Re: git stopped working with 1.7.1
that's really a patch :) but no choice.. I create a post on java.net Hope to get more feedback, to get this problem solve soon. Unix users are making fun of us :) Tomás Staig wrote: survivant wrote: Does anyone one have an answer for this problem ? (I saw lot of threads in different forums.. but never an answer.) I'm trying to clone a repository from my computer to my computer. I,m using cygwin 1.7. I,m able to clone all my others projects without problems. This project is 900k.. and the others are better than that. I have no idea what going on. I deleted the project and repush it.. always the same problems. $ GIT_TRACE=1 git clone g...@localhost:SmallProject.git trace: built-in: git 'clone' 'g...@localhost:SmallProject.git' Cloning into SmallProject... trace: run_command: 'ssh' 'g...@localhost' 'git-upload-pack '\''SmallProject.git'\''' DEBUG:gitosis.serve.main:Got command git-upload-pack 'SmallProject.git' DEBUG:gitosis.access.haveAccess:Access check for 'jerabi' as 'writable' on 'SmallProject.git'... DEBUG:gitosis.access.haveAccess:Stripping .git suffix from 'SmallProject.git', new value 'SmallProject' DEBUG:gitosis.group.getMembership:found 'jerabi' in 'gitosis-admin' DEBUG:gitosis.group.getMembership:found 'jerabi' in 'dev_jerabi' DEBUG:gitosis.access.haveAccess:Access ok for 'jerabi' as 'writable' on 'SmallProject' DEBUG:gitosis.access.haveAccess:Using prefix 'repositories' for 'SmallProject' DEBUG:gitosis.serve.main:Serving git-upload-pack 'repositories/SmallProject.git' trace: run_command: 'index-pack' '--stdin' '-v' '--fix-thin' '--keep=fetch-pack 6900 on jerabi-THINK' trace: exec: 'git' 'index-pack' '--stdin' '-v' '--fix-thin' '--keep=fetch-pack 6900 on jerabi-THINK' remote: Counting objects: 344, done. trace: built-in: git 'index-pack' '--stdin' '-v' '--fix-thin' '--keep=fetch-pack 6900 on jerabi-THINK' remote: Compressing objects: 58% (134/230) remote: Compressing objects: 100% (230/230), done. fatal: The remote end hung up unexpectedly fatal: early EOF fatal: index-pack failed jer...@jerabi-think ~/workspaces/test $ $ git --version git version 1.7.2.3 Don't have a clue of why this happens (some people say it is because of openssh, that if you change it for putty's implementation it works well) but when that happens to me (while pulling, never tried with clone) I just try a couple of times until it succeeds: for (( i=0 ; i 100 ; i++ )); do git pull; done works all the time for me... try something similar but with git clone ... of course. And well, I'm working with a 4+GB repository, so I'm really thankful that the clone worked right away. I know it's not the best solution, but if it works... it's good enough as a temporal workaround. Hope it works for you. Cheers, Tomas. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple -- View this message in context: http://old.nabble.com/git-stopped-working-with-1.7.1-tp26905956p29853811.html Sent from the Cygwin list mailing list archive at Nabble.com. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
Eric Blake ebl...@redhat.com wrote: On 03/11/2010 04:03 PM, Charles Wilson wrote: If it is not being worked on... I agree with Brian. This is a serious impediment to using Cygwin 1.7. I would think it would be of top priority. Only people who experience the problem can diagnose, debug, and fix it. git works fine for me under cygwin-1.7.1 (Vista, NTFS), so... Has the cygwin git maintainer been able to verify/reproduce the error? If not, then there's no way he can help you. I have sometimes seen the problem, but it has never been a show-stopper for me (I fall back to http cloning), and, so far, it has not percolated high enough to the top of my TODO list to be worth the investment of my time on it. I notice it is still broken (as of last week when I updated to the latest versions of everything and tested it). Any further news on when will be fixed? As for it not being a show-stopper for you, I just want to point out that there are many of us out here that do not have the luxury of switching to HTTP cloning when this fails. So, for us, it's a show-stopper that requires us to use putty to use git at all. Thanks. Kevin -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On 9/27/2010 12:38 PM, Kevin Layer wrote: Eric Blakeebl...@redhat.com wrote: On 03/11/2010 04:03 PM, Charles Wilson wrote: If it is not being worked on... I agree with Brian. This is a serious impediment to using Cygwin 1.7. I would think it would be of top priority. Only people who experience the problem can diagnose, debug, and fix it. git works fine for me under cygwin-1.7.1 (Vista, NTFS), so... Has the cygwin git maintainer been able to verify/reproduce the error? If not, then there's no way he can help you. I have sometimes seen the problem, but it has never been a show-stopper for me (I fall back to http cloning), and, so far, it has not percolated high enough to the top of my TODO list to be worth the investment of my time on it. I notice it is still broken (as of last week when I updated to the latest versions of everything and tested it). Any further news on when will be fixed? As for it not being a show-stopper for you, I just want to point out that there are many of us out here that do not have the luxury of switching to HTTP cloning when this fails. So, for us, it's a show-stopper that requires us to use putty to use git at all. I came up with a way to reproduce this if a developer wanted to try and fix it. That post can be seen here: http://cygwin.com/ml/cygwin/2010-07/msg00506.html git/ssh on cygwin are pretty much broken at this point... -Bill -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On Mon, Sep 27, 2010 at 02:24:05PM -0400, Bill Hoffman wrote: On 9/27/2010 12:38 PM, Kevin Layer wrote: Eric Blakeebl...