From mboxrd@z Thu Jan 1 00:00:00 1970 MIME-Version: 1.0 In-Reply-To: References: Date: Thu, 2 Feb 2012 04:05:57 -0800 Message-ID: From: Akshat Kumar To: Fans of the OS Plan 9 from Bell Labs <9fans@9fans.net> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Subject: Re: [9fans] [kenfs] the mercurial repo problem Topicbox-Message-UUID: 625fa40a-ead7-11e9-9d60-3106f5b1d025 Ah, sorry to double-post, but in order to get this working, one must: mkdir repo hg init repo cd repo; hg pull -u https://... echo '[paths] default =3D https://...' > .hg/hgrc Because it is forbidden to directly clone on a non-empty directory (an lnfs mounted dir contains a .longnames index file). ak On Thu, Feb 2, 2012 at 3:56 AM, Akshat Kumar wrote: > This is a problem perhaps two of us > have on Plan 9 (hi Erik): mercurial > repositories contain files with long > names and/or spaces. > > I realised that `lnfs' is a neat solution > to this problem (though you have to > remember to run it each time you're > in the repo) for the time-being. In > particular, it doesn't hog up ram (the > way ramfs does... :), and it doesn't > rely on the repo .hg dir already being > there (the way hgfs does), which is > what usually contains the long names. > And we can use `unlnfs' to export. > > Just for future reference -- I noticed that > this didn't come up when we last discussed > this issue. > > As always - > > BUGS > =A0 =A0 =A0 =A0 =A0This exists only to shame us into getting a real long = name > =A0 =A0 =A0 =A0 =A0file server working. > (Yeah, yeah - fossilheads can SUCK EGGS :-) > > > Best, > ak