{"thread":{"id":"11762","subject":"Is there a reason to keep walker.c ?","startedAt":"2008-01-27T20:45:35Z","lastAt":"2008-01-28T19:45:00Z","messageCount":8,"participants":["Mike Hommey","Johannes Schindelin","Jakub Narebski","Daniel Barkalow"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"66686","messageId":"20080127204535.GA4702@glandium.org","threadId":"11762","inReplyTo":null,"subject":"Is there a reason to keep walker.c ?","fromName":"Mike Hommey","fromEmail":"mh@glandium.org","sentAt":"2008-01-27T20:45:35Z","receivedAt":"2008-01-27T20:45:35Z","isPatch":false,"sender":{"key":"mh@glandium.org","avatar":"https://avatars.githubusercontent.com/u/1038527?v=4"},"body":"Hi,\n\nWhile working on the http code refactoring, I got to wonder if the\nwalker.c \"wrapper\", that is only used for the http transport, is still\nworth keeping. If there are plans for others transport to use this code,\nobviously, it would be worth keeping, but on the contrary, I think it\nwould simplify the http transport code even more. What do you think ?\n\nCheers,\n\nMike\n"},{"id":"66688","messageId":"alpine.LSU.1.00.0801272045490.23907@racer.site","threadId":"11762","inReplyTo":"20080127204535.GA4702@glandium.org","subject":"Re: Is there a reason to keep walker.c ?","fromName":"Johannes Schindelin","fromEmail":"johannes.schindelin@gmx.de","sentAt":"2008-01-27T20:46:59Z","receivedAt":"2008-01-27T20:46:59Z","isPatch":false,"sender":{"key":"johannes.schindelin@gmx.de","avatar":"https://avatars.githubusercontent.com/u/127790?v=4"},"body":"Hi,\n\nOn Sun, 27 Jan 2008, Mike Hommey wrote:\n\n> While working on the http code refactoring, I got to wonder if the \n> walker.c \"wrapper\", that is only used for the http transport, is still \n> worth keeping. If there are plans for others transport to use this code, \n> obviously, it would be worth keeping, but on the contrary, I think it \n> would simplify the http transport code even more. What do you think ?\n\nReally, I was waiting for somebody needing ftp and/or sftp support badly \nenough, so let's keep it.\n\nI mean, one of those guys asking for ftp push support _got_ to just start \nscratching that itch, right?\n\nCiao,\nDscho\n"},{"id":"66689","messageId":"20080127205155.GA5476@glandium.org","threadId":"11762","inReplyTo":"alpine.LSU.1.00.0801272045490.23907@racer.site","subject":"Re: Is there a reason to keep walker.c ?","fromName":"Mike Hommey","fromEmail":"mh@glandium.org","sentAt":"2008-01-27T20:51:55Z","receivedAt":"2008-01-27T20:51:55Z","isPatch":false,"sender":{"key":"mh@glandium.org","avatar":"https://avatars.githubusercontent.com/u/1038527?v=4"},"body":"On Sun, Jan 27, 2008 at 08:46:59PM +0000, Johannes Schindelin wrote:\n> Hi,\n> \n> On Sun, 27 Jan 2008, Mike Hommey wrote:\n> \n> > While working on the http code refactoring, I got to wonder if the \n> > walker.c \"wrapper\", that is only used for the http transport, is still \n> > worth keeping. If there are plans for others transport to use this code, \n> > obviously, it would be worth keeping, but on the contrary, I think it \n> > would simplify the http transport code even more. What do you think ?\n> \n> Really, I was waiting for somebody needing ftp and/or sftp support badly \n> enough, so let's keep it.\n> \n> I mean, one of those guys asking for ftp push support _got_ to just start \n> scratching that itch, right?\n\nThough, technically, ftp push could work with the curl code.\n\nMike\n"},{"id":"66691","messageId":"m37ihv9d52.fsf@localhost.localdomain","threadId":"11762","inReplyTo":"20080127205155.GA5476@glandium.org","subject":"Re: Is there a reason to keep walker.c ?","fromName":"Jakub Narebski","fromEmail":"jnareb@gmail.com","sentAt":"2008-01-27T21:23:13Z","receivedAt":"2008-01-27T21:23:13Z","isPatch":false,"sender":{"key":"jnareb@gmail.com","avatar":"https://avatars.githubusercontent.com/u/2706?v=4"},"body":"Mike Hommey <mh@glandium.org> writes:\n\n> On Sun, Jan 27, 2008 at 08:46:59PM +0000, Johannes Schindelin wrote:\n>> On Sun, 27 Jan 2008, Mike Hommey wrote:\n>> \n>>> While working on the http code refactoring, I got to wonder if the \n>>> walker.c \"wrapper\", that is only used for the http transport, is still \n>>> worth keeping. If there are plans for others transport to use this code, \n>>> obviously, it would be worth keeping, but on the contrary, I think it \n>>> would simplify the http transport code even more. What do you think ?