Re: Subversion web development question.
I'd definitely go with SVN for a code repo. I use a couple different SVN servers on various teams I work with at my clients. I also set one up for myself for code I'm working without other coders, mainly so I could get at it from home, on the road, or some client's site; a laptop or two, a desktop or two Very convenient. You might also consider integrating it with the Trac issue tracker . It has a very nice SVN repo code browser, takes bug/issue/feature tickets, offers a wiki (e.g., for writing project plans, docs, whatever). It's integrated in the sense that you can check in code into SVN and say in your log message something like fixes #37 and Track will notice and close the open ticket #37 for you. You can reference code within Trac too. It's lightweight and gets out of your way. I prefer it to other trackers and trouble ticket systems I've used like shudder Remedy /shudder, Jira, and even the venerable RT. Even if all you use is Trac's code browser it's a win, but the other stuff is real helpful with no bloat. FWIW, it's all written in Python. (a language I prefer to PHP and Java) ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Subversion web development question.
Forgive me if this question is stupid or has been answered before, more than likely this info exists, but I'm not looking in the right place. I work in a I'd recommend searching the subversion mailing lists as well. It's been covered, but it's not stupid :) small corporation that is heavily dependent on web apps in which I am the sole developer. Currently we're using ASP on Win2K with SQL Server 2000. We've used RedHat for years on our mail / DNS servers, and have migrated to FreeBSD for both. Since ASP has pretty much been rendered obsolete by .Net, we are eventually going to migrate our web apps. Rather than continue down the road of MS technology, we have decided to migrate to PHP MySQL based on Not to start a war, but if you're switching now, you might want to look into Ruby on Rails as well. Our development infrastructure to date works in this manner. Since I am the sole developer, and will be for the long term future, there has never been Since you're changing things now, assume that you are not the sole developer anymore. You'll be thankful you did when the next developer comes along :) any real need to incorporate SourceSafe. Instead we run a separate development server to do all the coding. An FTP site has been set up to repository on the server. I currently use HomeSite for it's built-in FTP client. Coding is done locally, saved back to the server, and tested on the server in a browser. Testing is not done locally because our needs are such that there is a large number of static text files need to be accessed by our apps. The parameters to access these files are hard to duplicate to a workstation, and it has proved (for us anyway) to be easier to test these apps on a server that is a mirrored environment of the live server. This sound very similar to our situation.. My question is can something like this be replicated on FreeBSD w/ Subversion? Sure. I would like to setup a versioning system, but am at a loss on how the development process would operate. I have found information on how to update files on the live server from the development server, but not much in the way of how to set up a development server to get Subversion to update the files in the Apache directories. It would be trivial for me to simply set up FTP sites that map to the Apache directories, and change the permission structure to allow access to these directories, but I'd rather not create a security headache for myself down the road. Can Subversion be set up to check out a file, commit it back to the server, and test it in a browser from a workstation? Or does the file need to be moved from the Subversion directories to the Apache directories by someone w/ root privileges every time a file has been edited? If your development server is local (ie. on the LAN) to your workstation, I would setup Samba to share your web tree to your workstation. Once this is done you can simply create/edit/delete files right from your workstation and the changes take affect immediately on the development server. The second step is to add subversion. To do this, create the subversion repository on your development server (or on a dedidcated server somewhere) and set it up to allow remote access. Then, go get a graphical SVN client like TortioseSVN (for windows only, but mac clients exist too) and configure it to check out your repository onto that share you mounted earlier. At this point you can manage your files on your workstation and immediately check your results. Then when happy, you can commit your changes with SVN and they get committed to the repository. You don't need subversion to move files into your web tree. That's not what it does. Some other thoughts... If you like the command line, you can skip the graphical SVN client and just do it directly on the development server. That's what I do. If you did that you could also skip Samba and keep on using your FTP client, although I personally hate having to FTP files back and forth. Search the net for books on subversion... there are a couple that are freely available. OReilly also has made their samba book free as well. Good luck! Sorry if this question has been answered somewhere else, it's just that most of the info I seem to be finding is based on application development for locally run binaries in which the app is tested locally before being committed back to the server. Best regards, Greg Groth ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED] ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Subversion web development question.
