aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* CMakeLists.txt: fix copy-pasted variable namePřemysl Eric Janouch2020-10-301-2/+2
|
* json-rpc-shell.adoc: minor improvementsPřemysl Eric Janouch2020-10-301-0/+8
| | | | Documented envvars and added a note about XDG paths.
* CMakeLists.txt: make this build in OpenBSDPřemysl Eric Janouch2020-10-301-1/+19
|
* CMakeLists.txt: omit end{if,foreach} expressionsPřemysl Eric Janouch2020-10-301-7/+7
| | | | Their usefulness was almost negative.
* Bump minimum CMake version to 3.0Přemysl Eric Janouch2020-10-303-23/+8
| | | | A nice, round number. This allows us to remove some boilerplate.
* CMakeLists.txt: fix an outdated commentPřemysl Eric Janouch2020-10-301-1/+1
|
* Bump libertyPřemysl Eric Janouch2020-10-301-0/+0
|
* json-rpc-test-server: only return regular filesPřemysl Eric Janouch2020-10-171-1/+6
| | | | They can be symlinked.
* json-rpc-test-server: comment on some CGI detailsPřemysl Eric Janouch2020-10-171-8/+33
| | | | | There are some unresolved issues in the CGI clients that needed a more precise description.
* json-rpc-test-server: fix function namesPřemysl Eric Janouch2020-10-161-4/+4
| | | | Very obviously copied and pasted from the shell.
* json-rpc-test-server: try to send a 408Přemysl Eric Janouch2020-10-151-14/+15
| | | | | | | | | Also send "Connection: close" when we're closing the connection. With HTTP/1.1 there come some responsibilities. Surprisingly enough, the forward declaration is desirable and the invocation a clean-up.
* WebSocket: send a User-Agent headerPřemysl Eric Janouch2020-10-151-0/+2
|
* json-rpc-test-server: add a simple co-process modePřemysl Eric Janouch2020-10-153-7/+179
| | | | | | A disgusting copy-paste but it will have to do for now. Closes #6
* json-rpc-test-server: WS: fix failures to upgradePřemysl Eric Janouch2020-10-151-5/+8
| | | | Similar to ad1aba9, only here we return 426 to the client.
* WebSocket: adapt to common "await" infrastructurePřemysl Eric Janouch2020-10-141-79/+9
|
* Add a backend for co-processesPřemysl Eric Janouch2020-10-143-11/+439
| | | | | | | | | Targets language servers. In this first stage, we don't need to support bi-directionality, although it's a requirement for finishing this task. Updates #4
* This software is no longer simplePřemysl Eric Janouch2020-10-144-11/+11
|
* WebSocket: some clean-upPřemysl Eric Janouch2020-10-141-37/+26
|
* Split out a http-parser wrapperPřemysl Eric Janouch2020-10-141-111/+152
|
* Simplify the FAIL macroPřemysl Eric Janouch2020-10-141-5/+1
|
* WebSocket: fix upgrade processingPřemysl Eric Janouch2020-10-141-14/+6
| | | | | When http-parser sets the upgrade field, it checks for status code 101 and even resolves our TODO about checking the entire Connection header.
* json-rpc-shell.adoc: document the M-Enter bindingPřemysl Eric Janouch2020-10-141-1/+1
|
* WebSockets -> WebSocketPřemysl Eric Janouch2020-10-144-15/+15
| | | | | This is the correct name of the protocol, usage of the word "WebSockets" should be limited.
* Support reading OpenRPC documents from a filePřemysl Eric Janouch2020-10-133-20/+31
| | | | Bump liberty, it generated incorrect help messages.
* Factor out init_backend()Přemysl Eric Janouch2020-10-131-27/+31
| | | | The main() function is still way too long.
* json-rpc-test-server: implement rpc.discoverPřemysl Eric Janouch2020-10-131-2/+35
|
* CleanupPřemysl Eric Janouch2020-10-131-14/+17
| | | | Prevent the last fuck-up from happening again.
* Add libedit autocompletion back inPřemysl Eric Janouch2020-10-131-0/+4
| | | | I've mistakenly removed it in the M-Enter change.