@redhat.com wrote: On 03/11/2010 04:03 PM, Charles Wilson wrote: If it is not being worked on... I agree with Brian. This is a serious impediment to using Cygwin 1.7. I would think it would be of top priority. Only people who experience the problem can diagnose, debug, and fix it. git works fine for me under cygwin-1.7.1 (Vista, NTFS), so... Has the cygwin git maintainer been able to verify/reproduce the error? If not, then there's no way he can help you. I have sometimes seen the problem, but it has never been a show-stopper for me (I fall back to http cloning), and, so far, it has not percolated high enough to the top of my TODO list to be worth the investment of my time on it. I notice it is still broken (as of last week when I updated to the latest versions of everything and tested it). Any further news on when will be fixed? As for it not being a show-stopper for you, I just want to point out that there are many of us out here that do not have the luxury of switching to HTTP cloning when this fails. So, for us, it's a show-stopper that requires us to use putty to use git at all. I came up with a way to reproduce this if a developer wanted to try and fix it. That post can be seen here: http://cygwin.com/ml/cygwin/2010-07/msg00506.html git/ssh on cygwin are pretty much broken at this point... And, when I tried this, it pointed to an actual problem in git rather than a problem in Cygwin so that limits what is meant by developer. Isn't git normally used for source control management by programmers? If this is such a bad problem why isn't one of those programmers looking into it and proposing a fix? That's one of the reasons for free software - you can fix the problems yourself. And, before someone inevitably opines that I'm just being mean, I have to point out that this really is how many open source projects work. People who are not maintainers actually do propose fixes. cgf -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On 9/27/2010 3:59 PM, Christopher Faylor wrote: And, when I tried this, it pointed to an actual problem in git rather than a problem in Cygwin so that limits what is meant by developer. Isn't git normally used for source control management by programmers? If this is such a bad problem why isn't one of those programmers looking into it and proposing a fix? That's one of the reasons for free software - you can fix the problems yourself. And, before someone inevitably opines that I'm just being mean, I have to point out that this really is how many open source projects work. People who are not maintainers actually do propose fixes. I agree. Also, I admit that I am fishing for a developer that might be interested in fixing this. I also wanted to make the connection between the threads so that if someone did want to take a look at this, they would know how to reproduce it. Christopher, I did have a question on the old thread that went un-answered. I was wondering exactly what process you used to determine that it was git that was having a stack issue. It would be helpful as a starting point for a developer to try and fix this issue. I would be interested in repeating your experiment myself if I had that information. Thanks. -Bill -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On Mon, Sep 27, 2010 at 05:56:04PM -0400, Bill Hoffman wrote: On 9/27/2010 3:59 PM, Christopher Faylor wrote: And, when I tried this, it pointed to an actual problem in git rather than a problem in Cygwin so that limits what is meant by developer. Isn't git normally used for source control management by programmers? If this is such a bad problem why isn't one of those programmers looking into it and proposing a fix? That's one of the reasons for free software - you can fix the problems yourself. And, before someone inevitably opines that I'm just being mean, I have to point out that this really is how many open source projects work. People who are not maintainers actually do propose fixes. I agree. Also, I admit that I am fishing for a developer that might be interested in fixing this. I also wanted to make the connection between the threads so that if someone did want to take a look at this, they would know how to reproduce it. Christopher, I did have a question on the old thread that went un-answered. I was wondering exactly what process you used to determine that it was git that was having a stack issue. It would be helpful as a starting point for a developer to try and fix this issue. I would be interested in repeating your experiment myself if I had that information. At this point, I don't remember how I came to the conclusion. If you think you need special insight from me to debug this problem then you're not approaching the problem correctly. There is no reason to wait weeks for me to tell you how to repeat my experiment when you could just start debugging the problem yourself. cgf -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On 9/27/2010 7:03 PM, Christopher Faylor wrote: Christopher, I did have a question on the old thread that went un-answered. I was wondering exactly what process you used to determine that it was git that was having a stack issue. It would be helpful as a starting point for a developer to try and fix this issue. I would be interested in repeating your experiment myself if I had that information. At this point, I don't remember how I came to the conclusion. OK, you forgot. Thanks anyway. -Bill -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
Does anyone one have an answer for this problem ? (I saw lot of threads in different forums.. but never an answer.) I'm trying to clone a repository from my computer to my computer. I,m using cygwin 1.7. I,m able to clone all my others projects without problems. This project is 900k.. and the others are better than that. I have no idea what going on. I deleted the project and repush it.. always the same problems. $ GIT_TRACE=1 git clone g...@localhost:SmallProject.git trace: built-in: git 'clone' 'g...@localhost:SmallProject.git' Cloning into SmallProject... trace: run_command: 'ssh' 'g...@localhost' 'git-upload-pack '\''SmallProject.git'\''' DEBUG:gitosis.serve.main:Got command git-upload-pack 'SmallProject.git' DEBUG:gitosis.access.haveAccess:Access check for 'jerabi' as 'writable' on 'SmallProject.git'... DEBUG:gitosis.access.haveAccess:Stripping .git suffix from 'SmallProject.git', new value 'SmallProject' DEBUG:gitosis.group.getMembership:found 'jerabi' in 'gitosis-admin' DEBUG:gitosis.group.getMembership:found 'jerabi' in 'dev_jerabi' DEBUG:gitosis.access.haveAccess:Access ok for 'jerabi' as 'writable' on 'SmallProject' DEBUG:gitosis.access.haveAccess:Using prefix 'repositories' for 'SmallProject' DEBUG:gitosis.serve.main:Serving git-upload-pack 'repositories/SmallProject.git' trace: run_command: 'index-pack' '--stdin' '-v' '--fix-thin' '--keep=fetch-pack 6900 on jerabi-THINK' trace: exec: 'git' 'index-pack' '--stdin' '-v' '--fix-thin' '--keep=fetch-pack 6900 on jerabi-THINK' remote: Counting objects: 344, done. trace: built-in: git 'index-pack' '--stdin' '-v' '--fix-thin' '--keep=fetch-pack 6900 on jerabi-THINK' remote: Compressing objects: 58% (134/230) remote: Compressing objects: 100% (230/230), done. fatal: The remote end hung up unexpectedly fatal: early EOF fatal: index-pack failed jer...@jerabi-think ~/workspaces/test $ $ git --version git version 1.7.2.3 -- View this message in context: http://old.nabble.com/git-stopped-working-with-1.7.1-tp26905956p29824574.html Sent from the Cygwin list mailing list archive at Nabble.com. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
