gmane.text.refdb.general

Re: docbook5, getting started

Markus Hoenicka — 2012-09-12T14:15:33

Tim Arnold <jtim.arnold< at >gmail.com> was heard to say:


I see. This is caused by a limitation of  
/usr/local/share/refdb/xsl/citations.xsl. This stylesheet is used to  
extract the citations from the source document and write them into the  
foo.id.xml file. For reasons that I don't quite understand right now,  
the script limits itself to processing book and article elements. As a  
short term fix you could simply extend the regexp in line 39 like this  
(untested!):

         <xsl:template match="article|d:article|chapter|d:chapter">

I'll try to figure out if there was some obscure reason for this  
limitation. If not I'll adapt the stylesheet to the current DocBook V5  
top-level elements.

regards,
Markus

Re: docbook5, getting started

2012-09-11T21:07:09

 > Hi Markus,
 > Here are some notes:
 > Using refdbxp 0.9.9-1 built from svn revision 531 markus< at >mhoenicka.de
 > 
 > (1) invalid xml generated
 > This citation:
 >     <citation role="REFDB">bart_m66</citation>
 > 
 > After the command
 > refdbxp -t db50x < test1.xml > test1_full.xml
 > becomes (note the non-terminated xref)
 >     <citation role="REFDB"><xref linkend="IDbart_m66-X"></citation>
 > 

This is an odd inconsistency. As refdbxp need not be aware of the
DocBook or TEI version you're using, you can only choose between
DocBook SGML, DocBook XML, and TEI XML. The first one is the default
if no type or an incorrect type is specified. Try "-t db31x" instead
and you should get valid XML. See the man page for further
details. The default stems from the days of yore when XML was all
fresh and new and no one used it because the tools weren't as good as
they are now. XML should probably be the default these days, but I'm
always careful not to break old behaviour.

 > 
 > (2) documentation typo
 > Page 102

Re: docbook5, getting started

2012-09-10T19:12:58

 > failed to access style data

Hi Tim,

this usually means that you did not add that particular style to the
database. The style data have to be available to the application
server refdbd rather than to any of the clients. You'll have to run a
command along the lines of:

refdba -C addstyle APA.xml

although this should have been taken care of during the initial setup
of RefDB with refdb-init. You can list the available styles with the
command:

refdbc -C liststyle

regards,
Markus

docbook5, getting started

2012-09-08T14:17:34

 > I'll follow up on this on the list, but could you please send me your
 > input file test.xml and test.id.xml (if that gets proper created by
 > runbib) for further analysis off-list?

Thanks for providing the input data and the additional debug
information. I think the problem here is that you do not use the
proper citation syntax for raw bibliographies. The citation syntax
that you used is the short notation which would require a
preprocessing with refdbxp before you run runbib, using a cooked
bibliography. If you want to process the raw bibliography with the
stock DocBook stylesheets, you'd have to use the standard DocBook
citation syntax, like this:

<citation role="REFDB"><xref linkend="kuma_v92"/></citation>

This syntax, and some hints on further processing these documents, is
explained in some more detail in the manual (chapter 10, section
"create bibliographies",
http://refdb.sourceforge.net/manual/ch10s03.html).

best regards,
Markus

docbook5, getting started

2012-09-08T07:57:41

 > In my docbook test.xml file I have a couple of citations like this:
 > <citation role="REFDB">kuma_v92</citation>.

Hi,

I'll follow up on this on the list, but could you please send me your
input file test.xml and test.id.xml (if that gets proper created by
runbib) for further analysis off-list?

regards,
Markus

docbook5, getting started

Tim Arnold — 2012-09-07T14:54:31

hi, I must be doing something wrong since I can't get to the first
step of creating a DocBook5.0 raw bibliography.
I converted my bibtex database to ris format and I can use refdba, so
my installation is okay I think.

