Re: Sourcegear Vault
Sorry Lynn, I haven't checked the posting here in a while. I did have it working, but not what I would call well. The Vault CLI is extremely slow when it needs to login to the Vault server, and it does this once for each file in the review. I wanted to create a method of logging into the Vault server on first contact and retain that connection. I was trying to integrate a test .Net utility with (something I found for Python, which name is escaping me now) when what little free-time I had evaporated. I'm hopeful I'll be able to return to this project early next year. John On Oct 28, 1:26 pm, Lynn Roth lynnr...@gmail.com wrote: John, Were you ever able to get Review Board working well with Vault? Thanks, Lynn On Tue, Aug 18, 2009 at 3:47 PM, Christian Hammond chip...@chipx86.comwrote: We won't support Review Board running on IronPython. That's a no-go. However, a utility program for communicating with the server could be written using IronPython (or whatever) and Review Board can execute that, much like we do with git and p4. Christian -- Christian Hammond - chip...@chipx86.com Review Board -http://www.review-board.org VMware, Inc. -http://www.vmware.com On Tue, Aug 18, 2009 at 12:14 PM, schuijo schu...@gmail.com wrote: Any chance Review-Board will work with IronPython? (yes, the Vault libraries are .NET) On Aug 18, 11:01 am, schuijo schu...@gmail.com wrote: Christian, It does, but requires a login first. It has a SOAP interface defined, but SourceGear highly recommends NOT using it. I have a Python question for you (sorry, until a couple weeks ago I never even heard of Python ;-) It seems the (time-wise) expensive steps in this Vault integration are the Java VM for the Java client, and the login step each time the CLC is called. I'm thinking that creating some form of a DLL interface that will stay logged into Vault would greatly improve our response times. Does Python have support for DLL integration? COM? Thanks! On Aug 17, 1:23 pm, Christian Hammond chip...@chipx86.com wrote: Hi John, Yeah, that seems a bit expensive, slow, and unscalable... Does Vault happen to have built-in support for looking at raw files from a web page, given a filename and revision? If so, it can take advantage of the new raw URL field I'm adding for Git. If not, we'll need to figure out a solution that doesn't require Java. Christian -- Christian Hammond - chip...@chipx86.com Review Board -http://www.review-board.org VMware, Inc. -http://www.vmware.com On Mon, Aug 17, 2009 at 8:31 AM, schuijo schu...@gmail.com wrote: Thanks Christian, that seemed to be the case. Just a note to anyone attempting to integrate with Sourcegear's Vault: I originally attempted to use their Java CLC for portability, but each file in the review needed to contact the Vault server twice and each contact would (re)spawn the Java VM which was very expensive time wise. I'm currently testing the system with the Win32 CLC and it is slow, but usable. I'm considering looking into the possiblity of batching the contacts to the Vault server...time permitting. Thanks again for all your help Christian!! John On Aug 12, 7:21 pm, Christian Hammond chip...@chipx86.com wrote: Hi, It's possible that when you use the cache function, it's returning cached data from some older, broken attempt. If you reenable the caching and then fully clear the cache and try again, does it work? Christian On Wednesday, August 12, 2009, schuijo schu...@gmail.com wrote: More: I traced the source of that parameter to get_original_file() in diffutils.py. When I bypass the cache lookup and just return the result of the fetch_file() sub-function I can display diffs properly, although response is very slow as expected. Any ideas where I should look to determine why the cache contents would be wrong? (I do have the memcached server installed) Thanks! On Aug 12, 5:54 pm, schuijo schu...@gmail.com wrote: I think I am very close now, but I have a problem in diffutils.py. The second parameter in the parse() function (file I believe) contains some invalid data, where does this get populated? Thanks! On Aug 11, 2:29 pm, Christian Hammond chip...@chipx86.com wrote: Hi, Path is the path to the file in the repository. If the path in the diff is always going to be absolute, then you can completely ignore base_path and just use path. We use base_path for Subversion, where the filenames in the diff are relative to the current directory rather than the root of the repository. We then append path to base_path to generate that absolute path
Re: Sourcegear Vault
Christian, It does, but requires a login first. It has a SOAP interface defined, but SourceGear highly recommends NOT using it. I have a Python question for you (sorry, until a couple weeks ago I never even heard of Python ;-) It seems the (time-wise) expensive steps in this Vault integration are the Java VM for the Java client, and the login step each time the CLC is called. I'm thinking that creating some form of a DLL interface that will stay logged into Vault would greatly improve our response times. Does Python have support for DLL integration? COM? Thanks! On Aug 17, 1:23 pm, Christian Hammond chip...@chipx86.com wrote: Hi John, Yeah, that seems a bit expensive, slow, and unscalable... Does Vault happen to have built-in support for looking at raw files from a web page, given a filename and revision? If so, it can take advantage of the new raw URL field I'm adding for Git. If not, we'll need to figure out a solution that doesn't require Java. Christian -- Christian Hammond - chip...@chipx86.com Review Board -http://www.review-board.org VMware, Inc. -http://www.vmware.com On Mon, Aug 17, 2009 at 8:31 AM, schuijo schu...