Hi again, with U5 abandoned, I have shifted my porting efforts over to the U 4.7.2 branch. I was really hoping to get something done by the end of Codegarden 12, but there was just way too much to do...
However, what I can say now is that I am pretty confident that we will be able to do a good port over to Linux. I have been working on this, and most of the code changes are pretty minor (or, looking back now seem minor) - but there are a fair number of them. In a nutshell, I have a got a partially working version now, and am working towards, let's say an alpha release. It is probably too early to share a version of the code b/c there are still a lot of errors, and it may end up frustrating people.
The good news: In the past two years, since I started on looking at porting Umbraco over to Linux, mono has developed by leaps and bounds - this is now making a successful port possible.
The bad news: Linux is case sensitive. There are a lot of minor changes to be done here, and some discipline to be had in the future in casing conventions. Mono does have the MONO_IOMAP=all option - but this is too slow for any production system. Hence, all casing needs to be vetted for Linux. Mono offers some inspection tools but it is still work, lots of it, and requires consistency for future releases. I had hoped to have a univeral release of Umbraco but we will probably end up with two branches - one for Windows and another for Linux.
The small print: I have had success compiling Umbraco with mono 2.11.2 (using a parallel development environment), and forcing all assemblies to .NET 4.0. I have not had any success under .NET 3.5 - indeed with 3.5 you get a lot of resource not found errors (that are not case related).
The even smaller print: Other than case errors, I am at present dealing with some MYSQL errors, and other unspecified oddities (E.g. as you can get it in London: "The train is delayed, and the cause is unknown, and under investigation")
But the overall outlook remains optimistic. I will check in again.
Friday, 29 June 2012
Thursday, 14 June 2012
Umbraco 5 on Linux
With the passing of version 5, I will no longer be spending anytime on getting Umbraco 5 to compile under Linux. I will be shifting my efforts to Umbraco 4 - but I am not making any promises.
Umbraco 5 R.I.P. - Editorial
There were already some 'Codeplex smells'... And my own efforts at porting Umbraco v5 over to Linux kept pointing to the extremely complex, perhaps overly complex, Hive.
Yes, the Umbraco Hive concept, an all encompassing aggregating data source that was going to be extremely fast, easy-to-use, and that was to replace the original XML based storage concept, proved in the end impossible to implement.
As the Hive was central to everything in version 5, the failure of the Hive, in my opinion, condemned the application to early extinction.
Umbraco 5 had been in the pipeline for about 2 years. It joins the long list of prominent Edsels, which the computing industry has produced... It is surprising that even in this Agile age, such a big mistake went unchecked, grew, and forced the project to be abandoned.
Sometimes it is difficult to let go - especially when we are passionate about something. In this case, both the Umbraco community and the Umbraco team had placed such high hopes on version 5 that it took quite a while before anybody said, "It is time to stop."
While the passing of version 5, and the circumstances of its passing form a sad note, it is also valuable to notice that the Umbraco team has acknowledged that version 5 was essentially a "dud." And they have done the right thing.
Furthermore, going back to version 4 is not such a bad thing, as version 4 is a very strong product, and will even give us XSLT back.
Yes, the Umbraco Hive concept, an all encompassing aggregating data source that was going to be extremely fast, easy-to-use, and that was to replace the original XML based storage concept, proved in the end impossible to implement.
As the Hive was central to everything in version 5, the failure of the Hive, in my opinion, condemned the application to early extinction.
Umbraco 5 had been in the pipeline for about 2 years. It joins the long list of prominent Edsels, which the computing industry has produced... It is surprising that even in this Agile age, such a big mistake went unchecked, grew, and forced the project to be abandoned.
Sometimes it is difficult to let go - especially when we are passionate about something. In this case, both the Umbraco community and the Umbraco team had placed such high hopes on version 5 that it took quite a while before anybody said, "It is time to stop."
While the passing of version 5, and the circumstances of its passing form a sad note, it is also valuable to notice that the Umbraco team has acknowledged that version 5 was essentially a "dud." And they have done the right thing.
Furthermore, going back to version 4 is not such a bad thing, as version 4 is a very strong product, and will even give us XSLT back.
Breaking News from Codegarden 2012
It is now official! Umbraco 5 has been iced.
Umbraco 5.2 will be the last release of the MVC version of the successful open source CMS, bringing to an end Umbraco's foray into the MVC realm. Further development on version 5 of Umbraco will be terminated.
Instead, development efforts will henceforth focus on the well-established Webforms based Umbraco 4, which is currently as version 4.7.2.
A roadmap for development on version 4 was shown, and this roadmap also includes incorporation, over time, of certain v5 elements back into the v4 branch.
However, whether we will at some point see an MVC port of version 4 remains to be seen.
Subscribe to:
Posts (Atom)