survivant wrote: Does anyone one have an answer for this problem ? (I saw lot of threads in different forums.. but never an answer.) I'm trying to clone a repository from my computer to my computer. I,m using cygwin 1.7. I,m able to clone all my others projects without problems. This project is 900k.. and the others are better than that. I have no idea what going on. I deleted the project and repush it.. always the same problems. $ GIT_TRACE=1 git clone g...@localhost:SmallProject.git trace: built-in: git 'clone' 'g...@localhost:SmallProject.git' Cloning into SmallProject... trace: run_command: 'ssh' 'g...@localhost' 'git-upload-pack '\''SmallProject.git'\''' DEBUG:gitosis.serve.main:Got command git-upload-pack 'SmallProject.git' DEBUG:gitosis.access.haveAccess:Access check for 'jerabi' as 'writable' on 'SmallProject.git'... DEBUG:gitosis.access.haveAccess:Stripping .git suffix from 'SmallProject.git', new value 'SmallProject' DEBUG:gitosis.group.getMembership:found 'jerabi' in 'gitosis-admin' DEBUG:gitosis.group.getMembership:found 'jerabi' in 'dev_jerabi' DEBUG:gitosis.access.haveAccess:Access ok for 'jerabi' as 'writable' on 'SmallProject' DEBUG:gitosis.access.haveAccess:Using prefix 'repositories' for 'SmallProject' DEBUG:gitosis.serve.main:Serving git-upload-pack 'repositories/SmallProject.git' trace: run_command: 'index-pack' '--stdin' '-v' '--fix-thin' '--keep=fetch-pack 6900 on jerabi-THINK' trace: exec: 'git' 'index-pack' '--stdin' '-v' '--fix-thin' '--keep=fetch-pack 6900 on jerabi-THINK' remote: Counting objects: 344, done. trace: built-in: git 'index-pack' '--stdin' '-v' '--fix-thin' '--keep=fetch-pack 6900 on jerabi-THINK' remote: Compressing objects: 58% (134/230) remote: Compressing objects: 100% (230/230), done. fatal: The remote end hung up unexpectedly fatal: early EOF fatal: index-pack failed jer...@jerabi-think ~/workspaces/test $ $ git --version git version 1.7.2.3 Don't have a clue of why this happens (some people say it is because of openssh, that if you change it for putty's implementation it works well) but when that happens to me (while pulling, never tried with clone) I just try a couple of times until it succeeds: for (( i=0 ; i 100 ; i++ )); do git pull; done works all the time for me... try something similar but with git clone ... of course. And well, I'm working with a 4+GB repository, so I'm really thankful that the clone worked right away. I know it's not the best solution, but if it works... it's good enough as a temporal workaround. Hope it works for you. Cheers, Tomas. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
The problem still exists but it appears infrequently with small repos. But larger repositories constantly fail even with the environment variable below defined. I currently use the putty ssh client for git. Nevertheless I consider this a major issue since the ssh protocol is the most important use case for git and as far as I know the cygwin version is the recommended one under windows. Corinna Vinschen-2 wrote: On Jan 18 00:17, ol42 wrote: But obviously the environment variable has still an effect! Without binmode I get the following error $ git clone ssh://o...@simulacron/home/git/gen-dsp.git Initialized empty Git repository in /home/ol/tmp/gen-dsp/.git/ remote: Counting objects: 979, done. remote: Compressing objects: 100% (966/966), done. fatal: The remote end hung up unexpectedly fatal: early EOF fatal: index-pack failed If I have CYGWIN=binmode defined it obviously works --- $ git clone ssh://o...@simulacron/home/git/gen-dsp.git Initialized empty Git repository in /home/ol/tmp/gen-dsp/.git/ remote: Counting objects: 979, done. remote: Compressing objects: 100% (966/966), done. remote: Total 979 (delta 483), reused 0 (delta 0) Receiving objects: 100% (979/979), 950.64 KiB | 1269 KiB/s, done. Resolving deltas: 100% (483/483), done. Checking out files: 100% (196/196), done. Dunno why you see two different results, but the CYGWIN=binmode setting most certainly doesn't exist anymore in Cygwin 1.7. Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Project Co-Leader cygwin AT cygwin DOT com Red Hat -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple -- View this message in context: http://old.nabble.com/git-stopped-working-with-1.7.1-tp26905956p28606567.html Sent from the Cygwin list mailing list archive at Nabble.com. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
Hi Brian, I had the same issue, lets see if my cause is your cause too. After I've replaced(or removed!) %windir%\system32\cygz.dll with the newer one in Cygwin, git 1.7.0.4 worked well. Perhaps I've manually copied an older cygz.dll to there sometimes ago.I can't remember. Some sub-commands of git like git-upload-pack tries to load a wrong/conflicting version of dll from somewhere. Note that I have the registry value to allow loading DLL from the same directory preceding the path but %windir%\system32 is always the first in the PATH environment variable. HTH Nothize Kevin Layer wrote: Brian L. br...@blucz.com wrote: From this thread I gather that little or no progress has been made on addressing the root cause of this issue. It still seems to be broken with the latest 1.7 series packages. What is the proper method for getting this bug in front of the guys who maintain the openssh package? The cygwin website suggests that the mailing list is the proper channel, but it doesn't seem like anyone's seriously looking into this as an openssh issue. It's clearly affecting more than a few users and the workaround, while reliable, is quite lame. Can this please get a little bit more priority? and another month goes by... If there is work going on behind the scenes, it would be nice to hear from the developer(s) working on it. If it is not being worked on... I agree with Brian. This is a serious impediment to using Cygwin 1.7. I would think it would be of top priority. Thanks. Kevin -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple -- View this message in context: http://old.nabble.com/git-stopped-working-with-1.7.1-tp26905956p28163005.html Sent from the Cygwin list mailing list archive at Nabble.com. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On 4/7/2010 5:12 AM, nothize wrote: After I've replaced(or removed!) %windir%\system32\cygz.dll with the newer one in Cygwin, git 1.7.0.4 worked well. Perhaps I've manually copied an older cygz.dll to there sometimes ago.I can't remember. The Cygwin setup program never installs files into system32. You most likely have some other piece of software which embedded Cygwin and helpfully broke your general Cygwin installation by installing Cygwin DLLs to system32 for you. If you want to have a functional Cygwin installation, you would be better off removing the offending software and ensuring that all Cygwin DLLs are removed from system32. Note that I have the registry value to allow loading DLL from the same directory preceding the path but %windir%\system32 is always the first in the PATH environment variable. This is a well-documented behavior of Windows: http://msdn.microsoft.com/en-us/library/ms682586%28VS.85%29.aspx Many of the git subcommands now live in /usr/lib/git-core, so none of the Cygwin DLLs exist in the same directory as the subcommand binaries. If SafeDllSearchMode is disabled, the current directory is searched next, but no one is likely to have copies of Cygwin DLLs in every directory in which they may be running commands. The next fallback in both safe and non-safe mode is the system directory (system32). When the matching DLL is found there, it preempts the use of the DLL you intended to find via PATH. This is why it is BAD, BAD, BAD to install third-party DLLs into system32. :-) -Jeremy -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
Brian L. br...@blucz.com wrote: From this thread I gather that little or no progress has been made on addressing the root cause of this issue. It still seems to be broken with the latest 1.7 series packages. What is the proper method for getting this bug in front of the guys who maintain the openssh package? The cygwin website suggests that the mailing list is the proper channel, but it doesn't seem like anyone's seriously looking into this as an openssh issue. It's clearly affecting more than a few users and the workaround, while reliable, is quite lame. Can this please get a little bit more priority? and another month goes by... If there is work going on behind the scenes, it would be nice to hear from the developer(s) working on it. If it is not being worked on... I agree with Brian. This is a serious impediment to using Cygwin 1.7. I would think it would be of top priority. Thanks. Kevin -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On 3/11/2010 1:50 PM, Kevin Layer wrote: Brian L.br...@blucz.com wrote: From this thread I gather that little or no progress has been made on addressing the root cause of this issue. It still seems to be broken with the latest 1.7 series packages. What is the proper method for getting this bug in front of the guys who maintain the openssh package? The cygwin website suggests that the mailing list is the proper channel, but it doesn't seem like anyone's seriously looking into this as an openssh issue. It's clearly affecting more than a few users and the workaround, while reliable, is quite lame. Can this please get a little bit more priority? and another month goes by... If there is work going on behind the scenes, it would be nice to hear from the developer(s) working on it. If it is not being worked on... I agree with Brian. This is a serious impediment to using Cygwin 1.7. I would think it would be of top priority. As always, it's worth noting that one way to jump-start activity in an area of open source that is of interest to you is dig in and offer some insights to the problem that's affecting you. I guess this goes without saying though, sometimes, it's beneficial to say things that don't need to be said. ;-) -- Larry Hall http://www.rfk.com RFK Partners, Inc. (508) 893-9779 - RFK Office 216 Dalton Rd. (508) 893-9889 - FAX Holliston, MA 01746 _ A: Yes. Q: Are you sure? A: Because it reverses the logical flow of conversation. Q: Why is top posting annoying in email? -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
Kevin Layer wrote: Brian L. wrote: It's clearly affecting more than a few users and the workaround, while reliable, is quite lame. Can this please get a little bit more priority? and another month goes by... If there is work going on behind the scenes, it would be nice to hear from the developer(s) working on it. If it is not being worked on... I agree with Brian. This is a serious impediment to using Cygwin 1.7. I would think it would be of top priority. Only people who experience the problem can diagnose, debug, and fix it. git works fine for me under cygwin-1.7.1 (Vista, NTFS), so... Has the cygwin git maintainer been able to verify/reproduce the error? If not, then there's no way he can help you. -- Chuck -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On 03/11/2010 04:03 PM, Charles Wilson wrote: If it is not being worked on... I agree with Brian. This is a serious impediment to using Cygwin 1.7. I would think it would be of top priority. Only people who experience the problem can diagnose, debug, and fix it. git works fine for me under cygwin-1.7.1 (Vista, NTFS), so... Has the cygwin git maintainer been able to verify/reproduce the error? If not, then there's no way he can help you. I have sometimes seen the problem, but it has never been a show-stopper for me (I fall back to http cloning), and, so far, it has not percolated high enough to the top of my TODO list to be worth the investment of my time on it. -- Eric Blake ebl...@redhat.com+1-801-349-2682 Libvirt virtualization library http://libvirt.org volunteer cygwin git maintainer signature.asc Description: OpenPGP digital signature
Re: git stopped working with 1.7.1
I ran into the same problem today. After looking at http://gitdiscoveries.wordpress.com/2010/02/21/git-and-cygwin-part-2/, I tried installing GIT 1.7.0 from source, but same as the author, it did not fix the problem either, so I'm guessing that the rest of the people in this thread are right, and it is an OpenSSH issue with cygwin 1.7.1. I really hope somebody will come up with a good workaround, other than using that plink fix. -- View this message in context: http://old.nabble.com/git-stopped-working-with-1.7.1-tp26905956p27677295.html Sent from the Cygwin list mailing list archive at Nabble.com. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