@gmail.com wrote: Thanks Christian, that seemed to be the case. Just a note to anyone attempting to integrate with Sourcegear's Vault: I originally attempted to use their Java CLC for portability, but each file in the review needed to contact the Vault server twice and each contact would (re)spawn the Java VM which was very expensive time wise. I'm currently testing the system with the Win32 CLC and it is slow, but usable. I'm considering looking into the possiblity of batching the contacts to the Vault server...time permitting. Thanks again for all your help Christian!! John On Aug 12, 7:21 pm, Christian Hammond chip...@chipx86.com wrote: Hi, It's possible that when you use the cache function, it's returning cached data from some older, broken attempt. If you reenable the caching and then fully clear the cache and try again, does it work? Christian On Wednesday, August 12, 2009, schuijo schu...@gmail.com wrote: More: I traced the source of that parameter to get_original_file() in diffutils.py. When I bypass the cache lookup and just return the result of the fetch_file() sub-function I can display diffs properly, although response is very slow as expected. Any ideas where I should look to determine why the cache contents would be wrong? (I do have the memcached server installed) Thanks! On Aug 12, 5:54 pm, schuijo schu...@gmail.com wrote: I think I am very close now, but I have a problem in diffutils.py. The second parameter in the parse() function (file I believe) contains some invalid data, where does this get populated? Thanks! On Aug 11, 2:29 pm, Christian Hammond chip...@chipx86.com wrote: Hi, Path is the path to the file in the repository. If the path in the diff is always going to be absolute, then you can completely ignore base_path and just use path. We use base_path for Subversion, where the filenames in the diff are relative to the current directory rather than the root of the repository. We then append path to base_path to generate that absolute path. Christian -- Christian Hammond - chip...@chipx86.com Review Board -http://www.review-board.org VMware, Inc. -http://www.vmware.com On Tue, Aug 11, 2009 at 10:54 AM, schuijo schu...@gmail.com wrote: Christian, Ok...I think I've severely bitten off more than I can chew, but I'm trying to forge my way through this. I've been modifying post-review to add Vault support, and appear to have it working to the point where it seems to be attempting to contact the Vault server while uploading the diff. The point I'm a little lost/confused on is how to represent Vault in the RepositoryInfo class. What exactly are path and base_path and how are they used? (hopefully this will help me to determine what need to be populated in there for Vault) Thanks! On Jul 28, 4:05 pm, Christian Hammond chip...@chipx86.com wrote: There are some threads on the mailing list about doing this, but they're not exactly step-by-step tutorials. The best reference right now is the scmtools/*.py files. Basically, you'll create a subclass of SCMTool that does the following: 1) Grabs a file from a repository, given a file path and revision. 2) Provide a DiffParser subclass that handles pulling out filenames and revisions and any other necessary data from a diff (most of the code for all this is common, so you just hook into things -- see the other files for examples). 3) If Vault has a concept of server-side changesets (you register a changeset
Re: Sourcegear Vault
More: I traced the source of that parameter to get_original_file() in diffutils.py. When I bypass the cache lookup and just return the result of the fetch_file() sub-function I can display diffs properly, although response is very slow as expected. Any ideas where I should look to determine why the cache contents would be wrong? (I do have the memcached server installed) Thanks! On Aug 12, 5:54 pm, schuijo schu...@gmail.com wrote: I think I am very close now, but I have a problem in diffutils.py. The second parameter in the parse() function (file I believe) contains some invalid data, where does this get populated? Thanks! On Aug 11, 2:29 pm, Christian Hammond chip...@chipx86.com wrote: Hi, Path is the path to the file in the repository. If the path in the diff is always going to be absolute, then you can completely ignore base_path and just use path. We use base_path for Subversion, where the filenames in the diff are relative to the current directory rather than the root of the repository. We then append path to base_path to generate that absolute path. Christian -- Christian Hammond - chip...@chipx86.com Review Board -http://www.review-board.org VMware, Inc. -http://www.vmware.com On Tue, Aug 11, 2009 at 10:54 AM, schuijo schu...@gmail.com wrote: Christian, Ok...I think I've severely bitten off more than I can chew, but I'm trying to forge my way through this. I've been modifying post-review to add Vault support, and appear to have it working to the point where it seems to be attempting to contact the Vault server while uploading the diff. The point I'm a little lost/confused on is how to represent Vault in the RepositoryInfo class. What exactly are path and base_path and how are they used? (hopefully this will help me to determine what need to be populated in there for Vault) Thanks! On Jul 28, 4:05 pm, Christian Hammond chip...@chipx86.com wrote: There are some threads on the mailing list about doing this, but they're not exactly step-by-step tutorials. The best reference right now is the scmtools/*.py files. Basically, you'll create a subclass of SCMTool that does the following: 1) Grabs a file from a repository, given a file path and revision. 2) Provide a DiffParser subclass that handles pulling out filenames and revisions and any other necessary data from a diff (most of the code for all this is common, so you just hook into things -- see the other files for examples). 