On 7/24/2006 12:16 PM, Philip Hallstrom wrote: Forgive me if this question is stupid or has been answered before, more than likely this info exists, but I'm not looking in the right place. I work in a I'd recommend searching the subversion mailing lists as well. It's been covered, but it's not stupid :) Will do, thanks. small corporation that is heavily dependent on web apps in which I am the sole developer. Currently we're using ASP on Win2K with SQL Server 2000. We've used RedHat for years on our mail / DNS servers, and have migrated to FreeBSD for both. Since ASP has pretty much been rendered obsolete by .Net, we are eventually going to migrate our web apps. Rather than continue down the road of MS technology, we have decided to migrate to PHP MySQL based on Not to start a war, but if you're switching now, you might want to look into Ruby on Rails as well. Our development infrastructure to date works in this manner. Since I am the sole developer, and will be for the long term future, there has never been Since you're changing things now, assume that you are not the sole developer anymore. You'll be thankful you did when the next developer comes along :) That's the plan :). any real need to incorporate SourceSafe. Instead we run a separate development server to do all the coding. An FTP site has been set up to repository on the server. I currently use HomeSite for it's built-in FTP client. Coding is done locally, saved back to the server, and tested on the server in a browser. Testing is not done locally because our needs are such that there is a large number of static text files need to be accessed by our apps. The parameters to access these files are hard to duplicate to a workstation, and it has proved (for us anyway) to be easier to test these apps on a server that is a mirrored environment of the live server. This sound very similar to our situation.. My question is can something like this be replicated on FreeBSD w/ Subversion? Sure. I would like to setup a versioning system, but am at a loss on how the development process would operate. I have found information on how to update files on the live server from the development server, but not much in the way of how to set up a development server to get Subversion to update the files in the Apache directories. It would be trivial for me to simply set up FTP sites that map to the Apache directories, and change the permission structure to allow access to these directories, but I'd rather not create a security headache for myself down the road. Can Subversion be set up to check out a file, commit it back to the server, and test it in a browser from a workstation? Or does the file need to be moved from the Subversion directories to the Apache directories by someone w/ root privileges every time a file has been edited? If your development server is local (ie. on the LAN) to your workstation, I would setup Samba to share your web tree to your workstation. Once this is done you can simply create/edit/delete files right from your workstation and the changes take affect immediately on the development server. Just for clarification, the dev server is not on the local LAN, it resides on the DMZ. Way back when, our dev server was on the LAN, live server on the DMZ, but we ran into a lot of issues getting some MS specific technologies working when the code was moved from the dev server to the live server. Code written for path statements had to be rewritten to run in the DMZ. Since then, the dev server has resided on the DMZ. The other reason is for remote accessibility. Having the dev server on the DMZ allows me the ability to code from home without having to punch a hole in the firewall to our internal LAN. The second step is to add subversion. To do this, create the subversion repository on your development server (or on a dedidcated server somewhere) and set it up to allow remote access. Then, go get a graphical SVN client like TortioseSVN (for windows only, but mac clients exist too) and configure it to check out your repository onto that share you mounted earlier. At this point you can manage your files on your workstation and immediately check your results. Then when happy, you can commit your changes with SVN and they get committed to the repository. You don't need subversion to move files into your web tree. That's not what it does. Fair enough, so should I assume then that the files need to be moved using SSH (for example) after su-ing to root? If so, I suppose I would be better off then just setting up Apache on my desktop, testing locally, then moving the files back to the server? Some other thoughts... If you like the command line, you can skip the graphical SVN client and just do it directly on the development server. That's what I do. If you did that you could also skip Samba and keep on using your FTP client,
Re: Subversion web development question.