In my docbook test.xml file I have a couple of citations like this:
<citation role="REFDB">kuma_v92</citation>.

runbib -d mybib -r -t db50x test.xml
I get
Variable $SGML_CATALOG_FILES not set
Parse error at line 4 column 2:
syntax error

The terminal where the refdbd instance is running (refdbd -s -e 0 -l
7) says this when I exec the runbib in another terminal:
---------------------
sqlite3
/usr/local/var/lib/refdb/db

mybib
SELECT meta_app,meta_type,meta_dbversion FROM t_meta WHERE meta_type='risx'
connected to database server using database:
mybib
no character encoding conversion required
output encoding is:
UTF-8
XML parse error

child exited with code 1
server waiting n_max_fd=4
---------------------


So I try the step-by-step way:
xsltproc --xinclude /usr/local/share/refdb/xsl/citations

Re: Fatal error: Call-time pass-by-reference has been removed in /usr/local/share/refdb/www/index.php on line 1531

Markus Hoenicka — 2012-06-06T09:20:20

Hi,

as far as I can see this problem is caused by a change between PHP 5.3  
and 5.4 regarding the mechanism how variables are passed by reference.  
I do not have a PHP interface up and running at this time, so the  
following is untested. Could you please check whether the following  
change in /usr/local/share/refdb/www/index.phpfixes the problem? The  
line in question is 1786 in my sources, although you mention 1531  
instead.

- if (!yaz_ccl_parse($id[$i], $term, &$cclresult)) {
+ if (!yaz_ccl_parse($id[$i], $term, $cclresult)) {

(note the missing ampersand). According to what I read about this  
issue on the web, PHP should be able to figure out by itself how to  
pass the variable appropriately, and the syntax using the ampersand  
has been removed in PHP 5.4.

regards,
Markus


"O. Hartmann" <ohartman< at >zedat.fu-berlin.de> was heard to say:

Fatal error: Call-time pass-by-reference has been removed in /usr/local/share/refdb/www/index.php on line 1531

O. Hartmann — 2012-06-06T07:19:18

A couple of days ago all of our FreeBSD boxes where forced to update
their ports recompiling all PNG dependencies.

After those updates, on ALL of our boxes providing the RefDB
webinterface, this webinterface, PHP5, has been gone.

I put one of the servers via php.ini into "development" mode reporting
errors, since the servers in questions do show a white page with nothing
otherwise.

I get this:

Fatal error: Call-time pass-by-reference has been removed in
/usr/local/share/refdb/www/index.php on line 1531

The line in question seems to refer to YAZ:

if (!yaz_ccl_parse($id[$i], $term, &$cclresult)) {
... }

I tried to recompile refdb via
portmaster -f refdb

as well as PHP

portmaster -f php5-5

but there is no solution doing it that way.

Since RefDB worked fine before, I guess this is a unconvenience brought
in by the port's updating orgy.

Regards
Oliver

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the

Re: DOI and edition

Stephan Gatzka — 2012-02-03T06:20:34

Hello!

Thanks a lot. Now it works perfectly for me.

Regards,

Stephan


------------------------------------------------------------------------------
Try before you buy = See our experts in action!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-dev2_______________________________________________
Refdb-users mailing list
Refdb-users< at >lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/refdb-users

Re: DOI and edition

2012-02-01T00:09:27

Hi,

Stephan Gatzka writes:
 > So for me it seems that the RIS VL tag will be found in "refdb_edition" 
 > and not in "refdb_volume".
 > 

thanks for your analysis which is pretty much correct. I've checked in
a fix (svn revision 755). This version returns your test dataset as
DocBook 5.0 including the edition information.

best regards,
Markus

Re: DOI and edition

Stephan Gatzka — 2012-01-31T20:28:35

Hi!


Thanks for your hint with the trailing space, but again, I'm not really 
a RIS specialist...


I just ran the code through the debugger.

In backend-db31.c:879 is the call to get_refdb_volume which itself calls

result = dbi_result_get_string(dbires, "refdb_volume");

This call always returns NULL.

In comparison, in backend-risx.c:864 is also the call to 
get_refdb_volume. But if this call fails, it calls get_refdb_edition in 
the else clause. This call succeeds.

So for me it seems that the RIS VL tag will be found in "refdb_edition" 
and not in "refdb_volume".

Hope this helps.

