Re: commit failed due to "backwards-sliding source views" - what do I do?
On Tue, Apr 28, 2020 at 8:16 AM Chaya wrote: > > Has anyone found solution to this problem? > I am getting this error while running git svn fetch command. > Tried many things to solve this issue, but no luck. > If anyone has found solution , it would be very helpful if you post it here. > Thanks! > > -- > Sent from: > http://subversion.1072662.n5.nabble.com/Subversion-Users-f84423.html I did not find much about this error. Just this SO question, without a real answer: https://stackoverflow.com/questions/24904329/git-svn-clone-failed-due-to-svndiff-backward-sliding-source-view Seems to me the problem is specific to git-svn, so I'm thinking you should contact people in the git-svn community. They might know more or have experience with this kind of problem. -- Johan
Re: commit failed due to "backwards-sliding source views" - what do I do?
On Tue, Apr 28, 2020 at 2:16 AM Chaya wrote: > > Has anyone found solution to this problem? > I am getting this error while running git svn fetch command. > Tried many things to solve this issue, but no luck. > If anyone has found solution , it would be very helpful if you post it here. > Thanks! A search through the mailing list archive at https://svn.haxx.se/users/ (the lists.apache.org archive doesn't seem to go farther back than 2009?) shows some hits for backwards-sliding windows and such, but all the ones I've found are very old posts from 2006 and 2008. Perhaps the "best" one I found was this report from 2006 that the OP later reported was caused by some antivirus software: https://svn.haxx.se/users/archive-2006-08/0398.shtml My best guesses would be to make sure you're using recent versions of svn and git, check whether this error is due to any history-rewriting gymnastics on the git side of things, and if all else fails, look into whether antivirus or some other software is playing a role here. Nathan
Lists history (was: Re: commit failed due to "backwards-sliding source views" - what do I do?)
Nathan Hartman wrote on Tue, 28 Apr 2020 10:09 -0400: > On Tue, Apr 28, 2020 at 2:16 AM Chaya wrote: > > > > Has anyone found solution to this problem? > > I am getting this error while running git svn fetch command. > > Tried many things to solve this issue, but no luck. > > If anyone has found solution , it would be very helpful if you post it here. > > Thanks! > > A search through the mailing list archive at > https://svn.haxx.se/users/ (the lists.apache.org archive doesn't seem > to go farther back than 2009?) Subversion moved to Apache in 2009. Neither lists.a.o nor mail-archives.a.o would be expected to go farther back than that. Before the migration, the lists were @subversion.tigris.org (except private@, which was svn-full-committ...@red-bean.com), and instead of being part of ASF we had our own Subversion Corporation (see https://svn.haxx.se/org/, and the archive.org snapshots of subversion.org). The tigris lists used to be independently archived in several places, but at least two of them (tigris and gmane) have closed since then. It wouldn't be a bad idea to download haxx.se's archives to Apache hardware so we have an independent backup. In fact, Infra might even be willing to inject the pre-migration data into the mail-archives.a.o/lists.a.o archives, so we have unified archives. [When we migrated legal@ allowed us to add the historical release tarballs to archive.a.o/dist/ even though they contained LGPL dependency tarballs (which aren't normally allowed in Apache releases). In light of this, I can't imagine backfilling the lists would be a problem.] The first step, though, would be to get the raw mbox's from haxx.se. The owner of that domain is a full committer, so it should be easy… Cheers, Daniel
Re: commit failed due to "backwards-sliding source views" - what do I do?
Johan Corveleyn wrote on Tue, 28 Apr 2020 14:06 +0200: > On Tue, Apr 28, 2020 at 8:16 AM Chaya wrote: > > > > Has anyone found solution to this problem? > > I am getting this error while running git svn fetch command. Is it a public repository, so we can try to reproduce the error ourselves? What's the URL scheme (http(s) or svn(+ssh))? Version numbers of svn on the server and client and git on the client? > > Tried many things to solve this issue, but no luck. > > If anyone has found solution , it would be very helpful if you post it here. > > Thanks! > > I did not find much about this error. Just this SO question, without a > real answer: > https://stackoverflow.com/questions/24904329/git-svn-clone-failed-due-to-svndiff-backward-sliding-source-view > > Seems to me the problem is specific to git-svn, so I'm thinking you > should contact people in the git-svn community. They might know more > or have experience with this kind of problem. If the git-svn maintainers have questions or bug reports for us, we'd be happy to engage with them. Cheers, Daniel
Re: commit failed due to "backwards-sliding source views" - what do I do?
Hi, I am using the latest versions of svn and git. And what do you mean by "history-rewriting gymnastics" ? Could you please explain? -- Sent from: http://subversion.1072662.n5.nabble.com/Subversion-Users-f84423.html
Re: commit failed due to "backwards-sliding source views" - what do I do?
1. Unfortunately, it's not a public repository. 2. The SVN project repo is actually file based. So I had to convert it to svn protocol before cloning to git. 3. Using latest version of svn (version 1.13.0 (r1867053)) and git ( version 1.13.0 (r1867053)) -- Sent from: http://subversion.1072662.n5.nabble.com/Subversion-Users-f84423.html
Re: commit failed due to "backwards-sliding source views" - what do I do?
On Wed, Apr 29, 2020 at 1:14 AM Chaya wrote: > Hi, > I am using the latest versions of svn and git. And what do you mean by > "history-rewriting gymnastics" ? Could you please explain? Hi, I don't use the git-svn bridge myself, but I have read about it (admittedly, quite some time ago). I remember there were suggestions to keep the git history linear so that it will match the way Subversion works, and recommendations of things not to do in git because they could prevent the bidirectional bridge from working properly. I don't remember the details; I wrote "history-rewriting gymnastics" in a hurry. Nathan
