Conversation
github-actions
bot
added
the
severity: ABI change
|
Unable to assist at the moment. Functions need to be exported before I can test. |
|
Not sure what you mean. Commits pushed a few days ago exported the SSL_Llistwn_ex function |
The OSSL_QUIC_method isn't exported. I need that to use the SSL_Listen_ex. |
|
yes it is, see commit c3ea941: |
|
Sorry for the late response - Easter holidays. |
|
Hmmm. I don't see the export. Am I on the right branch? Exports: |
|
@nhorman - what is the current status of this PR? |
|
@mattcaswell apologies, this fell off my radar. Need to look at it again, but my last recollection was that the test I wrote for it passed, but definitely needs some cleanup/further testing. It should be enough for @Splediferous to experiment with however. @Splediferous If I build it on linux, I see the symbol as being exported: |
|
I built this on windows, and the export is there: 
|
|
I'm still not seeing it. |
|
whats the head commit of the tree you're building this library from? |
|
I don't use Git. |
|
But...The changes we're testing aren't in the release. They're only in the draft PR I gave you. They need to be tested prior to putting them in a release, thats why you were provided with the PR. |
|
I think we've had a communications problem. I asked about it in #27250 (comment) and didn't get a response. When the final 3.6 was released I thought it had been overlooked and created a feature request. Where can I download the source to test? Can you link a repository where I can download a zip or tar package? Sorry for the confusion :( |
|
I'm not sure I see the problem with using git, despite what you use with your normal development. But if its just not feasible to use git to clone this pr, then you can use githubs direct url feature to get an archive of the source branch to build from: |
|
I guess you don't have an IT department then. The souce you linked does, indeed, have the export. I will test and let you know the outcome by 20/10/2025 at the latest. |
|
your it department wont let you install git? Thats......wow. I'm sorry to hear that. |
|
I have successfully created a client connection with a modified Listener example from the demo folder. Anyhoo. I've also successfully created both a client and server connection using the OSSL_QUIC_method(). I'll be integrating the new changes to use the default QUIC channel this week with a view to working with multiple channels at a later date. Is there an ETA for including OSSL_QUIC_method() in the main branch? |
|
@mattcaswell it sounds like @Splediferous has confirmed this draft pr is working for them. Obviously I need to rebase, and finish the docs on this. Do you have any early comments on what else may need to be included here before I take this out of draft for official review? |
mattcaswell
left a comment
mattcaswell
left a comment
There was a problem hiding this comment.
I've not reviewed this in detail, but the obvious missing thing here is documentation.
I'm also a bit confused by the test (see comment below)
|
This pull request is ready to merge |
|
merged, thank you! |
github-project-automation
bot
moved this from Waiting Merge
to Done
in Development Board
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from #27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from #27397)
The flag defaults to 0 (unknown), and gets set to 1 (using SSL_accept_ex), or -1 (using SSL_accpet_connection) Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from #27397)
Based on the value of the using_peeloff flag, we need to choose if we're going to create a user_ssl when creating a new channel Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from #27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from #27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from #27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from #27397)
Changes for SSL_listen_ex squashed creation of qlog objects even when configured. Fix that up so qlog objects are created regardless of weather we use SSL_accept_connection or SSL_listen_ex Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from #27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from #27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from #27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from #27397)
folow the using_peeloff get/set routines to just a set routine that fails if the set is for a mode that doesn't match the current mode Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from #27397)
Various review fixups to clarify meaning of variables and fix unwinding of operations should we encounter errors in some operations. Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from #27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from openssl#27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from openssl#27397)
The flag defaults to 0 (unknown), and gets set to 1 (using SSL_accept_ex), or -1 (using SSL_accpet_connection) Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from openssl#27397)
Based on the value of the using_peeloff flag, we need to choose if we're going to create a user_ssl when creating a new channel Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from openssl#27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from openssl#27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from openssl#27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from openssl#27397)
Changes for SSL_listen_ex squashed creation of qlog objects even when configured. Fix that up so qlog objects are created regardless of weather we use SSL_accept_connection or SSL_listen_ex Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from openssl#27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from openssl#27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from openssl#27397)
Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from openssl#27397)
folow the using_peeloff get/set routines to just a set routine that fails if the set is for a mode that doesn't match the current mode Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from openssl#27397)
Various review fixups to clarify meaning of variables and fix unwinding of operations should we encounter errors in some operations. Reviewed-by: Saša Nedvědický <[email protected]> Reviewed-by: Matt Caswell <[email protected]> (Merged from openssl#27397)
6 participants
Rough draft of SSL_listen_ex api
Allows for the creation of a 'blank' SSL object using OSSL_QUIC_method(), which can be passed to SSL_listen_ex to allow for quic to poll for inbound connections in a fashion simmilar to DTLS.
Very rough draft, still needs documentation, but it has a rudimentary test in place, which seems to work
Checklist