\n>> \n>> Really, I was waiting for somebody needing ftp and/or sftp support badly \n>> enough, so let's keep it.\n>> \n>> I mean, one of those guys asking for ftp push support _got_ to just start \n>> scratching that itch, right?\n> \n> Though, technically, ftp push could work with the curl code.\n\nIIRC git fetch works with FTP transport.  Somebody would have to write\nreplacement for WebDAV authentication for ftp / sftp / ftps to have\nproper ftp push support.  There were request, but AFAIR no code.\n\nAre you thinking about POP / IMAP transport, or XMPP one ;-PPP ?\n-- \nJakub Narebski\nPoland\nShadeHawk on #git\n"},{"id":"66692","messageId":"alpine.LNX.1.00.0801271603280.13593@iabervon.org","threadId":"11762","inReplyTo":"20080127204535.GA4702@glandium.org","subject":"Re: Is there a reason to keep walker.c ?","fromName":"Daniel Barkalow","fromEmail":"barkalow@iabervon.org","sentAt":"2008-01-27T21:23:17Z","receivedAt":"2008-01-27T21:23:17Z","isPatch":false,"sender":{"key":"barkalow@iabervon.org","avatar":"https://avatars.githubusercontent.com/u/55364219?v=4"},"body":"On Sun, 27 Jan 2008, Mike Hommey wrote:\n\n> Hi,\n> \n> While working on the http code refactoring, I got to wonder if the\n> walker.c \"wrapper\", that is only used for the http transport, is still\n> worth keeping. If there are plans for others transport to use this code,\n> obviously, it would be worth keeping, but on the contrary, I think it\n> would simplify the http transport code even more. What do you think ?\n\nIt would be a good base for sftp (i.e. dumb file access over ssh). In \nfact, I think stuff should ideally be moved into walker.c such that the \nHTTP-specific code just handles access to files by filename and the logic \nof what files to request in what order is in walker.c. I think this would \nget the simplification you're looking for while making it easy to add sftp \nor any other situation where you have only slow remote filesystem-like \naccess to the repository.\n\n\t-Daniel\n*This .sig left intentionally blank*\n"},{"id":"66713","messageId":"20080128071749.GA12488@glandium.org","threadId":"11762","inReplyTo":"alpine.LNX.1.00.0801271603280.13593@iabervon.org","subject":"Re: Is there a reason to keep walker.c ?","fromName":"Mike Hommey","fromEmail":"mh@glandium.org","sentAt":"2008-01-28T07:17:49Z","receivedAt":"2008-01-28T07:17:49Z","isPatch":false,"sender":{"key":"mh@glandium.org","avatar":"https://avatars.githubusercontent.com/u/1038527?v=4"},"body":"On Sun, Jan 27, 2008 at 04:23:17PM -0500, Daniel Barkalow wrote:\n> On Sun, 27 Jan 2008, Mike Hommey wrote:\n> \n> > Hi,\n> > \n> > While working on the http code refactoring, I got to wonder if the\n> > walker.c \"wrapper\", that is only used for the http transport, is still\n> > worth keeping. If there are plans for others transport to use this code,\n> > obviously, it would be worth keeping, but on the contrary, I think it\n> > would simplify the http transport code even more. What do you think ?\n> \n> It would be a good base for sftp (i.e. dumb file access over ssh). In \n> fact, I think stuff should ideally be moved into walker.c such that the \n> HTTP-specific code just handles access to files by filename and the logic \n> of what files to request in what order is in walker.c. I think this would \n> get the simplification you're looking for while making it easy to add sftp \n> or any other situation where you have only slow remote filesystem-like \n> access to the repository.