Re: [RFC PATCH] lib: document new database_open API

2020-07-08 Thread Floris Bruynooghe
On Sat 04 Jul 2020 at 14:01 -0300, David Bremner wrote: > Floris Bruynooghe writes: > >>> + * >>> + * - in the environment variable NOTMUCH_DATABASE, if non-empty >>> + * >>> + * - by $XDG_DATA_HOME/notmuch/$NOTMUCH_PROFILE where XDG_DATA_HOME >>> + * defaults to "$HOME/.local/share" and

Re: [RFC PATCH] lib: document new database_open API

2020-07-04 Thread David Bremner
Floris Bruynooghe writes: >> + * >> + * - in the environment variable NOTMUCH_DATABASE, if non-empty >> + * >> + * - by $XDG_DATA_HOME/notmuch/$NOTMUCH_PROFILE where XDG_DATA_HOME >> + * defaults to "$HOME/.local/share" and NOTMUCH_PROFILE defaults to >> + * "default" > > I like the profile

Re: [RFC PATCH] lib: document new database_open API

2020-07-04 Thread Floris Bruynooghe
On Fri 03 Jul 2020 at 10:43 -0300, David Bremner wrote: > Several aspects of this are potentially controversial: > > 1) The use of environment variables as fallback. I understand the > discomfort with having a library function check the environment, but > this seems to be functionality people

Re: [RFC PATCH] lib: document new database_open API

2020-07-03 Thread David Bremner
David Bremner writes: > There is not much code to back this so far. This is just me thinking > out loud at this point. The location calculation is done (and also > easy). The challenging part is probably updating > notmuch_database_get_config to do what this comment promises. There's always

[RFC PATCH] lib: document new database_open API

2020-07-03 Thread David Bremner
Several aspects of this are potentially controversial: 1) The use of environment variables as fallback. I understand the discomfort with having a library function check the environment, but this seems to be functionality people want, and it is better to implement it once. 2) The use of both NULL