Re: MG++ build collection fails

From Puneet Pawaia
DateThu, 09 Jan 2003 10:09:47 +0530
Subject Re: MG++ build collection fails
In-Reply-To (3E00F247-4899FEF1-cs-waikato-ac-nz)
Hi All,

The configuration that I am using is as follows
Windows XP Home SP1 (administrative rights login), Greenstone 2.38, Active
Perl 5.6.1 build 631, IE6 SP1

I attempted to build the demo collection using mg++ with the configuration
provided with the compilation after adding the buildtype information and
changing the indexes information

public true
buildtype mgpp <-- changed
indexes text <-- changed

;indexes section:text section:Title document:text <-- commented out
;defaultindex section:text <-- commented out

plugin GAPlug
plugin HTMLPlug -description_tags -input_encoding iso_8859_1
plugin ArcPlug
plugin RecPlug -use_metadata_files

classify Hierarchy -hfile sub.txt -metadata Subject -sort Title
classify Hierarchy -hfile AZList.txt -metadata AZList -sort Title
-buttonname Title -hlist_at_top
classify Hierarchy -hfile org.txt -metadata Organization -sort Title
#classify Hierarchy -hfile keyword.txt -metadata Keyword -sort Title
-buttonname Howto
classify List -metadata Keyword -buttonname Howto

format SearchVList
"<td valign=top>[link][icon][/link]</td><td>{If}{[parent(All':
'):Title],[parent(All': '):Title]:}[link][Title][/link]</td>"
format VList
"<td valign=top>[link][icon][/link]</td><td
valign=top>[highlight]{Or}{[Title],Untitled}[/highlight]<i><small>{If}{[Date],<br>publication date: [Date]}{If}{[NumPages],<br>no. of pages: [NumPages]}{If}{[Source],<br>source ref: [Source]}</small></i></td>"

format CL4VList "<br>[link][Keyword][/link]"

format DocumentText "<h3>[Title]</h3>\n\n<p>[Text]"
format DocumentImages true
format DocumentButtons "Expand Text|Expand Contents|Detach|Highlight"
format HelpBookDocs true

collectionmeta collectionname "greenstone demo"
collectionmeta collectionextra "This is a demonstration collection for
the Greenstone digital library software. It contains a small subset (11
books) of the Humanity Development Library"
collectionmeta collectionextra [l=fr] "C'est une collection pour
d??monstration du logiciel Greenstone. Elle contient une petite partie du
projet de biblioth??ques humanitaires et de d??veloppement (11 livres)."
collectionmeta iconcollectionsmall
collectionmeta iconcollection "/gsdl/collect/demo/images/demo.gif"
collectionmeta .section:Title "section titles"
collectionmeta .section:Title [l=fr] "titres des sections"
collectionmeta .document:text "entire books"
collectionmeta .document:text [l=fr] "livres entiers"
collectionmeta .section:text "chapters"
collectionmeta .section:text [l=fr] "chapitres"

This compilation resulted in the following errors

The collection could not be built ( failed).
The build log contains the following information:
GAPLug: processing HASHf6dc.dirdoc.xml
GAPLug: processing HASH0163.dirdoc.xml
GAPLug: processing HASHebf5.dirdoc.xml
GAPLug: processing HASH0182.dirdoc.xml
*** creating auxiliary files
build: ERROR: failed
Please restart the collector and try again.

I would like to bring to your notice here that when I used my own sample of
500 files for building a collection using mg++, I got 4 file errors at that
time too. When I reduced the number of files to 100, I still got 4 file
errors and when I further reduced the files to 10, again 4 file errors were
reported. Could this be significant ?


At 03:40 AM 12/19/2002, you wrote:
>The plugin error thing doesn't sound like an mgpp problem - plugins are used
>before you get to mg or mgpp - have you looked at the log file for the mg
>build? this may contain the same error, in which case this is not the reason
>your build failed.
>were there any other messages about mgpp stuff? did you use the collector or
> how did you know that the build had failed? - error messages? or
>did you try to view the collection and it wasn't there? did you get
>anything in
>the building or index directories? are the archives the same for the mgpp and
>mg builds - they should be.
>have you managed to build anything with mgpp before on the same computer? eg
>the demo collection? perhaps its not your collection, but rather mgpp doesn't
>work for some reason.
>If you send me your collect.cfg I can check it just in case that is the
>Stefan Boddie wrote:
> > Hi,
> >
> > Sorry for the delay in responding to this. Supposing you've altered your
> > collect.cfg file correctly (and it sounds like you have) then the
> collection
> > should build equally well with mgpp as with mg. A couple of possible
> causes:
> >
> > 1. I think there have been some minor bug fixes to mgpp since gsdl-2.38 was
> > released. I can't remember the details but maybe Kath (our mgpp guru) can
> > let us know if any of these bugs might cause this problem. Kath: Remember
> > the changes we made when building the protemix collection? That was since
> > gsdl-2.38 was released, right?
> >
> > 2. I haven't tested it much yet but I have some vague suspicions that the
> > Greenstone building code has some problems with perl 5.8. Can you let me
> > know what version of perl you have? Does dropping back to perl 5.6.1 solve
> > the problem?
> >
> > regards,
> > Stefan.
> >
> > ----- Original Message -----
> > From: "Puneet Pawaia" <>
> > To: <>
> > Sent: Friday, December 13, 2002 5:32 PM
> > Subject: MG++ build collection fails
> >
> > > gsdl 2.38, Windows XP Home
> > >
> > > Hi All
> > >
> > > I am attempting to build a collection from a set of 500 html files. These
> > > files are very basic in nature in that they contain the following
> > elements.
> > > A top level heading tagged as H1, a second level heading tagged as h2, a
> > > third level heading tagged as h3 and some lines of text.
> > >
> > > When I build the collection using mg, the collection is built
> > successfully.
> > >
> > > However, when I add a 'buildtype mgpp' line and modify the indexes and
> > > defaultindex lines, the build fails. The log reports that there are 4
> > > doc.xml files in different subdirectories that could not be processed by
> > > any plug-in. When I attempt to go to the referenced subdirs in the
> > > tempbuild directory, these directories are not there.
> > >
> > > How can I resolve these errors.
> > >
> > > TIA
> > >
> > > Regards
> > > Puneet
> > >
> > >
> > >