From this thread I gather that little or no progress has been made on addressing the root cause of this issue. It still seems to be broken with the latest 1.7 series packages. What is the proper method for getting this bug in front of the guys who maintain the openssh package? The cygwin website suggests that the mailing list is the proper channel, but it doesn't seem like anyone's seriously looking into this as an openssh issue. It's clearly affecting more than a few users and the workaround, while reliable, is quite lame. Can this please get a little bit more priority? On Mon, Jan 25, 2010 at 6:17 AM, Andre Loker lo...@gmx.de wrote: alland wrote: I have the same issue trying to clone a repo with a recently updated cygwin. The problem is not present in any other configurations I tried to clone with, including git/ssh built for MSys on the same computer. I can confirm the problems myself (cygwin 1.7.1, git 1.6.6.1). Also, using plink/pageant instead of OpenSSH works as a workaround for me for the moment, though it's not really convenient. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
alland wrote: I have the same issue trying to clone a repo with a recently updated cygwin. The problem is not present in any other configurations I tried to clone with, including git/ssh built for MSys on the same computer. I can confirm the problems myself (cygwin 1.7.1, git 1.6.6.1). Also, using plink/pageant instead of OpenSSH works as a workaround for me for the moment, though it's not really convenient. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
I have the same issue trying to clone a repo with a recently updated cygwin. The problem is not present in any other configurations I tried to clone with, including git/ssh built for MSys on the same computer. -- View this message in context: http://old.nabble.com/git-stopped-working-with-1.7.1-tp26905956p27300928.html Sent from the Cygwin list mailing list archive at Nabble.com. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On Jan 18 00:17, ol42 wrote: But obviously the environment variable has still an effect! Without binmode I get the following error $ git clone ssh://o...@simulacron/home/git/gen-dsp.git Initialized empty Git repository in /home/ol/tmp/gen-dsp/.git/ remote: Counting objects: 979, done. remote: Compressing objects: 100% (966/966), done. fatal: The remote end hung up unexpectedly fatal: early EOF fatal: index-pack failed If I have CYGWIN=binmode defined it obviously works --- $ git clone ssh://o...@simulacron/home/git/gen-dsp.git Initialized empty Git repository in /home/ol/tmp/gen-dsp/.git/ remote: Counting objects: 979, done. remote: Compressing objects: 100% (966/966), done. remote: Total 979 (delta 483), reused 0 (delta 0) Receiving objects: 100% (979/979), 950.64 KiB | 1269 KiB/s, done. Resolving deltas: 100% (483/483), done. Checking out files: 100% (196/196), done. Dunno why you see two different results, but the CYGWIN=binmode setting most certainly doesn't exist anymore in Cygwin 1.7. Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Project Co-Leader cygwin AT cygwin DOT com Red Hat -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
But obviously the environment variable has still an effect! Without binmode I get the following error $ git clone ssh://o...@simulacron/home/git/gen-dsp.git Initialized empty Git repository in /home/ol/tmp/gen-dsp/.git/ remote: Counting objects: 979, done. remote: Compressing objects: 100% (966/966), done. fatal: The remote end hung up unexpectedly fatal: early EOF fatal: index-pack failed If I have CYGWIN=binmode defined it obviously works --- $ git clone ssh://o...@simulacron/home/git/gen-dsp.git Initialized empty Git repository in /home/ol/tmp/gen-dsp/.git/ remote: Counting objects: 979, done. remote: Compressing objects: 100% (966/966), done. remote: Total 979 (delta 483), reused 0 (delta 0) Receiving objects: 100% (979/979), 950.64 KiB | 1269 KiB/s, done. Resolving deltas: 100% (483/483), done. Checking out files: 100% (196/196), done. Regards Otto Ken Brown-6 wrote: On 1/16/2010 9:44 AM, ol42 wrote: CYGWIN=binmode This is obsolete. See http://www.cygwin.com/cygwin-ug-net/using-cygwinenv.html#cygwinenv-implemented-options Ken -- View this message in context: http://old.nabble.com/git-stopped-working-with-1.7.1-tp26905956p27207290.html Sent from the Cygwin list mailing list archive at Nabble.com. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
It is obviously a problem with LF to CRLF conversion of cygwin. It can be deacted by inserting the line CYGWIN=binmode on top of the cygwin.bat file. The same issue is happening with CVS. Refer also to http://www.gjt.org/mlist/jcvs/msg01888.html Kevin Layer wrote: Eric Blake e...@byu.net wrote: Yes, as maintainer of the git package on cygwin, I've seen sporadic failures of the git protocol, which I have always ended up working around by switching over to an http protocol. I'm assuming that it might either be BLODA or an actual bug in cygwin1.dll regarding handling of socket fds across forks, but haven't ever been bothered by it enough to try and debug it, because the workaround of http always works. Also, given that other versions of SSH work with the 1.7 git, I would think this is pretty good evidence that it's a bug in cygwin1.dll. Date: Thu, 31 Dec 2009 14:04:16 -0500 Message-ID: 6ee4c8380912311104i1869d73dk818e041f2d8b2...@mail.gmail.com Subject: Re: git stopped working with 1.7.1 From: Brian L. br...@blucz.com To: cygwin@cygwin.com ... The problem seems to be rooted in cygwin's openssh package. If I point GIT_SSH at plink.exe (from the putty distribution) and configure pageant properly for keyless login then git works fine. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple -- View this message in context: http://old.nabble.com/git-stopped-working-with-1.7.1-tp26905956p27189941.html Sent from the Cygwin list mailing list archive at Nabble.com. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On 1/16/2010 9:44 AM, ol42 wrote: CYGWIN=binmode This is obsolete. See http://www.cygwin.com/cygwin-ug-net/using-cygwinenv.html#cygwinenv-implemented-options Ken -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
Eric Blake e...@byu.net wrote: According to Kevin Layer on 1/8/2010 11:01 AM: Has anyone in the cygwin group reproduced this? Yes, as maintainer of the git package on cygwin, I've seen sporadic failures of the git protocol, which I have always ended up working around by switching over to an http protocol. I'm assuming that it might either be BLODA or an actual bug in cygwin1.dll regarding handling of socket fds across forks, but haven't ever been bothered by it enough to try and debug it, because the workaround of http always works. Some of us can't use git w/HTTP, so I hope this gets fixed sooner rather than later. Kevin -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
