootOmar Polo
We should open the log file inside the chroot; missed in 4acf495f41d2c26136e99072293ca8b9bff91dc0. See https://github.com/omar-polo/gmid/issues/24
2024-03-03revert 9f1cce3d0e53209180eabbcd48878c1e989101e7Omar Polo
we actually should open the log file in the chroot, the bug is in the code.
2024-03-03fix log access doc: path is not relative to the chrootOmar Polo
Reported by Colin Henein, thanks! See https://github.com/omar-polo/gmid/issues/24
2024-02-02add a note regarding the usage of the bundled libtlsOmar Polo
See https://codeberg.org/op/gmid/issues/2
2024-01-30sync changelogOmar Polo
2024-01-30contrib/vim: fix indentAnna “CyberTailor”
2024-01-30rename the @common_opt macro back to @commonOmar Polo
now common is no longer a reserved keyword
2024-01-30turn log styles into strings from yacc point of viewOmar Polo
having styles as reserved keywords means that variables / macros can't be called `common', `condensed', etc... which is not great and not obvious either. Instead, let's keep the log styles as strings and match on them. This also allows to have a slightly better error message in case of a typo. See: https://codeberg.org/op/gmid/issues/1
2024-01-30remove stray spaceOmar Polo
2024-01-26tweak and update freebsd taskOmar Polo
2024-01-26rework the grammar so that ; is accepted after variables and optionsOmar Polo
See Codeberg issue #1.
2024-01-26add a test that uses @-style macrosOmar Polo
See Codeberg issue #1.
2024-01-24change the default PUBKEY for the verify-release targetOmar Polo
doesn't play well with minor releases such as 2.0.1 since for them I reuse the 2.0 key.
2024-01-24set next versionOmar Polo
2024-01-24prepare release 2.0.12.0.1Omar Polo
2024-01-24changelog for 2.0.1Omar Polo
2024-01-21convert remaining code to the imsg gettersOmar Polo
Now gmid doesn't touch anymore the internals of the imsg structs.
2024-01-21convert the remaining bit of crypto.c to the ibuf_* APIsOmar Polo
2024-01-21please macosOmar Polo
for some reason that's not entirely clear to me, __dead doesn't seem to work on macos, so clang thinks datalen is used un-initialized. meh
2024-01-21update imsg test: gmid now requires the new API tooOmar Polo
2024-01-21convert crypto.c to the new imsg APIOmar Polo
generated by cgit v1.2.3 (git 2.26.2) at 2024-11-24 11:03:25 +0000