Re: Re: Wolfram Workbench
- To: mathgroup at smc.vnet.net
- Subject: [mg68218] Re: [mg68145] Re: [mg67990] Wolfram Workbench
- From: Igor Antonio <igora at wolf-ram.com>
- Date: Sat, 29 Jul 2006 01:00:52 -0400 (EDT)
- Organization: Wolfram Research, Inc.
- References: <200607201004.GAA09748@smc.vnet.net> <200607260626.CAA20404@smc.vnet.net>
- Reply-to: igora at wolf-ram.com
- Sender: owner-wri-mathgroup at wolfram.com
Hi Chris, Chris Chiasson wrote: > Thanks for posting this message. I usually ignore all WRI product > announcements... > > I have never used a revision control system before. I have wanted to > because my website files are XML, and should be easily manageable with > one. Unfortunately, (fully featured) Mathematica notebooks are > resistant to version control, or so I read on the group a long time > ago. You can certainly manage your notebook files using a version control software and use meaningful comments in the commits to help you "manage" the files. However, you won't be able to take "full" advantage of version control programs, such as comparing different revisions. This is because you would be comparing the cell expressions, not their content. In addition, things like: - Cached data in the file - Notebook properties, such as WindowSize, WindowMargins, etc... will be shown as changes between notebooks and you probably wouldn't care about them. > >>From watching the demos at > > http://www.wolfram.com/products/workbench/workflow/ > > it appears WRI has developed a new type of .m and .nb file pairing. In > the new pairing, the .m file appears to the the primary document, > while the notebook is for testing code and generating outputs. Let's > hope no one accidentally uses the option to create a .m file from > initialization files on one of those notebooks. That might overwrite > the source file. > > How should people go about importing old code (developed with the > previous .nb and .m file usage) into this framework? > I'm assuming that in addition to your package code, you also have tests inside your notebook to test the package (tests that are not in initialization cells, so they don't placed in the .m file). There is no specific procedure on how to "import your old code" since development on eclipse is focused on writing the .m files directly. My suggestion is that you should first understand how Mathematica projects work in eclipse, then figure out how to "move" things over. You could simply create a Mathematica Basic Project, copy your auto-created .m files into the project, create a new notebook in that project, and copy your test cases from your old notebook to the new notebook. At that point, you can discard your old notebook. When you do a right-click on the new notebook, Mathematica > Run..., it will load all the .m files from that project. If you want to take it a step further, you can convert your tests from your old notebook to specific unit tests in Eclipse. This is a manual process that would involve lots of cutting and pasting because the unit tests must follow a specific syntax. > I have many more questions, especially with respect to revision > control, package generation and loading, handling of the notebook > cache, etc, but I will download the trial first... > > Do the Workbench developers feel like dropping any knowledge? Ask away in the newsgroup. I'm not personally involved with the development of the workbench, but we can all respond to users' questions/comments. :-) There's no trial version of the workbench. Eclipse itself is free and so is the Wolfram plugin to Eclipse. The Wolfram Workbench (a pre-packaged Eclipse with the Wolfram plugin + icons) is also free. At this time, it's only available to premier subscribers since it's still pre-release version. Igor -- Igor C. Antonio Wolfram Research, Inc. http://www.wolfram.com To email me personally, remove the dash.
- References:
- Wolfram Workbench
- From: "Mark Westwood" <markc.westwood@gmail.com>
- Re: Wolfram Workbench
- From: "Chris Chiasson" <chris@chiasson.name>
- Wolfram Workbench