Eric Blake e...@byu.net wrote: Yes, as maintainer of the git package on cygwin, I've seen sporadic failures of the git protocol, which I have always ended up working around by switching over to an http protocol. I'm assuming that it might either be BLODA or an actual bug in cygwin1.dll regarding handling of socket fds across forks, but haven't ever been bothered by it enough to try and debug it, because the workaround of http always works. Also, given that other versions of SSH work with the 1.7 git, I would think this is pretty good evidence that it's a bug in cygwin1.dll. Date: Thu, 31 Dec 2009 14:04:16 -0500 Message-ID: 6ee4c8380912311104i1869d73dk818e041f2d8b2...@mail.gmail.com Subject: Re: git stopped working with 1.7.1 From: Brian L. br...@blucz.com To: cygwin@cygwin.com ... The problem seems to be rooted in cygwin's openssh package. If I point GIT_SSH at plink.exe (from the putty distribution) and configure pageant properly for keyless login then git works fine. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
Has anyone in the cygwin group reproduced this? -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
According to Kevin Layer on 1/8/2010 11:01 AM: Has anyone in the cygwin group reproduced this? Yes, as maintainer of the git package on cygwin, I've seen sporadic failures of the git protocol, which I have always ended up working around by switching over to an http protocol. I'm assuming that it might either be BLODA or an actual bug in cygwin1.dll regarding handling of socket fds across forks, but haven't ever been bothered by it enough to try and debug it, because the workaround of http always works. -- Don't work too hard, make some time for fun as well! Eric Blake e...@byu.net signature.asc Description: OpenPGP digital signature
Re: git stopped working with 1.7.1
Kevin Layer layer at franz.com writes: layer at hobart128 /c/tmp $ git clone git:/repo/git/acl acl.test Initialized empty Git repository in /c/tmp/acl.test/.git/ remote: Counting objects: 9205, done. remote: Compressing objects: 100% (3300/3300), done. fatal: The remote end hung up unexpectedly fatal: early EOFs: 62% (5708/9205) fatal: index-pack failed layer at hobart128 /c/tmp $ This is on XP 64-bit. Also happens on Windows 7 64-bit (a fresh install of 1.7.1). Exactly the same thing happens here with 1.7.1 under 64-bit Windows 7. However my Cygwin is not a new install but an upgrade of a previous 1.5 installation -- in which Git worked just fine. Of course, just as you say, this was an older version of Git but I don't think this explains the problem because msysGit 1.6.5.1, which I installed just to test if it works, does work fine. So the problem is either some bug in Git 1.6.4 which was not present in 1.5.whatever version that was included in Cygwin 1.5 and is not present in Git 1.6.5 neither or a bug in Cygwin itself, which seems much more likely. And, just to remove any doubts, there are no problems at the server side. I don't use plink, just plain old Cygwin ssh, and the connection is well established and fetching using git from another Linux box works fine too. So the problem is clearly local to Windows machine with Cygwin and is almost certainly due to some bug or change of behaviour in Cygwin itself. Unfortunately I have no clue about how to debug it. I can run git fetch with -v option but it doesn't seem to tell me anything useful. I include a dump of its output below but, again, I don't expect anybody to make much sense out of it. What I'd like to know is how to debug this further because currently my Windows machine is simply unusable because I have everything in various Git repositories which I can't update any longer. Is going back to Cygwin 1.5 really the only solution? Thanks in advance for any ideas, VZ P.S. An example git fetch output: % git fetch -vv Server supports multi_ack Server supports side-band-64k Server supports ofs-delta Marking f05b21d2620cc3870f056f7ead52b9d748aec5ca as complete Marking 48e67403e6836895d4b8519a3d00f7ee35da27f8 as complete Marking 48e67403e6836895d4b8519a3d00f7ee35da27f8 as complete Marking 48e67403e6836895d4b8519a3d00f7ee35da27f8 as complete Marking 2c4fbf10d7f8f7e7aa47bf71c08de09d26742c99 as complete Marking 84572aa72060cfa4014de80baa2c53d58f2bce15 as complete Marking 29f2ec01e35f6557ccebe86fd197aa918f5767ce as complete Marking 884506852efdd169130f30731b1826645896ae72 as complete Marking 608a6ee830a421fc1930bbcdaa518ae544a17b91 as complete Marking c7d2a36cc3c5dbeaf878e51068990f1caac81e20 as complete Marking 0816d63c262e58ede3c36121187ad5e45d446f9c as complete Marking 94751ccb93bff9e01887be4098fa4afe4fa7b637 as complete Marking 0aedb4c06c0eaedc770466137f5be41bbcd9e208 as complete Marking 644b203810e25b8d81c8a4a3effdb8e7fb1f1b68 as complete Marking 2665a1035bb3052987db9682504a5a11be7a4fa6 as complete Marking dc18d12ba60ae2567cf7dbc3f35419b1a68a8190 as complete Marking cb4047c1ac9744848b5bd076394d2cd1948b6298 as complete Marking 2a55b4b480b58563d5b2cb86863339413e681fa3 as complete want db67a9b73270783452b5a9995a20358e567fbd1d (refs/heads/master) have f05b21d2620cc3870f056f7ead52b9d748aec5ca have 48e67403e6836895d4b8519a3d00f7ee35da27f8 have 2c4fbf10d7f8f7e7aa47bf71c08de09d26742c99 have 84572aa72060cfa4014de80baa2c53d58f2bce15 have 29f2ec01e35f6557ccebe86fd197aa918f5767ce have 884506852efdd169130f30731b1826645896ae72 have 608a6ee830a421fc1930bbcdaa518ae544a17b91 have c7d2a36cc3c5dbeaf878e51068990f1caac81e20 have 0816d63c262e58ede3c36121187ad5e45d446f9c have 94751ccb93bff9e01887be4098fa4afe4fa7b637 have 0aedb4c06c0eaedc770466137f5be41bbcd9e208 have 644b203810e25b8d81c8a4a3effdb8e7fb1f1b68 have 2665a1035bb3052987db9682504a5a11be7a4fa6 have dc18d12ba60ae2567cf7dbc3f35419b1a68a8190 have cb4047c1ac9744848b5bd076394d2cd1948b6298 have 2a55b4b480b58563d5b2cb86863339413e681fa3 have bef431623bdf4d2b7deb1812e5d81bb1902c0893 have 1b8fbf665d981f9d0e7ef51da9f2c50553f0277d have 775e2d9f6dbf807992ffd8cf9e8aa4db2c293a5d have 91e1f7c43873ff94496071ca986bbc1347382d74 have 150df99282735e7bccec31c7293e5ef160636876 have 9c8fa09b2c35da056e067599ba3b806774c4a231 have 61322b31cd5f9620b2b8a254438204c31c7f367b have 950caad6ad2b47c5353b2a5957975cd4df76d095 have c9d2d80542f87455347cf0def54dfa4e3745cfa2 have c4b6ed5c0f2798267164b2097cb7939bd7e9015d have b8b1cd1b2a6cae1149e3a9ab8ce1bb9a576b15a3 have 94f9c85dd4405381cfdad150b0ca0fa17977c541 have 63efd028897262949bf5e00f355804ad3e34c22d have 8015275517d233802c3bc4de099a4b637a913aab have 8e83bb623df97863a21adfea8434bd3257035e4f have 6bbb9692703f48121e8bf5af4dd784a4eec3119a have 750d2cd80284d69e5a00938405e253025a71105b have 83925804159e829c6be5e5aac4d7c3605eaa5d00 have e52f04e32d1d12fc3d2c80d669aa0ede054ba3aa have 342bf4dea699c2e45b088afe6c763c7bca255ae6 have
