jwcolby
jwcolby at colbyconsulting.com
Fri Jul 16 12:12:56 CDT 2010
Michael, Very cool and I want to use this. John W. Colby www.ColbyConsulting.com Michael Bahr wrote: > John here is an example from another SCM that we were using where I made a > program so users can use the same format to populate code files with the > header information: > > ###################################################################### > # Start Header Info > ###################################################################### > #$Workfile: $ > #$Revision: $ > #$Author: $ > #$Date: $ > #$Modtime: $ > ###################################################################### > # End Header Info > ###################################################################### > 0000:://Hello World--This is test of the Emergency Broadcast System!! > 0001:://PROC WAIT3000 > 0002:://\\ > 0003::LOOP:: > 0004::WAIT 3000 > > and this is the result > > ###################################################################### > # Start Header Info > ###################################################################### > #$Workfile: WAIT300.SCP $ > #$Revision: 1.0 $ > #$Author: mbahr $ > #$Date: Mar 24 2006 14:34:56 $ > #$Modtime: Mar 24 2006 13:23:52 $ > ###################################################################### > # End Header Info > ###################################################################### > 0000:://Hello World--This is test of the Emergency Broadcast System!! > 0001:://PROC WAIT3000 > 0002:://\\ > 0003::LOOP:: > 0004::WAIT 3000 > > in this example the keywords are encased in dollar-signs and get modified > everytime the file is checked-in. Now you have traceability with files > either doing a GET or Check-out. > > If you have not found yet then it is buried somewhere. > > Mike... > >> >What needs to be done is you must add a header in the form of comments >> that use the keywords or >> artifacts I believe, then SVN will use the keywords with the appropriate >> information. >> >> Ohhhh. I haven't found anything that says that. >> >> John W. Colby >> www.ColbyConsulting.com >> >> >> Michael Bahr wrote: >>> Yes you are correct, SVN does not implicitely add anything. What needs >>> to >>> be done is you must add a header in the form of comments that use the >>> keywords or artifacts I beleive, then SVN will use the keywords with the >>> appropriate information. >>> >>> Mike... >>> >>>> Michael, >>>> >>>> I mostly understand all of that. I use SVN with Visual Studio here in >>>> my >>>> office. I have a >>>> programmer that comes in and works and I do C# coding. We both check >>>> out >>>> / in pretty much daily. I >>>> have had to reconcile differences a couple of times. I don't do >>>> anything >>>> fancy with it though. >>>> >>>> AFAICT SVN does not add anything into the file itself - author, date >>>> etc. >>>> Given that it could >>>> corrupt a file structure (word or excel) I doubt that it can do that. >>>> I >>>> think the header info is >>>> kept in the repository somewhere. We shall see. >>>> >>>> I exported a FE to text files and checked them in, then checked them >>>> back >>>> out. I then made a simple >>>> edit to the code behind form for one form and reexported / checked in >>>> that >>>> one form. Basically I >>>> did exactly that, export over the top of an existing text file. >>>> >>>> SVN tags files on disk with icons and the icon changed from a green >>>> check >>>> to a red ! until I checked >>>> the change in. Unfortunately I it appears that Tortoise doesn't use >>>> the >>>> same property to find the >>>> repository for the diff application because when i tried to use that >>>> widget it says it can't find >>>> the repository. Sigh. >>>> >>>> But the concept seems sound. Really the hardest part seems like it >>>> would >>>> be getting the file back >>>> in to the MDB. That is probably not difficult, it would just have to >>>> be >>>> deleted and then the text >>>> file sucked in and saved. >>>> >>>> John W. Colby >>>> www.ColbyConsulting.com >>>> >>>> >>>> Michael Bahr wrote: >>>>> John, SVN works the same way as VSS, ClearCase, PVCS, etc. You add >>>>> your >>>>> files or project the first time, this is the base line revision. This >>>>> revision is annotated like 0.001 or something like that. >>>>> >>>>> Then you do a "Checkout" which should add a marker in SVN to indicate >>>>> that >>>>> those files have been checked-out. By checking out you are allowed to >>>>> "Check-In" the files. This does not overwrite or destroy the previous >>>>> file, just adds another copy to the revisioning process. This >>>>> increases >>>>> the revision by one like 0.002. Now you can do a differential between >>>>> 0.001 and 0.002. BTW, differentials only work on text file, not >>>>> binary. >>>>> If you do a "Get" and you make changes you can not do a check-in, you >>>>> must >>>>> do a "Check-Out" first. So over time you will have many revisions of >>>>> files and at some point you will want to "release" the files or >>>>> project >>>>> as >>>>> Version 1.000. Eveything before version 1.000 would