Andrea Arcangeli

> method: POST_JSON > headers: > cache-control: "no-cache" > X-Gotify-Key: XXXXXXXXXXX The cache-control of the http POST HA sends to gotify server isn't going to alter the caching behavior of...

My workflow is about the opposite and my problem is that I'm not getting enough caching. To avoid wasting space, I give gotify max 1 min to fetch the image...

> You could include the current timestamp Depending on the use case for HA, filename_{{ this.context.id }}.jpg or {{ trigger.event.data.camera + '_' + trigger.event.context.id }}.jpg may be simpler than adding...

Just a minor update: it looks like the problem with coil is two parallel requests on the same url lead to data corruption despite the diskcache is serialized (by jvm...

I see in #337 you already predicted there could be a race condition in the disk cache. I wouldn't entirely rule out a bug in coil DiskLruCache, if nothing else...

> I see in #337 you already predicted there could be a race condition in the disk cache. I didn't notice yet there's two instances of the CoilHandler pointing to...

Same here, I also tried to pass priority as string and didn't work. Corollary problem is the click url needs to be null in json to be ignored by the...

I'm calling it from a my code, so I can restore the cached context even if I restart the editor. I didn't realize gptel-add-file always uses the full path so...

It's ok, it's no big deal to add the absolute file path is already used by gptel-add and cannot cause "interactive" dupicates. For the script, if there's a good reason...

Ok I replaced it with this, should reduce some overhead. ``` (defun gptel-context-move-to-tail () "Move the current buffer's file to the tail of the context." (when (and gptel-context (buffer-file-name)) (let...