Re: git stopped working with 1.7.1
2010/1/5 Vadim Zeitlin: Exactly the same thing happens here with 1.7.1 under 64-bit Windows 7. However my Cygwin is not a new install but an upgrade of a previous 1.5 installation -- in which Git worked just fine. Does the locale setting make a difference? For example, try it with the following: - LANG=en_US.UTF-8 - LANG=en_US.ISO-8859-1 - LANG=C - LANG=(nothing) Andy -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
I wrote: Unfortunately I have no clue about how to debug it. ... Is going back to Cygwin 1.5 really the only solution? Sorry, I should have read the thread more carefully before writing this, the message from Brian L does provide a solution: using plink.exe instead of Cygwin ssh works fine so there is at least a temporary workaround. Still no idea about what's wrong with ssh though, interactive logins work just fine. FWIW: P.S. An example git fetch output: % git fetch -vv Server supports multi_ack Server supports side-band-64k Server supports ofs-delta ... got ack (2) 8ccfe563ad6b9ce0c549b168e135d09a25818caf got ack (1) 8ccfe563ad6b9ce0c549b168e135d09a25818caf remote: Counting objects: 1617, done. remote: Compressing objects: 100% (1033/1033), done. With plink the lines until now are exactly the same. But then, instead of fatal: The remote end hung up unexpectedly fatal: early EOFs: 15% (156/1038) fatal: index-pack failed I get the expected Receiving objects: 100% (1038/1038), 400.62 KiB, done. Resolving deltas: 100% (960/960), completed with 564 local objects. and everything works correctly. Thanks a lot to Brian for narrowing down the problem and providing a workaround! VZ -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
David Antliff david.antl...@gmail.com wrote: On Thu, Dec 24, 2009 at 07:55, Kevin Layer la...@franz.com wrote: la...@hobart128 /c/tmp $ git clone git:/repo/git/acl acl.test Initialized empty Git repository in /c/tmp/acl.test/.git/ remote: Counting objects: 9205, done. remote: Compressing objects: 100% (3300/3300), done. fatal: The remote end hung up unexpectedly fatal: early EOFs: 62% (5708/9205) fatal: index-pack failed I'm no git expert, but that looks to me like the remote side (where the repository is stored) is experiencing the error while it's preparing data for transfer, and your local git is simply reporting the remote error. It also looks like the remote side is actually the same machine but you're using the git:// protocol to access it without specifying a remote server. I've never tried this and would have expected instead to see something like: I'm not using the git protocol. Note the single slash. The machine is named `git', which is what is confusing you. Anything of the form foo:/path uses SSH, which is what this is using. The server is hanging up, yes, but perhaps the client isn't sending the right responses. Since it looks like the remote is on the same machine as your shell, do you have write access to the actual repository? If so, you could run git-fsck on the repository to make sure it's intact. No, it's on a different machine, but there is absolutely nothing wrong with the repo. I've tried *THREE* Windows machines now. On all, git stops working when cygwin is upgraded. Works before. Doesn't work after. Period. It is not a problem on the server. There are 10+ of people using it every minute of the day and no one has had a problem. Ever. What about other repositories, do they behave the same way, or is your problem restricted to this one? It's random (at what percentage) where it dies (with this repo). It randomly dies with other repos, too. I did install 1.5.25 on the 64-bit Windows 7 machine and git works fine. I've sucked an enormous amount of data off the git server with it, too, and not one problem. Kevin -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
Brian L. br...@blucz.com wrote: I'm seeing very similar bad behavior from cygwin+git on win7 x64 as well as winxp x86. This bug is not confined to 64 bit platforms. This bug is new in 1.7.x--I have cygwin 1.5 installs on both of these machines that do not exhibit this failure. The problem seems to be rooted in cygwin's openssh package. If I point GIT_SSH at plink.exe (from the putty distribution) and configure pageant properly for keyless login then git works fine. Good sleuthing! -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On Mon, Jan 4, 2010 at 15:06, Kevin Layer wrote: I'm not using the git protocol. Note the single slash. The machine is named `git', which is what is confusing you. Anything of the form foo:/path uses SSH, which is what this is using. Yes, you're right, I'm so used to seeing git://server syntax that I was a bit thrown by the git:/path syntax. -- David. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
I'm seeing very similar bad behavior from cygwin+git on win7 x64 as well as winxp x86. This bug is not confined to 64 bit platforms. This bug is new in 1.7.x--I have cygwin 1.5 installs on both of these machines that do not exhibit this failure. The problem seems to be rooted in cygwin's openssh package. If I point GIT_SSH at plink.exe (from the putty distribution) and configure pageant properly for keyless login then git works fine. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