\n\nI like this idea. I'll probably implement that, then.\n\nMike\n"},{"id":"66755","messageId":"20080128185828.GA14804@glandium.org","threadId":"11762","inReplyTo":"20080128071749.GA12488@glandium.org","subject":"http-push as a builtin ? (Was: Is there a reason to keep walker.c ?)","fromName":"Mike Hommey","fromEmail":"mh@glandium.org","sentAt":"2008-01-28T18:58:28Z","receivedAt":"2008-01-28T18:58:28Z","isPatch":false,"sender":{"key":"mh@glandium.org","avatar":"https://avatars.githubusercontent.com/u/1038527?v=4"},"body":"On Mon, Jan 28, 2008 at 08:17:49AM +0100, Mike Hommey wrote:\n> On Sun, Jan 27, 2008 at 04:23:17PM -0500, Daniel Barkalow wrote:\n> > On Sun, 27 Jan 2008, Mike Hommey wrote:\n> > \n> > > Hi,\n> > > \n> > > While working on the http code refactoring, I got to wonder if the\n> > > walker.c \"wrapper\", that is only used for the http transport, is still\n> > > worth keeping. If there are plans for others transport to use this code,\n> > > obviously, it would be worth keeping, but on the contrary, I think it\n> > > would simplify the http transport code even more. What do you think ?\n> > \n> > It would be a good base for sftp (i.e. dumb file access over ssh). In \n> > fact, I think stuff should ideally be moved into walker.c such that the \n> > HTTP-specific code just handles access to files by filename and the logic \n> > of what files to request in what order is in walker.c. I think this would \n> > get the simplification you're looking for while making it easy to add sftp \n> > or any other situation where you have only slow remote filesystem-like \n> > access to the repository.\n> \n> I like this idea. I'll probably implement that, then.\n\nBTW, would there be objections to have http-push as a builtin ?\n\nMike\n"},{"id":"66756","messageId":"alpine.LNX.1.00.0801281438450.13593@iabervon.org","threadId":"11762","inReplyTo":"20080128185828.GA14804@glandium.org","subject":"Re: http-push as a builtin ? (Was: Is there a reason to keep walker.c ?)","fromName":"Daniel Barkalow","fromEmail":"barkalow@iabervon.org","sentAt":"2008-01-28T19:45:00Z","receivedAt":"2008-01-28T19:45:00Z","isPatch":false,"sender":{"key":"barkalow@iabervon.org","avatar":"https://avatars.githubusercontent.com/u/55364219?v=4"},"body":"On Mon, 28 Jan 2008, Mike Hommey wrote:\n\n> On Mon, Jan 28, 2008 at 08:17:49AM +0100, Mike Hommey wrote:\n> > On Sun, Jan 27, 2008 at 04:23:17PM -0500, Daniel Barkalow wrote:\n> > > On Sun, 27 Jan 2008, Mike Hommey wrote:\n> > > \n> > > > Hi,\n> > > > \n> > > > While working on the http code refactoring, I got to wonder if the\n> > > > walker.c \"wrapper\", that is only used for the http transport, is still\n> > > > worth keeping. If there are plans for others transport to use this code,\n> > > > obviously, it would be worth keeping, but on the contrary, I think it\n> > > > would simplify the http transport code even more. What do you think ?\n> > > \n> > > It would be a good base for sftp (i.e. dumb file access over ssh). In \n> > > fact, I think stuff should ideally be moved into walker.c such that the \n> > > HTTP-specific code just handles access to files by filename and the logic \n> > > of what files to request in what order is in walker.c. I think this would \n> > > get the simplification you're looking for while making it easy to add sftp \n> > > or any other situation where you have only slow remote filesystem-like \n> > > access to the repository.\n> > \n> > I like this idea. I'll probably implement that, then.\n> \n> BTW, would there be objections to have http-push as a builtin ?\n\nNot from me. Actually, it would be ideal to call its functions directly \nfrom transport.c and deprecate the separate command. (And possibly \nseparate the control structure from the HTTP code and move the former into \nwalker.c where it could be used by sftp)\n\n\t-Daniel\n*This .sig left intentionally blank*\n"}]}