| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2023-07-25 | allow to change the logging style; introduce some new ones | Omar Polo | |
| add `log style <style>'; The old default is called `legacy' now, a new default format is added called `condensed', and `common' and `combined' to mimick Apache httpd and nginx (respectively) are also added. | |||
| 2023-07-25 | update/sync/correct ChangeLog | Omar Polo | |
| 2023-07-25 | add missing -include titan.d | Omar Polo | |
| 2023-07-24 | still respect `log' when in debug mode | Omar Polo | |
| 2023-07-24 | add a manpage for titan(1) | Omar Polo | |
| 2023-07-24 | titan: parse the response code and exit accordingly | Omar Polo | |
| Exit with 0 if the response code was in the 2x or 3x range, or with 2 for other codes. It already exits with 1 upon any other error (including parsing errors.) Print the redirect code on 3x to stdout and the meta to stderr for the 1x, 4x, 5x and 6x ranges. | |||
| 2023-07-24 | bump man date | Omar Polo | |
| 2023-07-24 | run tls_connect_socket() after dropping the "inet dns" pledge promises | Omar Polo | |
| 2023-07-24 | read from stdin if no file is given | Omar Polo | |
| 2023-07-24 | sync changelog | Omar Polo | |
| 2023-07-24 | open the log files inside the chroot | Omar Polo | |
| 2023-07-24 | reopen log files upon SIGUSR2 | Omar Polo | |
| 2023-07-24 | add a test for the file logging | Omar Polo | |
| 2023-07-24 | add ability to log to files with log access <path> | Omar Polo | |
| 2023-07-24 | logger use dprintf and a fd instead of a FILE | Omar Polo | |
| simplifies further handling. The stdio layer introduces its own buffering and for the logs I'd like to avoid it. fflush(3) is an option, but using a raw fd and dprintf(2) requires less code. | |||
| 2023-07-23 | sync changelog | Omar Polo | |
| 2023-07-23 | remove unused IMSG types | Omar Polo | |
| 2023-07-23 | avoid use-after-free | Omar Polo | |
| 2023-07-23 | fix a reduce/reduce conflict | Omar Polo | |
| location -> error and locopt -> fastcgi -> error both end up with a optnl that can be reduced to the empty string. | |||
| 2023-07-23 | add a test with fastcgi, locations and forceful disabling | Omar Polo | |
| 2023-07-23 | add `fastcgi off' to forceful skip fastcgi for a route | Omar Polo | |
| 2023-07-23 | adjust syntax in fastcgi test; add another test for the old syntax | Omar Polo | |
| 2023-07-23 | revamp fastcgi configuration: make it per-location | Omar Polo | |
| this revamps the syntax in the configuration to better match httpd(8) (and in general be less weird) and to allow per-location fastcgi configurations. the bare `param' is now deprecated, but for compatibility it acts like `fastcgi param' would do now. Same story for `fastcgi <pathÂ>'. | |||
| 2023-07-23 | move struct envlist and alist up | Omar Polo | |
| 2023-07-23 | plug memleak | Omar Polo | |
| 2023-07-23 | add titan to .gitignore | Omar Polo | |
| 2023-07-23 | revert previous | Omar Polo | |
| Somehow the compat for __dead is not working properly on macos (cirrus ci) since it complains that parse_debug() does not return a value in all control paths when it uses usage() (marked as __dead) to catch a wrong usage. | |||
| 2023-07-22 | use __dead instead of __attribute__((noreturn)) | Omar Polo | |
| 2023-07-22 | titan: add compat shims to compile on !OpenBSD | Omar Polo | |
| 2023-07-22 | gg: warn when the TLS layer is not closed properly | Omar Polo | |
| various servers are not handling correctly the close notify so for the moment don't turn this into an hard error but just warn. Hopefully, given some time, most servers will be fixed. while here, drop the gotos and just use a break to exit the main loop. | |||
| 2023-07-22 | add titan(1) -- a draft titan client | Omar Polo | |
| 2023-07-22 | split out iri.h from gmid.h | Omar Polo | |
| 2023-07-22 | fmt | Omar Polo | |
| 2023-07-22 | drop engine support | Omar Polo | |
| 2023-07-22 | remove the useless logging methods | Omar Polo | |
| it makes more clear where the magic is. adapted from the smtpd' ca.c diff. | |||
| 2023-07-02 | add compat for endian (now required by imsg) | Omar Polo | |
| 2023-07-02 | update imsg | Omar Polo | |
| 2023-07-01 | rename do_accept() -> server_accept() | Omar Polo | |
| 2023-07-01 | enrich the description of the server process | Omar Polo | |
| 2023-07-01 | typo | Omar Polo | |
| 2023-07-01 | change log_request to take the code and meta unpacked | Omar Polo | |
| don't know what i was smoking when I wrote log_request() like that... | |||
| 2023-07-01 | improve fcgi test: send more than one chunk of data | Omar Polo | |
| 2023-07-01 | parse (and log) the header from fastcgi | Omar Polo | |
| 2023-07-01 | multiple -v don't make it more verbose anymore | Omar Polo | |
| 2023-07-01 | change on fatalx -> log_warnx | Omar Polo | |
| we already check the validity of the format string, but still avoid a gratious fatal() at runtime. | |||
| 2023-07-01 | rename fmt_sbuf -> fmtbuf; make the buffer explicit | Omar Polo | |
| 2023-07-01 | add some ideas | Omar Polo | |
| 2023-07-01 | avoid needless strlen() | Omar Polo | |
| 2023-07-01 | simplify request handling | Omar Polo | |
| get rid of check_path(), it's overly complicated. Instead, inline open_file() in client_read() and rework open_dir() to just use openat() instead of the complicate dance it was doing. Simplify open_dir() too in the process: if the directory entry for the index is not a regular file, pretend it doesn't exist. | |||
| 2023-07-01 | use a function-local buffer for the canonical redirect | Omar Polo | |
 |
index : slackcoder/gmid | |
| Small and secure gemini server | Mirror |
| aboutsummaryrefslogtreecommitdiff |