For your insensitive search issue, two suggestions:
Try variable completion-ignore-case
Raise GitHub issue on defcustom org-roam-completion-ignore-case
Your issue indicates it’s not working on v2 now. It’s either still to be implemented or to be deprecated.
For your second issue of org-roam-buffer, it looks to me that you have all the elements to be displayed; however, they are displayed in the wrong order. It looks to me the same problem as this one on Slack
Suggestions:
Ask Milan what he might have done to solve it; he didn’t come back to Slack
Add more description how others might reproduce it on their end
I don’t have the same issue, so have no idea. You would really need to provide a step-by-step description and your environment (e.g. Doom, Spacemacs, what commit of v2 you are on, etc.).
Not sure what you are expecting by sharing your config.
For the insensitive issue, try to see if my suggestions work. For the org-roam-buffer order issue, it looks like the cause is in the other parts of your config. I would turn off everything else and see if I could locate it.
Other than these suggestions, I don’t think I can do much else. I’m on vacation with no laptop.
Thank you for your help. I found that it was a conflict with the emacs-dashboard package. For now I disabled the dashboard but I wonder if anyone else has the same conflict.
I don’t know if my desire is the same as the OP’s, but, I would like org-roam-node-find’s search over nodes to be case-insensitive. Right now, I have a node called “Attention Is All You Need”, and when I just tried to find it by typing “Attention is”, I got no matches, no matter the value of completion-ignore-case.
I guess what I want is something stronger than what is controlled by that variable – I want the case of what I type to also be ignored.
You might respond, then why don’t you always just type in lower case? Which is a reasonable response, but, I also have this use case in mind: when I copy and paste text from somewhere, and there are the titles of research papers in the text, I would like to be able to select the text of the paper’s title and call org-roam-node-find on it, to see whether I already have a node for that paper or not, and that paper’s title’s cases may or may not match the case of the target node.
Two possibilities. Your completion framework and completion-styles.
The latter is less likely because before I changed completion-ignore-case to t it was case sensitive.
For completion frameowork, I use Vertico, so it’s very close to bare-bone completion.
For completion-styles, Orderless does most of the work; it is set to the value as follows: (basic orderless partial-completion).
The value of completion-styles in my Emacs is (basic partial-completion emacs22 helm).
So I guess the next thing for me to do will be to investigate how those completions work, and see if there’s another completion that I would like to use over the ones in the existing list. Thanks!
Is this behavior the result of helm-case-fold-search being set to smart?
pattern “foo” match “foo” and “Foo”
pattern “Foo” match only "Foo
It’s a 2012 exchange but I read it as helm’s general behavior even when you are searching within minibuffer. I don’t use it so I can’t investigate any further…