Regards,

Stephan

------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d____________________________

Re: DOI and edition

2012-01-31T00:17:37

Hi,

Stephan Gatzka writes:
 > M3  - DOI:10.1007/3-540-27522-3
 > ER  -

Unless your mail agent unvoluntarily edited your input file, the "ER"
tag lacks a trailing space. The RIS spec is pretty clear about the
fact that each tag ends with two spaces, a dash, and another
space. The parser does not recognize "ER  -" as a tag and reads it as
another line of the preceding M3 field. This does not quite explain why
the "ER  -" string is duplicated though, but you should try again
using a correctly terminated RIS entry, i.e. "ER  - ".

 > What I'm missing is the edition tag and I also have these ER  - intermixed.
 > 

I see that the edition information is missing from the DocBook
output. I'll look into this as soon as time permits.

best regards,
Markus

Re: DOI and edition

Stephan Gatzka — 2012-01-28T20:52:51

Hello!

I've found some time to compile the svn version (754) and now I got the 
DOI correctly, Thanks a lot.

Nevertheless, I experienced some other problems.

That's the RIS entry I imported into an empty database:

TY  - BOOK
ID  - SCHOLZ_SOFTWARE_ENGINEERING_EMBEDDED
BT  - Softwareentwicklung eingebetteter Systeme
AU  - Scholz,Peter
PY  - 2005///
VL  - 2
CY  - Berlin, Heidelberg
PB  - Springer Verlag
SN  - 3-540-23405-5
M3  - DOI:10.1007/3-540-27522-3
ER  -

When I use
refdbc: getref :ID:=1 -t ris -s ALL

TY  - BOOK
ID  - SCHOLZ_SOFTWARE_ENGINEERING_EMBEDDED
AU  - Scholz,Peter
RP  - NOT IN FILE
BT  - Softwareentwicklung eingebetteter Systeme
CY  - Berlin, Heidelberg
PB  - Springer Verlag
M3  - DOI:10.1007/3-540-27522-3
ER  -
ER  -
PY  - 2005///
SN  - 3-540-23405-5
VL  - 2
ER  -
999:1 retrieved:0 failed

Now I see these weird ER tags, I expected them only at the adn of the 
wohle entry. Maybe that's also the reason why the docbook output also 
looks somehow strange:

refdbc: getref :ID:=1 -t db50x -s ALL

Re: DOI and edition

Stephan Gatzka — 2012-01-27T06:17:21


That's no problem for me.

I will try it out this evening.

Many thanks!

Regards,

Stephan

------------------------------------------------------------------------------
Try before you buy = See our experts in action!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-dev2_______________________________________________
Refdb-users mailing list
Refdb-users< at >lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/refdb-users

DOI and edition

2012-01-26T22:55:00

Hi,

Stephan Gatzka writes:
 > Is it possible or planned to include a
 > <biblioid class="doi">xxxx</biblioid>
 > when exporting to docbook5? I've imported RIS data sets including DO tags.
 > 

This has been implemented in the svn version a while ago.

 > The second question is which RIS tag is the correct one for the 
 > "edition" of a book and is this tag already exported to an <edition> tag 
 > in docbook?
 > 

The RIS spec suggest "VL" for this purpose. I've noticed that this
field was exported incorrectly in the docbook output (it was exported
to volumenum regardless of the reference type). I've checked in a
fix. Editions of books and related types should now be exported
correctly as of svn revision 754.

I'd recommend to use the current svn revision if you're able to
build RefDB from the sources. I've been using this version in a
production environment for months without running into any
problems.

regards,
Markus

DOI and edition

Stephan Gatzka — 2012-01-23T21:28:17

Hi!

I'm a somehow newbie to refdb, but I got it running and I'm impressed by 
the capabilities. I'm using the latest official version 0.9.9-1

Currently I've two questions I didn't found answers yet.

Is it possible or planned to include a
<biblioid class="doi">xxxx</biblioid>
when exporting to docbook5? I've imported RIS data sets including DO tags.

The second question is which RIS tag is the correct one for the 
"edition" of a book and is this tag already exported to an <edition> tag 
in docbook?