3) If Vault has a concept of server-side changesets (you register a changeset with a description, and other data, and the server always knows what you have checked out on the client) then you'll need to implement get_changeset(). So the general model is that this code will have three classes: 1) VaultTool 2) VaultDiffParser 3) VaultClient VaultTool will be a subclass of SCMTool and will be what Review Board talks to. VaultDiffParser will be a subclass of DiffParser and will override functions to parse revision info out of a diff. VaultClient will be a wrapper around the command line tool, which VaultClient will talk to. Now, let's talk diffs. Many revision control systems provide tools that generate diffs unsuitable for Review Board, and sometimes we have to work around them. If vault's tool generates a diff containing revision information for a file that can be used to pull data from the repository, then we're good. If not, you'll need to implement this in post-review. Basically, generate a custom diff, or post-process the generated diff, to stuff revision information in. Then VaultDiffParser will parse that out. You'll see this done for Perforce and many others in post-review. Christian -- Christian Hammond - chip...@chipx86.com Review Board -http://www.review-board.org VMware, Inc. -http://www.vmware.com On Tue, Jul 28, 2009 at 1:42 PM, schuijo schu...@gmail.com wrote: I couldn't locate a Python API for Vault, but they do have a command line tool. Is there a document or thread detailing how to implement the interface? Thanks! On Jul 28, 3:00 pm, Christian Hammond chip...@chipx86.com wrote: Not to my knowledge. Some have talked about wanting it, but I don't know if anyone's working on it. If you'd like to work on it, we'll provide guidance. The only real requirement is that there's a command line tool or Python API for grabbing a file from a remote server. Everything else can be done by us. It does help, though, if the tool/API can generate a suitable diff, or at least one we can post-process. Christian -- Christian Hammond - chip...@chipx86.com Review Board -http://www.review-board.org VMware, Inc. -http://www.vmware.com
Re: Sourcegear Vault
Christian, Ok...I think I've severely bitten off more than I can chew, but I'm trying to forge my way through this. I've been modifying post-review to add Vault support, and appear to have it working to the point where it seems to be attempting to contact the Vault server while uploading the diff. The point I'm a little lost/confused on is how to represent Vault in the RepositoryInfo class. What exactly are path and base_path and how are they used? (hopefully this will help me to determine what need to be populated in there for Vault) Thanks! On Jul 28, 4:05 pm, Christian Hammond chip...@chipx86.com wrote: There are some threads on the mailing list about doing this, but they're not exactly step-by-step tutorials. The best reference right now is the scmtools/*.py files. Basically, you'll create a subclass of SCMTool that does the following: 1) Grabs a file from a repository, given a file path and revision. 2) Provide a DiffParser subclass that handles pulling out filenames and revisions and any other necessary data from a diff (most of the code for all this is common, so you just hook into things -- see the other files for examples). 3) If Vault has a concept of server-side changesets (you register a changeset with a description, and other data, and the server always knows what you have checked out on the client) then you'll need to implement get_changeset(). So the general model is that this code will have three classes: 1) VaultTool 2) VaultDiffParser 3) VaultClient VaultTool will be a subclass of SCMTool and will be what Review Board talks to. VaultDiffParser will be a subclass of DiffParser and will override functions to parse revision info out of a diff. VaultClient will be a wrapper around the command line tool, which VaultClient will talk to. Now, let's talk diffs. Many revision control systems provide tools that generate diffs unsuitable for Review Board, and sometimes we have to work around them. If vault's tool generates a diff containing revision information for a file that can be used to pull data from the repository, then we're good. If not, you'll need to implement this in post-review. Basically, generate a custom diff, or post-process the generated diff, to stuff revision information in. Then VaultDiffParser will parse that out. You'll see this done for Perforce and many others in post-review. Christian -- Christian Hammond - chip...@chipx86.com Review Board -http://www.review-board.org VMware, Inc. -http://www.vmware.com On Tue, Jul 28, 2009 at 1:42 PM, schuijo schu...@gmail.com wrote: I couldn't locate a Python API for Vault, but they do have a command line tool. Is there a document or thread detailing how to implement the interface? Thanks! On Jul 28, 3:00 pm, Christian Hammond chip...@chipx86.com wrote: Not to my knowledge. Some have talked about wanting it, but I don't know if anyone's working on it. If you'd like to work on it, we'll provide guidance. The only real requirement is that there's a command line tool or Python API for grabbing a file from a remote server. Everything else can be done by us. It does help, though, if the tool/API can generate a suitable diff, or at least one we can post-process. Christian -- Christian Hammond - chip...@chipx86.com Review Board -http://www.review-board.org VMware, Inc. -http://www.vmware.com On Tue, Jul 28, 2009 at 12:12 PM, schuijo schu...@gmail.com wrote: Is anyone working on support for Sourcegear's Vault? Thanks! --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups reviewboard group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
Sourcegear Vault
Is anyone working on support for Sourcegear's Vault? Thanks! --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups reviewboard group. To post to this group, send email to reviewboard@googlegroups.com To unsubscribe from this group, send email to reviewboard+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/reviewboard?hl=en -~--~~~~--~~--~--~---
