Betreff: USVN versus UD6
Von: "Ulrich Merkel <UlrichMerkel@web.de>
An: "Petr ©imek" <firstname.lastname@example.org>
Datum: 20.11.09 08:36:44
To use SVN, we need some flat files with a fine granularity.
Because the repository information is held in a "real" database, we have to transfer the info to flat files.
- we have to create export files on our own (export)
- we have to load the information back to the database (import)
UD6 from march-hare:
The UD6 is a database driver. The repository information is stored directly as XML files.
This causes longer waits if you browse through "all components" because they have to open a lot of files.
They do not skip the step as you say, these XML files ARE the database.
From your arguments, I would recommend the UD6 because:
- it works dirtectly inside of the idf with no "additional" steps
- it exists over years (so shuld be minimal bug-risk)
Because this makes sense, I have no problem, if you decide to use UD6 and not spend time on the USVN any more.
P.S. Unfortunately, Uniface does not provide any plug-in support or hook mechanism
where you can append your own functionality to the idf internal processing.
> -----Ursprüngliche Nachricht-----
> Von: "Petr ©imek" <email@example.com>
> Gesendet: 19.11.09 19:58:43
> An: Ulrich Merkel <UlrichMerkel@web.de>
> Betreff: Re: Re: Re: Uniface svn integration (extended version 2)
> Hi Uli
> I went already through the documents you sent me. However I did not import the source code yet. Please correct me if I'm wrong.
> I assume it all works in this way - every developer has it's own Uniface repository (with sources in database), but inside of idf there is additional menu allowing developer to load component from the svn repository on the file system. Then the component is inside the Uniface database and once the coding and perhaps testing is ready, component can be loaded (again with the help of additional menu) back to the svn, so others can work on newer version.
> If I compare it with ud6 - they do a bit more - they completely skip the phase of storing the component in the database. Every modification of any source code (component, global object) can directly be mapped to modification of a file on the file system.
> I definitely do not want to say I'm not going to continue over here on this svn Uniface integration. Maybe this is the only option left for the team if my manager will not find budget for 50 ud6 licenses ...
> But anyway - do you know ho to relink Uniface and remap some basic Uniface functions (every data store, every data load) to your own functions? In such function you should always get pointer to internal Uniface data stream. (maybe trx). What you finally do with this stream is up to you - by default Uniface stores it into defined DICT tables.
> Hope there was a bit of sense in these few sentences above
-- Ulrich Merkel --
|Only registered users can see links on this forum!|
Register or Login on forum!