Kind regards,

Stephan

------------------------------------------------------------------------------
Try before you buy = See our experts in action!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-dev2_______________________________________________
Refdb-users mailing list
Refdb-users< at >lists.sourceforge.net
https://

RefDB chooses NOT to communicate with latestversionof PostgreSQL

2011-07-17T16:52:48

 > Refdb complains about the version number of the latest version of
 > PostgreSQL 9.0.4 and does not continue.
 > 
 > Is there real incompatibility?
 > 

Hi Frank,

I've set up PostgreSQL 9.03 on my FreeBSD 8.1 box. I've edited
pg_hba.conf to use password authentication for everything. I've then
set up a RefDB database according to

http://refdb.sourceforge.net/manual/ch04s08.html#sect2-manual-setup

I've used manual setup as I've already installed RefDB with MySQL as a
backend. Next I've added myself as a user. The only problem that I
encountered is that I cannot connect to PostgreSQL unless I provide a
password with the refdba adduser command, but this is expected. After
adding the user properly, I was able to use RefDB without any
problems.

Now if that procedure does not work on your box, we'll have to figure
out other potential problems. I've checked the PostgreSQL release
notes. There should not be any changes from 9.03 to 9.04 that pertain
to access problems.

To get an idea of what else may be the c

Re: RefDB chooses NOT to communicate with latest version of PostgreSQL

2011-07-15T21:15:15

Hi Frank,

Frank M Salter writes:
 > main database version is not supported

this in fact indicates that the libdbi driver somehow doesn't manage
to talk to the PostgreSQL engine properly.

 > At some point: I was using a number of terminal screens and did not
 > see the exact moment
 > when the following (transcribed by me) appeared:
 > =======================================================
 > *** glibc detected  *** refdb *** munmap_chunk(): invalid pointer
 > 0xb75f48e0  ***
 > =======  Backtrace:  =======
 > /lib/libc.so.6(+0x6f68a)[0xb76f068a]
 > /usr/local/lib/libdbi.so.1(dbi_result_free + 0x14c)[0xb782827c]
 > refdbd[0x804d9a5]
 > /lib/libc.so.6(__libc_start_main+0xe6)[0xb7697db6]
 > refdbd[0x804aa21]

This looks like a crash in libdbi. This may be caused by some
incompatibility between the driver and the engine, or it's just a
plain libdbi bug. However, I've never had any such crashes with any of
the drivers that I've used, so I still suspect the pgsql driver does
not work well with the latest Postg

RefDB chooses NOT to communicate with latestversionof PostgreSQL

2011-07-13T19:43:11

 > Refdb complains about the version number of the latest version of
 > PostgreSQL 9.0.4 and does not continue.
 > 
 > Is there real incompatibility?
 > 
 > Is there a workaround to allow access?
 > 
 > If not, what is the latest version of PostgreSQL with which RefDB will
 > work successfully?
 > 

Hi Frank,

it would be very helpful to actually see the error messages. In fact,
I suspect the problem is more likely related to the libdbi driver
rather than to RefDB itself. In any case, please use the refdbd -l
(and maybe -e and -L) command line switches or the corresponding
variables in /usr/local/etc/refdb/refdbdrc (don't forget to restart or
reload the server in that case) to switch on debug level
logging. You'll find the documentation right here:

http://refdb.sourceforge.net/manual/re02.html

Now run any refdbc or refdba command which demonstrates the problem,
and send the refdbd log to the list.

You should also try to locate the PostgreSQL log output. I can't recall
where it goes to by default, but have

RefDB chooses NOT to communicate with latest versionof PostgreSQL

Frank M Salter — 2011-07-12T13:57:17

Refdb complains about the version number of the latest version of
PostgreSQL 9.0.4 and does not continue.

Is there real incompatibility?

Is there a workaround to allow access?

If not, what is the latest version of PostgreSQL with which RefDB will
work successfully?

Regards

Frank Salter

------------------------------------------------------------------------------
AppSumo Presents a FREE Video for the SourceForge Community by Eric 
Ries, the creator of the Lean Startup Methodology on "Lean Startup 
Secrets Revealed." This video shows you how to validate your ideas, 
optimize your ideas and identify your business strategy.
http://p.sf.net/sfu/appsumosfdev2dev

Re: Problems with adduser

Torsten Bronger — 2011-06-19T09:09:41

Hallöchen!

Markus Hoenicka writes:


After some further testing, I now see what you mean, so ignore my
previous posting.  And adding an existing user with the same
password seems to be idempotent.  Howerver, password changes don't
happen, at least not in Postgres.

Tschö,
Torsten.