David Antliff david.antl...@gmail.com wrote: It may be a 64-bit issue, so I'll try a 32-bit machine, if I can scrounge one up. We are using WinXP 32-bit and have not seen this problem (yet). It would be really helpful (to me at least) to know whether you can reproduce the issue with 32-bit Windows... I imagine there are a fair number of people using git in Cygwin because the Windows alternatives (msys, etc) do not provide a POSIX-like toolchain environment. Using Cygwin also allows scripts around git to operate on Linux with no modifications. That is why we use Cygwin anyway, and Cygwin + git is very important to us, FWIW. I just tried 32-bit Windows XP Pro. I upgraded from 1.5.25 and installed git (it wasn't previously on that machine). Failed the same way. Kevin -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On Thu, Dec 24, 2009 at 07:55, Kevin Layer la...@franz.com wrote: la...@hobart128 /c/tmp $ git clone git:/repo/git/acl acl.test Initialized empty Git repository in /c/tmp/acl.test/.git/ remote: Counting objects: 9205, done. remote: Compressing objects: 100% (3300/3300), done. fatal: The remote end hung up unexpectedly fatal: early EOFs: 62% (5708/9205) fatal: index-pack failed I'm no git expert, but that looks to me like the remote side (where the repository is stored) is experiencing the error while it's preparing data for transfer, and your local git is simply reporting the remote error. It also looks like the remote side is actually the same machine but you're using the git:// protocol to access it without specifying a remote server. I've never tried this and would have expected instead to see something like: $ git clone git://localhost/repo/git/acl acl.test or $ git clone some-path/repo/git/acl acl.test I.e. there's something about your command that *I* don't understand. It's probably perfectly fine though, and the problem is with me. Are you intending to clone a repository on the same machine but via the git:// protocol? What if you just do this instead: $ git clone some-path/repo/git/acl acl.test Do you get the same error? Since it looks like the remote is on the same machine as your shell, do you have write access to the actual repository? If so, you could run git-fsck on the repository to make sure it's intact. What about other repositories, do they behave the same way, or is your problem restricted to this one? -- David. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On Mon, Dec 28, 2009 at 14:13, Kevin Layer wrote: This seems serious. Do people just not use cygwin git? It sounds very serious. I am a very interested user of git on Cygwin and I'm watching this thread with interest. However It may be a 64-bit issue, so I'll try a 32-bit machine, if I can scrounge one up. We are using WinXP 32-bit and have not seen this problem (yet). It would be really helpful (to me at least) to know whether you can reproduce the issue with 32-bit Windows... I imagine there are a fair number of people using git in Cygwin because the Windows alternatives (msys, etc) do not provide a POSIX-like toolchain environment. Using Cygwin also allows scripts around git to operate on Linux with no modifications. That is why we use Cygwin anyway, and Cygwin + git is very important to us, FWIW. -- David. -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
crickets I'm a little surprised there are *no* replies to this thread, other than mine. (It's not like there is no traffic on the mailing list.) Can someone tell me what I need to do to interest people in this? Let me summarize: git clone fails with this: remote: Compressing objects: 100% (3300/3300), done. fatal: The remote end hung up unexpectedly fatal: early EOFs: 62% (5708/9205) fatal: index-pack failed The clone is via ssh from a box on my local (10 Gig) network. This has happened on *two* machines. Both 64-bit. One XP. One Windows 7. Both have all recent Windows updates. The Windows 7 box is a new one. I installed Windows, the Windows 7 SDK and then cygwin. Git is the first thing I tried, and it failed. The XP box I have used for years. I had a working git until the other day when I upgraded to cygwin 1.7.1--it fails just like the Windows 7 box after the upgrade. This seems serious. Do people just not use cygwin git? It may be a 64-bit issue, so I'll try a 32-bit machine, if I can scrounge one up. Thanks. Kevin -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
On Sun, Dec 27, 2009 at 05:13:39PM -0800, Kevin Layer wrote: crickets I'm a little surprised there are *no* replies to this thread, other than mine. (It's not like there is no traffic on the mailing list.) It's the holiday season. Our support staff are probably out enjoying themselves. cgf -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
la...@hobart128 /c/tmp $ git clone git:/repo/git/acl acl.test Initialized empty Git repository in /c/tmp/acl.test/.git/ remote: Counting objects: 9205, done. remote: Compressing objects: 100% (3300/3300), done. fatal: The remote end hung up unexpectedly fatal: early EOFs: 62% (5708/9205) fatal: index-pack failed I should mention that the percentage at which it fails is random. Out of 40 times I've run the command above, it worked exactly 1 time. Usually, however, the failures are anwhere from 20% to 63% on the early EOFs line. Also, other people seem to be having the issue, too: http://stackoverflow.com/questions/1493645/git-fatal-remote-end-hung-up/1954050#1954050 -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
la...@hobart128 /c/tmp $ git clone git:/repo/git/acl acl.test Initialized empty Git repository in /c/tmp/acl.test/.git/ remote: Counting objects: 9205, done. remote: Compressing objects: 100% (3300/3300), done. fatal: The remote end hung up unexpectedly fatal: early EOFs: 62% (5708/9205) fatal: index-pack failed la...@hobart128 /c/tmp $ I have no problem ssh'ing to the `git' host, btw. I went back to 1.5.25-15 and git works again (of course, it's an older version of git, 1.6.1.2). I tried to install 1.6.1.2 on 1.7.1, but it didn't work. The clone complained that the index-pack program was missing. I did the usual ./configure make make install so I'm not sure what else to try. I still have cygwin 1.7.1 installed on a Windows 7 machine, so I can try 1.6.1.2 if someone can help me get it installed properly. Since Windows 7 isn't supported with 1.5, I really need to figure this out. Thanks for any help you can provide. Kevin -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Re: git stopped working with 1.7.1
I tried to install 1.6.1.2 on 1.7.1, but it didn't work. The clone complained that the index-pack program was missing. I did the usual ./configure make make install so I'm not sure what else to try. I copied git-index-pack.exe from the build directory to /usr/local/libexec/git-core/ and now 1.6.1.2 behaves the same as the cygwin-built git: la...@hobart256 ~ $ git clone git:/repo/git/acl acl Initialized empty Git repository in /home/layer/acl/.git/ remote: Counting objects: 9205, done. remote: Compressing objects: 100% (3300/3300), done. fatal: The remote end hung up unexpectedly fatal: early EOFs: 32% (2946/9205) fatal: index-pack failed la...@hobart256 ~ So, it would seem that something in cygwin, not git, has made this fail, since git 1.6.1.2 is working for me with cygwin 1.5.25. Kevin -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