* Bump version, update NEWSv1.1.0Přemysl Eric Janouch2020-10-132-1/+12
|
* libedit: bind M-Enter to newline-insert as wellPřemysl Eric Janouch2020-10-131-5/+12
|
* json-rpc-test-server: fix reading the request URIPřemysl Eric Janouch2020-10-131-1/+1
|
* Bump http-parserPřemysl Eric Janouch2020-10-131-0/+0
| | | | | Apparently it's reached maturity and there won't be any changes anytime soon, making this the perfect time for an upgrade.
* OpenRPC: avoid eating HTTP/transport errorsPřemysl Eric Janouch2020-10-131-15/+22
|
* Document OpenRPC tab completion supportPřemysl Eric Janouch2020-10-132-0/+5
| | | | | | | | Now that it's functional in both frontends, we can flaunt it. I still don't want to make it the default. Closes #1
* Implement tab completion under libeditPřemysl Eric Janouch2020-10-131-23/+137
| | | | | | | | | | | I haven't tested it with real wide characters but it will have to do. I wasn't even sure if this piece of crap could be coerced into doing this at first, so it's a win for me. It uses a variation of the code in degesch where we /don't/ want to print the list of candidates on partial failure. Updates #1
* json-rpc-test-server: fix a potential memory leakPřemysl Eric Janouch2020-10-131-0/+1
|
* json-rpc-shell.adoc: update WebSocket notesPřemysl Eric Janouch2020-10-101-2/+1
| | | | | https://github.com/open-rpc/client-js also uses WebSockets, although they don't seem to support notifications (in general).
* Bump libertyPřemysl Eric Janouch2020-10-102-59/+8
| | | | resolve_relative_runtime_unique_filename() used to have a bug.
* Readline: add trivial OpenRPC supportPřemysl Eric Janouch2020-10-101-6/+108
| | | | So far hidden under a switch and only for this frontend.
* Reorder help message entries a bitPřemysl Eric Janouch2020-10-101-6/+6
| | | | Should be both more useful and more alphabetic this way.
* Fix use of possibly uninitialised memoryPřemysl Eric Janouch2020-10-101-1/+1
|
* Split make_json_rpc_call() in halfPřemysl Eric Janouch2020-10-101-13/+24
|
* Slightly refactor make_json_rpc_call()Přemysl Eric Janouch2020-10-101-36/+29
|
* Readline: make M-Enter insert a newlinePřemysl Eric Janouch2020-10-091-0/+12
| | | | | | | Before, it was only possible with C-v C-j but it's too useful to require such an awkward method. There is a precedent in, e.g., zsh and fish for the new binding.
* Fix handling terminal resizes while the terminal is suspendedPřemysl Eric Janouch2020-10-091-0/+2
| | | | GNU Readline has a misfeature.
* Readline: fix a dormant bug in prompt changesPřemysl Eric Janouch2020-10-091-5/+3
| | | | For details, see a similar change in degesch from uirc3.
* Make a release, create NEWSv1.0.0Přemysl Eric Janouch2020-09-052-2/+6
|
* Streamline the manual page a bitPřemysl Eric Janouch2020-09-051-11/+11
| | | | I have consulted `man 7 man-pages` but overall it's a huge mess.
* Use saner defaultsPřemysl Eric Janouch2020-09-052-32/+39
| | | | | | | | So that most of the time users won't need to use any switches. --pretty-print has been inverted into jq's --compact-output, and --auto-id has been replaced with barely, if-at-all useful --null-as-id.
* Improve AsciiDoc compatibilityPřemysl Eric Janouch2020-09-051-25/+19
| | | | | | | | | | | | | I need two renderers to work: hswg/libasciidoc and asciidoctor in man page mode (and ideally in HTML as well). That should be covered now. The triple-plus thing was the first thing that showed good results, after trying backslashes, single-plus quoting and [] after ://. The change of the source code block kind could be considered as unification. I'm combining tabs with spaces within one document though, and I should get rid of the tabs in the rest of it then...
generated by cgit v1.2.3-70-g09d2 (git 2.47.0) at 2024-11-22 02:24:52 +0000