Installing Software

Hello,

from what i’ve discovered so far, applications are not a single binary but instead a directory starting with a “!”. This directory in turn contains all the necessary files to run the application. Thats correct, or?

I’m asking cause i try to install a Application (!PDF) to a freshly installed RISC OS 5.21 in RPCemu – but it doesn’t works the way i expected it. What i’ve done:

1. Downloaded the ZIP file on the Host and extracted the !PDF Folder to a location inside the HostFS
2. Tried to run the PDF application by left double clicking the !PDF Folder.

But instead of running the application, the Editor opens up and displays the content of the !Run file (HostFS:$.Daten.!PDF.!Run). What did i do wrong?

Yours,
Egon

from what i’ve discovered so far, applications are not a single binary but instead a directory starting with a “!”. This directory in turn contains all the necessary files to run the application. Thats correct, or?

Correct. When you first see the special directory (with the !) the !Boot file is automatically run and any actions in it are carried out. When you click on the directory the !Run file is run. What the run action is depends.

I’m asking cause i try to install a Application (!PDF) to a freshly installed RISC OS 5.21 in RPCemu – but it doesn’t works the way i expected it. What i’ve done:
1. Downloaded the ZIP file on the Host and extracted the !PDF Folder to a location inside the HostFS

Did you leave the files in the zip and the start RPCEmu, then run SparkFS (read only version in the disc build) and drag the file to the SparkFS icon on the iconbar?
Or did you open the zip via Windows? Opening via Windows is wrong.

2. Tried to run the PDF application by left double clicking the !PDF Folder.
But instead of running the application, the Editor opens up and displays the content of the Readme file (HostFS:$.Daten.Run). What did i do wrong?

See above – at a guess you extracted the file set using windows and thereby lost the filetype information.

Open the zip file using the RISC OS tool SparkFS and you will see the files in the ! prefixed directory have different file icons. Click the middle mouse button, or third button or scrollwheeel depending on the mouse over one of those files and then follow the menu option for file > then down to “Set type” and sideways >

it will be telling you the current filetype of that file. If the filetype is “Obey” then the system Obeys the instructions, if the file type is Text the system opens it using a text editor.
Since the default file type is Text, opening the zip file in windows and destroying the filetype meta data will cause the file to have the default type and open in a text editor.

Just a guess on the origin of your problem…

I think i can partially answer my own question after playing around a little more:

1. When installing SparkPlug (self extracting version) and then using it to extract the PDF application, it works.
2. Whats the difference between uncompressing it in RISC OS directly using SparkPlug and uncompressing it on the Host system and putting the files in the HostFS?

Yours,
Egon

Whats the difference between uncompressing it in RISC OS directly using SparkPlug and uncompressing it on the Host system and putting the files in the HostFS?

See previous mail. Basically Windows doesn’t understand the meta data element of the file and destroys it. To retain it in the hostFS system HostFS will append the three character representation of the filetype meta data as a ,xxx suffix e.g. ,feb for the obey files

Hi Egon,

I’m not sure. Maybe you can post your questions on www.arcsite.de .
If not, sorry. ;-)

Hello Steve,

thanks for your reply. It’s now clear why it didn’t worked. I didn’t knew that files in RISC OS not only comprise of data and name but also have an associated type which isn’t stored within the plain file but in the filesystem directly.

Yours,
Egon

I didn’t knew that files in RISC OS not only comprise of data and name but also have an associated type which isn’t stored within the plain file but in the filesystem directly.

Not unusual. Many people don’t know that the Windows habit of relying on the name suffix to denote a filetype is the odd man out. Because Windows systems are so common in general populace use people assume everything else works the same way.
Linux systems for example will happily execute a file named myfile.pdf or myfile.txt if the file is really an executable item. In short the name suffix is something for humans to read and of no import to most OS implementations.

Imagine a file as a box. The content is the data istself. The labels on the box give some information about the content. Thats called metadata. Data about the data. Like the filename, filetype, date created, date modified, access rights.
Different file system store metadata in different manner, not all compatible.