Re: [greenstone-users] Unknownplug problem

From John R. McPherson
DateMon, 14 Mar 2005 09:54:40 +1300
Subject Re: [greenstone-users] Unknownplug problem
In-Reply-To (ISP-GOSSp8u16sa5LMg0008ad81-isp-go-FPT-NET)
On Fri, 2005-03-11 at 13:27 +0700, Tran wrote:

> I'm using Unknownplug for importing *.djvu files. Using GLI I've
> entered two arguments:
> assoc_field: djvu
> file_format: *.djvu
>
> The strange thing has happended when I tried to build my collection:
> From four *.djvu files that I wanted to add into my collection only
> one file was recognised and processed by UnknownPlug:
> The file DownloadDewdaney-golovolom.djvu was not recognised
> by any plugin.
> The file DownloadDifferential equations for engineers and
> scientists.djvu was not recognised by any plugin.
> The file DownloadFishman,_Monte_Carlo.djvu was not recognised
> by any plugin.
> The file Download
> Hoffman;_Numerical_Methods_for_Engineers&Scientists,2001.djvu is
> being processed by UnknownPlug.
> The file DownloadPrincilpes of data Mining.djvu was not
> recognised by any plugin.
> The file Download
> Thompson_Computing_for_scientists_and_engineers.djvu was not
> recognised by any plugin.
>
> Using Unknownplug without any argument did help, too. The symptom was
> the the same.

These filenames work for me. I think the problem is your arguments to
UnknownPlug.
The file_format option is used for metadata only. Eg if you put
-file_format DJVU, then the greenstone documents will be given
FileFormat metadata with "DJVU" in them.

You want the -process_exp option to tell UnknownPlug which files it
should process: "-process_exp djvu" to process any files that have
"djvu" in the filename. Note that I think that this is case sensitive by
default. Try "-process_exp '(:i)djvu$' to match all filenames that end
with djvu, case-insensitively.

> I've reported that Unknownplug can recognised and processed just
> one of my djvu files. It seems that it's a backward compatibility
> problem. Means that Greenstone is designing for early version of
> Windows - therefore it recognises well files with less than 8
> characters in their name.

That definitely isn't the case. Greenstone is primarily developed on
linux and windows xp, and short filename problems (for windows 3.1) are
only an issue for the files distributed with greenstone. But I think
we've dropped support for win 3.1 now, since very few people are using
greenstone on it. None of that should affect greenstone's importing of
files though.

Can you email your collect.cfg file and an output log of it failing to
get your files when they have the original names?

John