Deomid Ryabkov
Deomid Ryabkov
> LGTM apologies, i had to amend my second commit - put const the wrong way up :)
added a test. however, while doing it i realized that actually, in practice the leak does not happen - the case of directory name is the only one that allocates...
> Generally speaking, we accept minor portability patches for platforms that we don't support officially, with the understanding that the contributor is responsible for testing. So this is acceptable in...
> That covers most of the use cases for the requirement here. Is this enough for you, or do you need different handshake sizes for the different connections? it still...
thanks for putting this together. this looks reasonable. having any part of this done would make be excellent!
yes, looks like a temporary variable was unused when building without debug. fixed. but unused variables aside, does the overall approach look good?
addressed comments, rebased
fwiw it covers all the changes i did in https://github.com/apache/mynewt-nimble/pull/1563
@macksal yep, in my library i buffer the full response and wait until NimBLE reads it out in MTU-1 byte sized chunks ([here](https://github.com/mongoose-os-libs/bt-common/blob/a8aae713f4fae16169f09c27e07f707a5f52a570/src/esp32xx/esp32xx_bt_gatts.c#L569)). this seems to be working so far...
that's fine, the problem is there is no way to do that currently. it's literally impossible to tell if the read being serviced is a part of long read or...