gmnisrv/src/server.c, branch mainFork of gmnisrv (Gemini server): full CA certificate chain (Lets Encrypt), TLS 1.3 only, handshake buffer-overflow fix
https://git.virebent.art/virebent/gmnisrv/atom?h=main2026-06-26T17:20:00+00:00Serve CA full chain (Let's Encrypt), TLS 1.3, fix handshake buffer overflow2026-06-26T17:20:00+00:00Gab Virebentgab@virebent.art2026-06-26T17:20:00+00:00urn:sha1:fb48d1308d2f63f8e9b23c5d1d921783fa0dacbe
- tls.c/config.h: read full PEM chain into STACK_OF(X509), send via SSL_set1_chain
- tls.c: minimum protocol TLS 1.2 -> TLS 1.3
- server.c: drain handshake output via local buffer loop instead of staging
into the fixed 4 KB client buffer (a full CA chain overflowed it -> assert)
- FORK.md: describe the fork
Close some files when too many are open2022-12-19T11:02:08+00:00Remy Noulinloader2x@gmail.com2022-11-22T07:03:22+00:00urn:sha1:95138564acbc347c0a5f4f86d9280e3ab1484fd2notify clients about malformed requests2022-08-15T13:50:27+00:00Julian Maingotjulianmaingot@gmail.com2022-07-31T00:24:36+00:00urn:sha1:132f2ececc012cf4744ca5b8044fb637fed96a0c
I think this has been broken for a while and I came across it while
implementing my own client. The server should notify the client that the
request is malformed but wasn't sending the response when SSL_read
returned SSL_ERROR_WANT_READ.
disconnect_client: fix index calculation2021-05-16T19:54:16+00:00Eyal Sawadyecs@d2evs.net2021-05-16T19:53:47+00:00urn:sha1:b9a92193e96bbe621ebc8430d8308d45a5b86cef
This was sometimes causing a gmnisrv_client to be overwritten by the
deinitialized client, which occasionally lead to a segfault when we
later tried to use it.
Fix use of wrong index during pollfd refresh2020-11-10T13:06:37+00:00Kenny Levinsenkl@kl.wtf2020-11-10T13:05:43+00:00urn:sha1:9fd9eef9d7f850ccc1f8d5fe4dc60df461a5dbc2
The right side of the refresh ended up using the index of the deleted
client instead of the index from the loop, which happens to work when
the destroyed client is either last or second-last.
Update pollfd pointers when destroying a client2020-11-09T02:34:34+00:00Kenny Levinsenkl@kl.wtf2020-11-08T22:59:41+00:00urn:sha1:fdd27bb4c1f1fb05481368c214e0f0f30a81b0ca
The client pollfd pointer would go stale when the server pollfd array
was moved to compensate for a destroyed client, which in turn led to
poll breakage.
Refresh the pollfd pointers after memmove.
Prevent double free of client resources2020-11-08T21:37:53+00:00Drew DeVaultsir@cmpwn.com2020-11-08T21:37:53+00:00urn:sha1:74077b6f951865420a78345d5e793bba08fc0edcSet POLLOUT only when there's something to write2020-11-08T14:38:17+00:00mbaysmbays@sdf.org2020-11-08T13:30:45+00:00urn:sha1:4b2c3afdb0f448372e1ffa58a88250a305ea85cbDon't memmove one past clients2020-11-08T14:37:59+00:00Tom Lebreuxme@tomlebreux.com2020-11-07T12:56:48+00:00urn:sha1:7aedbed7740c3ee5d60705a2ad74cdc760e60f2eFix not moving all fds2020-11-08T14:37:55+00:00Tom Lebreuxme@tomlebreux.com2020-11-07T05:12:00+00:00urn:sha1:d20caf389bd6cd8e6d8daaeb6d51a022fc643292