249 commits

This branch

This branch

All branches

Author	SHA1	Message	Date
Simon Cruanes	0bd8868172	update test outputs	2026-01-14 22:47:08 -05:00
Simon Cruanes	4aee136827	make tests compile	2026-01-14 22:31:46 -05:00
Simon Cruanes	2a866e60f8	fix fuchsia and tef backends to use trace_id	2026-01-14 22:30:00 -05:00
Simon Cruanes	5b83834af5	subscriber: restore trace_id ... we need it for async backends, clearly	2026-01-14 22:29:46 -05:00
Simon Cruanes	c89a031e43	core: enter_span takes an optional flavor ... it's a core extension, it should be easy to use	2026-01-14 22:29:32 -05:00
Simon Cruanes	2cfb3c67fa	port fuchsia to new subscriber	2026-01-14 22:09:33 -05:00
Simon Cruanes	7389ca5b45	fix warning	2026-01-14 22:09:25 -05:00
Simon Cruanes	322e1d0f44	subscriber: make ~new_span_id mandatory ... too easy to forget otherwise	2026-01-14 22:09:05 -05:00
Simon Cruanes	7a392e54d1	trace-tef: use new subscriber, no more global state	2026-01-14 21:56:09 -05:00
Simon Cruanes	a1837e402e	trace.subscriber: define custom span, simplify subscriber callbacks ... - custom span carries around all the required data, including mutable args - no more manual enter/exit callbacks - use record for callbacks - no more big Span_tbl	2026-01-14 21:56:09 -05:00
Simon Cruanes	40b44349e7	core: add on_init callback to collector	2026-01-14 21:56:09 -05:00
Simon Cruanes	8e2bb5bc83	remove event for now	2026-01-14 21:27:33 -05:00
Simon Cruanes	5018a9ead7	refactor core library ... much cleaner now, and simpler. - `span` is an open sum type, so it can carry as much info as we want from enter to exit. This means most collectors shouldn't even need a global table for spans! - merge manual spans and spans. It's just spans now. Keep old Trace_core.xxx_manual_xxx functions but deprecate them - add `extension_parameter` to carry additional info to collector (eg OTEL span kind, TEF sync/async, maybe span links, etc) - make the collector into a record of callbacks + a state, for efficiency (but also to make it easier to extend in the future with callbacks that have a default implementation )	2026-01-14 20:29:33 -05:00
Vincent Bernardoff	1c9a869148	ppx: call Stdlib.Printexc ... This is because some libraries, i.e. Janestreet Core, override the Printexc module.	2026-01-10 08:20:55 -05:00
Simon Cruanes	5141d4bde4	style	2026-01-09 16:04:03 -05:00
Simon Cruanes	ef35cc1d79	Merge pull request #38 from c-cube/simon/fix-ppxlib-0.37 ... adapt to ppxlib 0.37	2025-12-12 08:54:29 -05:00
Simon Cruanes	3f21ea67ce	update deps to ppxlib=0.37~	2025-12-09 12:24:52 -05:00
Simon Cruanes	0947d2d523	better error message	2025-12-09 12:24:52 -05:00
Simon Cruanes	b3da5bc41e	deal with Pexp_function	2025-12-09 12:24:48 -05:00
Simon Cruanes	9b4a3855fb	update test	2025-12-09 12:24:48 -05:00
Simon Cruanes	89eecf7ba3	breaking: use poly variants for user_data/span_flavor in subscriber ... no need to do redundant conversions.	2025-12-04 12:39:03 -05:00
Simon Cruanes	aeb2aff3b7	breaking: require subscribers to provide mk_span/mk_trace_id ... we want control over this!	2025-12-04 12:31:14 -05:00
Simon Cruanes	15edb582d0	fix: relax bound on ppxlib	2025-11-01 22:18:25 -04:00
Simon Cruanes	acae4ff88d	doc	2025-11-01 22:18:16 -04:00
Simon Cruanes	c2a1ee5904	format	2025-09-15 10:26:34 -04:00
Simon Cruanes	9a77dad2fd	chore: CI	2025-09-15 10:23:19 -04:00
Simon Cruanes	0c275b3aab	use at_exit in trace_tef/tldrs	2025-09-15 09:11:09 -04:00
Simon Cruanes	a81785f8c0	fix fuchsia: bound check	2025-07-26 01:20:40 -04:00
Simon Cruanes	d9cd7621f5	prepare for 0.10	2025-05-27 09:58:19 -04:00
Simon Cruanes	cd5785d938	format	2025-05-27 09:54:58 -04:00
Simon Cruanes	6853fa50f3	docs	2025-05-27 09:54:43 -04:00
Simon Cruanes	ba40156f22	chore: addd a tag to the opam package	2025-05-13 09:09:38 -04:00
Simon Cruanes	e6b17c5536	Merge pull request #36 from c-cube/simon/fuchsia-via-subscriber-2025-05-02 ... full refactoring of TEF and fuchsia backends	2025-05-08 09:44:58 -04:00
Simon Cruanes	d1759fea89	fix for 4.08	2025-05-07 22:32:29 -04:00
Simon Cruanes	4098e88c68	CI	2025-05-07 21:52:25 -04:00
Simon Cruanes	c3bd2f92a8	fix bench	2025-05-07 20:41:48 -04:00
Simon Cruanes	d7f0aff406	cleaner tracing errors	2025-05-07 17:19:47 -04:00
Simon Cruanes	86e65d2046	test: update and improve fuchsia tests	2025-05-07 17:19:47 -04:00
Simon Cruanes	7acc1b930f	detail	2025-05-07 15:35:11 -04:00
Simon Cruanes	190f70d7c9	feat fuchsia: full revamp of the library, modularized ... - separate exporter, writer, subscriber - use the subscriber span tbl to keep track of context - use a `Buf_chain.t` to keep multiple buffers in use, and keep a set of ready buffers - batch write the ready buffers and then recycle them	2025-05-07 15:33:34 -04:00
Simon Cruanes	a4779227fa	add .mli for rpool	2025-05-07 13:15:37 -04:00
Simon Cruanes	81096e0d3c	refactor TEF: split into exporter,writer,subscriber ... code is a lot cleaner now.	2025-05-07 13:08:22 -04:00
Simon Cruanes	4454975a61	feat util: remove b_queue, add Rpool ... to be used in various buffer pools.	2025-05-07 13:06:55 -04:00
Simon Cruanes	005626a2cd	feat: add trace.event, useful for background threads ... send these events into a queue and process them somewhere else.	2025-05-07 11:10:15 -04:00
Simon Cruanes	76703461ea	feat(trace.subscriber): add Span_tbl, and a depopt on picos_aux	2025-05-07 11:09:42 -04:00
Simon Cruanes	7cc16bc0b8	wip: test for fuchsia	2025-05-05 15:08:57 -04:00
Simon Cruanes	7405e3ae1b	wip: port fuchsia to subscriber infra	2025-05-05 15:08:57 -04:00
Simon Cruanes	ef50b578f1	refactor(subscriber): timestamps are int64ns now	2025-05-05 15:08:57 -04:00
Simon Cruanes	384dca93e2	fix: better retrocompat for enter_manual_{toplevel,sub}_span	2025-05-02 09:19:35 -04:00
Simon Cruanes	3c1360677a	subscriber: tee a whole array at once	2025-05-02 08:55:45 -04:00