Sandro Santilli

Results 324 comments of Sandro Santilli

It is the straight skeleton computation (with inexact kernel) that's taking up all that memory, specifically this snippet: ``` CGAL::create_interior_straight_skeleton_2( poly.outer_boundary().vertices_begin(), poly.outer_boundary().vertices_end (), poly.holes_begin (), poly.holes_end (), CGAL::Epick() ); ```...

No news. Could be memory fragmentation, yes.

See #334 With `/olm info all` I get `No matching devices found`. I guess @poljar needs help with this

the instructions don't work for me, I always get "Devince xxx of user yyy not found"

The error was fixed for me by making the `matrix_sso_helper` command available (without the .py extension) in PATH. Ie: cp matrix_sso_helper.py /usr/local/bin/matrix_sso_helper; chmod +x /usr/local/bin/matrix_sso_helper Next issue is that matrix.org...

I'm having the same problem, but only with `matrix.org` homeserver, so this sounds like a limitation of matrix.org itself, which requires SSO ?

This is the ironic screen I'm shown when trying to follow the weechat-matrix instructions: ``` Choose identity provider Log in to matrix.org Choose an identity provider to log in *...

The matrix-weechat instructions were: ``` 16:07:51 -- | matrix: The server requested a single sign-on, please open this URL in your browser. Note | that the browser needs to run...

Info from #matrix IRC channel on libera.chat: ``` 6:11:35 ilmari | strk: https://matrix-client.matrix.org/_matrix/client/v3/login returns sso, | token, password and appservice flows 16:12:08 ilmari | in that order. maybe weechat only...

The code seems to ASSUME server requires SSO if it gets a redirectURL typed response