Re: [RFC] New ports category "fs"
- Reply: Olivier Certner : "Re: [RFC] New ports category "fs""
- In reply to: Olivier Certner : "Re: [RFC] New ports category "fs""
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Wed, 06 Nov 2024 16:14:30 UTC
Hi Olivier, Am Wed, Nov 06, 2024 at 05:09:58PM +0100 schrieb Olivier Certner: > Hi Robert, > > I have a few remarks/questions. > > The move implemented two different operations at once: > 1. Grouping sysutils/fusefs-*, and removing the 'fusefs-' port prefix. > 2. Grouping all filesystems-related ports. > > Old sysutils/fusefs-* ports now do not have the 'fusefs-' prefix, but > the packages produced from them still have. Isn't that slightly confusing? > > That's maybe a non-existent case, but how would we handle a native > driver vs. a FUSE implementation in ports, now that the latter are not > prefixed? The main purpose of the new category is to get rid of the sysutils/fusefs- pseudo category. The prefix stays for installed packages so the PKGNAME doesn't change and binary package users do not get a surprise when they try to update. If for some reason there are multiple packages competing for the same name within the filesystems category, a disambiguating prefix may be added as usual. > Is there a plan to tag the old sysutils/fusefs-* ports with, e.g., > some 'fusefs' virtual category? That could come in handy when listing > which FUSE filesystems are available (in particular with 'make search > cat=fusefs'), even if there are other, though less user-friendly, ways > to obtain this list. Right now you can find these easily using "git grep '^USES=.*fuse', though in principle a fusefs (or just fuse) virtual category might be nice. Feel free to propose such a category. I wanted to keep it with just one new category as to raise the chances of this patch passing review (still took over a month). Yours, Robert Clausecker -- () ascii ribbon campaign - for an encoding-agnostic world /\ - against html email - against proprietary attachments