At this point you can manage your files on your workstation and immediately check your results. Then when happy, you can commit your changes with SVN and they get committed to the repository. You don't need subversion to move files into your web tree. That's not what it does. Fair enough, so should I assume then that the files need to be moved using SSH (for example) after su-ing to root? If so, I suppose I would be better off then just setting up Apache on my desktop, testing locally, then moving the files back to the server? Why would you need to be root? Here's what we do at work: The development server is at the data center (since we're all remote). Apache has been configured to run several name based virtual servers and looks for the document root in a 'cpr' in our home directory. (cpr is the initials of our project) For those of us that like the CLI and working directly on the server, the 'cpr' directory is also our subversion working directory. So I edit a file, test it, like it, commit it. repeat. Our designers don't like CLI. What we've done for them is to set them up with TortioseSVN on their workstations. They check out the repository on their desktop. They make changes. When they want to test, they run rsync (via cygwin and a shell script) that syncs only the files they've changed (and skips all the .svn sub dirs) up to the server. They test. They like, they commit using TortioseSVN. All of thise happens over a SSH tunnel. At no point in any of this do any of us need to be root If you can get your dev environment running locally it will be easier since you won't have to upload to test, but depending on your situation that might not be reasonable. -philip ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Subversion web development question.
Why would you need to be root? Here's what we do at work: Because /usr/local/www/apache22/data is owned by root. I guess I mistakenly figured that this is where the files should go. I know that you can configure Apache to point to any directory, but was unsure of the consequences of pointing it at directories outside of /usr/local/www/apache22/data. The development server is at the data center (since we're all remote). Apache has been configured to run several name based virtual servers and looks for the document root in a 'cpr' in our home directory. (cpr is the initials of our project) For those of us that like the CLI and working directly on the server, the 'cpr' directory is also our subversion working directory. So I edit a file, test it, like it, commit it. repeat. Our designers don't like CLI. What we've done for them is to set them up with TortioseSVN on their workstations. They check out the repository on their desktop. They make changes. When they want to test, they run rsync (via cygwin and a shell script) that syncs only the files they've changed (and skips all the .svn sub dirs) up to the server. They test. They like, they commit using TortioseSVN. All of thise happens over a SSH tunnel. At no point in any of this do any of us need to be root If you can get your dev environment running locally it will be easier since you won't have to upload to test, but depending on your situation that might not be reasonable. -philip What you suggest makes perfect sense, I wasn't sure if you could point Apache directly at a directory within Subversion's structure or not. I do have one more question. From a security standpoint would it be safer to point to a symlink within Apache's existing directory structure that points to /usr/home/svnDirectory, or configure Apache to point directly to /usr/home/svnDirectory? It's probably a moot point, but I'm still unsure of myself when it comes to messing around with permissions within FreeBSD and don't like making changes to the defaults without asking first. Many thanks, Greg Groth ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Subversion web development question.
Why would you need to be root? Here's what we do at work: Because /usr/local/www/apache22/data is owned by root. I guess I mistakenly figured that this is where the files should go. I know that you can configure Apache to point to any directory, but was unsure of the consequences of pointing it at directories outside of /usr/local/www/apache22/data. if it's your dev server, point it to wherever is most convenient. The development server is at the data center (since we're all remote). Apache has been configured to run several name based virtual servers and looks for the document root in a 'cpr' in our home directory. (cpr is the initials of our project) For those of us that like the CLI and working directly on the server, the 'cpr' directory is also our subversion working directory. So I edit a file, test it, like it, commit it. repeat. Our designers don't like CLI. What we've done for them is to set them up with TortioseSVN on their workstations. They check out the repository on their desktop. They make changes. When they want to test, they run rsync (via cygwin and a shell script) that syncs only the files they've changed (and skips all the .svn sub dirs) up to the server. They test. They like, they commit using TortioseSVN. All of thise happens over a SSH tunnel. At no point in any of this do any of us need to be root If you can get your dev environment running locally it will be easier since you won't have to upload to test, but depending on your situation that might not be reasonable. -philip What you suggest makes perfect sense, I wasn't sure if you could point Apache directly at a directory within Subversion's structure or not. I do have one I think you should read more about subversion as I think you are confusing a svn repository with a svn working directory... more question. From a security standpoint would it be safer to point to a symlink within Apache's existing directory structure that points to /usr/home/svnDirectory, or configure Apache to point directly to /usr/home/svnDirectory? It's probably a moot point, but I'm still unsure of myself when it comes to messing around with permissions within FreeBSD and don't like making changes to the defaults without asking first. Doesn't make a bit of difference. I'd avoid the symlink simply cause there isn't any reason to have it. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
