- Aug 30, 2021
-
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
work when launched manually Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
new canales.m3u Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
EPG_THREADS to control the number of threads downloading the epg, by default at 3, 8 seems to give trouble with a standard Movistar fiber connection/ONT/router Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
for the first time. Stop loading the metadata cache when it is always updated along the epg. We only want it for the epg backend to use. Gracias @jesusvpct por abrir el issue y seguir insistiendo hasta que te funcionase :) Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
- Aug 29, 2021
-
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
since it is essential for the entire proxy to work, make it more resillient by trying to update the epg until it is successful instead of just dying. Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
- Aug 25, 2021
-
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
- Aug 22, 2021
-
-
Javier Marcet authored
188 and 1316 Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
cleanup imports and apply autopep8. Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
drop walrus operator user, it hinders readability a lot. Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
- add -xerror global option so recordings with the minimal error are rejected. THis works so well that programs with error's in movistar's servers are rejected up front ad infinitum, I should add some kind of retry limit for the same pid - drop the redundant --default-language - raise ffmpeg verbose lovel to info Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
AttributeError when sending eof Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
once the recordings are done. The reason this happened is movistar_u7d needs to launch said processes as background processes or it will be forced to keep the record api calls open for the duration of the recordings. By default, posix mandates parent processes must take care of their children, i.e., they must wait() for them. When a child ends, the kernel sends a SIG_CHLD signal so the parent knows and can wait() for it. But there exists also the option to tell the kernel up front we are careless parents and to not expect us to take care of our children which is what we want movistar_u7d to do. vod.py is its only child, even if repeated, it's always the same child and movistar_u7d does not care at all about it when it is asked to record something, vod.py can take care of itself on its own. Hence, movistar_u7d.py now ignores the SIG_CHLD signal each time a vod recording process is launched. Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
This reverts commit d1a2a77f.
-
Javier Marcet authored
that they survive an autoreload without keeping the server port busy Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
methods, and ensure no future is left behind on autoreload Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
- Aug 17, 2021
-
-
Javier Marcet authored
by calling get_vod_url() from movistar_u7d, we already return a 404 when needed and cleanup all the vod processes, so leave the entire vod logic within vod. Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
good the stale vod.py process issue, also get rid of all pkills, and show faster a 404 when it happens. After this commit the proxy is able to cope with a wrk test and at the same time serve vod clients with basically no added delay at all. All in all kudos to the Sanic team and thanks for the help on discord :) Signed-off-by: Javier Marcet <javier@marcet.info>
-
Javier Marcet authored
be solved Signed-off-by: Javier Marcet <javier@marcet.info>
-