be your >>>>> developement >>>>> work for example. Now you have version control. >>>>> >>>>> Then you continue with your developement with revisions (check-out, >>>>> make >>>>> changes, check-in) until you do another release, for example Ver >>>>> 1.100. >>>>> Here is where you can do differentials between the various releases. >>>>> >>>>> Now SVN has (should) features that you would really want like header >>>>> information in the files. This is important information like date, >>>>> time, >>>>> author, description, revision number, and possibly version number. >>>>> SVN >>>>> probrably uses some keywords to that you must place in the header the >>>>> first time (the baseline) and SVN will automatically update these >>>>> keywords >>>>> everytime you do a check-in. The most important items in the header >>>>> are >>>>> the date and revision. So if you were to make a hard copy or have >>>>> several >>>>> soft copies hanging around of the source code you can easily identify >>>>> the >>>>> revision, otherwise it would be very difficult. Doing things your way >>>>> of >>>>> sucking out the objects each time I think would eliminate the header >>>>> information thus rendering the check-out files difficult to track. >>>>> >>>>> So I hope you understand the process of revision and version. Your >>>>> can >>>>> be >>>>> be done but I think it is a lot of work dealing with Access objects. >>>>> Using SVN for .Net projects should be very easy to use and very >>>>> beneficial. >>>>> >>>>> On another note, I would suggest that you come up with a plan for >>>>> version >>>>> control, or SCM (Software Configuration Management). This is for you >>>>> and >>>>> your customers. >>>>> >>>>> Mike... >>>>> >>>>>> Yea, but I want source control. >>>>>> >>>>>> John W. Colby >>>>>> www.ColbyConsulting.com >>>>>> >>>>>> >>>>>> Dan Waters wrote: >>>>>>> <I am currently faced with a set of FEs that I need to see the >>>>>>> differentials >>>>>>> of.> >>>>>>> >>>>>>> FMS has a utility you can purchase called Access Detective. It's >>>>>>> about >>>>>>> a >>>>>>> couple hundred dollars, and you'd quickly get that back in time >>>>>>> saved. >>>>>>> >>>>>>> HTH, >>>>>>> Dan >>>>>>> >>>>>>> >>>>>>> Charlotte Foust wrote: >>>>>>>> I guess the success will depend on what you hope to gain. Without >>>>>>>> a >>>>>>>> checkin/out from within Access, any changes to a database will have >>>>>>>> to >>>>>>>> be manually exported and then imported to subversion, right? I'm >>>>>>>> not >>>>>>>> familiar with the product, so how do you keep versions of the >>>>>>>> objects >>>>>>>> rather than overwriting, or does it matter? >>>>>>>> >>>>>>>> Charlotte >>>>>>>> >>>>>>>> On Tue, Jul 13, 2010 at 4:21 PM, jwcolby >>>>>>>> <jwcolby at colbyconsulting.com> >>>>>>> wrote: >>>>>>>>> I am trying a workaround. Max has written code that exports all >>>>>>>>> the >>>>>>> objects in an access database >>>>>>>>> to text files. It seems a short step to using that to get them >>>>>>>>> into >>>>>>> subversion. >>>>>>>>> John W. Colby >>>>>>>>> www.ColbyConsulting.com >>>>>>>>> >>>>>>>>> >>>>>>>>> Charlotte Foust wrote: >>>>>>>>>> How would you move items in an out of Access? Source Safe uses >>>>>>>>>> an >>>>>>>>>> Access add-in to handle source control at the object level. Is >>>>>>>>>> there >>>>>>>>>> something similar for subversion, or are you trying a workaround? >>>>>>>>>> >>>>>>>>>> Charlotte >>>>>>>>>> >>>>>>>>>> On Tue, Jul 13, 2010 at 1:35 PM, jwcolby >>>>>>>>>> <jwcolby at colbyconsulting.com> >>>>>>> wrote: >>>>>>>>>>> I am wondering whether we could put together subversion and >>>>>>>>>>> Access >>>>>>>>>>> to >>>>>>> do source control. >>>>>>>>>>> Max's eatbloat will export and import most things Access into a >>>>>>> directory structure. Subversion >>>>>>>>>>> could be used to move that stuff into version control. Once in >>>>>>> subversion, we might be able to do >>>>>>>>>>> differencing etc. >>>>>>>>>>> >>>>>>>>>>> Whaddayathink? >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> -- >>>>>>>>>>> John W. Colby >>>>>>>>>>> www.ColbyConsulting.com >>>>>>>>>>> -- >>>>>>>>>>> AccessD mailing list >>>>>>>>>>> AccessD at databaseadvisors.com >>>>>>>>>>> http://databaseadvisors.com/mailman/listinfo/accessd >>>>>>>>>>> Website: http://www.databaseadvisors.com >>>>>>>>>>> >>>>>>>>> -- >>>>>>>>> AccessD mailing list >>>>>>>>> AccessD at databaseadvisors.com >>>>>>>>> http://databaseadvisors.com/mailman/listinfo/accessd >>>>>>>>> Website: http://www.databaseadvisors.com >>>>>>>>> >>>>>> -- >>>>>> AccessD mailing list >>>>>> AccessD at databaseadvisors.com >>>>>> http://databaseadvisors.com/mailman/listinfo/accessd >>>>>> Website: http://www.databaseadvisors.com >>>>>> >>>> -- >>>> AccessD mailing list >>>> AccessD at databaseadvisors.com >>>> http://databaseadvisors.com/mailman/listinfo/accessd >>>> Website: http://www.databaseadvisors.com >>>> >>> >> -- >> AccessD mailing list >> AccessD at databaseadvisors.com >> http://databaseadvisors.com/mailman/listinfo/accessd >> Website: http://www.databaseadvisors.com >> > >