refdb-users — List for RefDB users

Re: [Refdb-users] mandatory RIS fields

[Z F <mai...@ya...>]

Hello Markus,

--- Markus Hoenicka <mar...@mh...> wrote:

> Hi,
> 
> Z F <mai...@ya...> was heard to say:
> 
> > PS in refdb 0.9.6-pre3, I do not see updatejo neither in refdba nor
> in
> > refdbc. Was it introduced later? It is a useful command.
> 
> Actually I was mistaken. This command is in refdbc, not in refdba.
> But I see it
> right here in my refdb-0.9.7-pre1 installation. Do you use 0.9.6-pre3
> or
> 0.9.7-pre3? 0.9.6-pre3 is definitely not up to date.

I hve 0.9.6-pre3 not 0.9.7-pre3.
I used the debian packages, remember? :-)

Lazar

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

Re: [Refdb-users] RIS import errors

[Z F <mai...@ya...>]

Hello Markus,

> Please make sure that the end-of-reference tag has the trailing
> space. That is,
> if "_" stands for a space, you have to use "ER__-_". If you use

Oh, I thought I checked that. As it turns out, on my initial
import I was careful, but when I was adding new references I did not
put the trailing spaces, even though I thought I did, I even thought
I checked that before writing the email. But after your message I
checked, and sure --- no trailing spaces after ER  -

This could probably explain an error I reported of several "ER  -" at
the end of refernces too. When a single reference is added, addref is
probably trying to fix the errors and it can not find "ER  - ", and
adds
it at the end, but "ER  -" (without a space) is treated as content of
the previous field.

Uh... I like the system even though I make so many mistakes in using
it.

> either vim or
> Emacs as an editor, please have a look at:
> 
> http://refdb.sourceforge.net/vim.html
> http://refdb.sourceforge.net/emacs.html
> 
> Both editors can do syntax coloring which makes it easy to spot typos
> like
> these.

I use nano as an editor :-)

Thank you very much for your help

Lazar

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

Re: [Refdb-users] mandatory RIS fields

[Markus H. <mar...@mh...>]

Hi,

Z F <mai...@ya...> was heard to say:

> PS in refdb 0.9.6-pre3, I do not see updatejo neither in refdba nor in
> refdbc. Was it introduced later? It is a useful command.

Actually I was mistaken. This command is in refdbc, not in refdba. But I see it
right here in my refdb-0.9.7-pre1 installation. Do you use 0.9.6-pre3 or
0.9.7-pre3? 0.9.6-pre3 is definitely not up to date.

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

Re: [Refdb-users] RIS import errors

[Markus H. <mar...@mh...>]

Hi,

Z F <mai...@ya...> was heard to say:

> I have created a RIS file manually.
>
> Major points are that each reference starts from an empty line
> followed by TY  - field and ends by ER  -
>
> In the middle, there are some other fields.
> The problem I have is that when I use addref to load the data into
> the database, all the references are recored as on, i.e. only one ID
> is issued. If I split the file into many so that each file contains
> a single reference without modification of the reference content,
> the addref operation is successful on every file.
>

Please make sure that the end-of-reference tag has the trailing space. That is,
if "_" stands for a space, you have to use "ER__-_". If you use either vim or
Emacs as an editor, please have a look at:

http://refdb.sourceforge.net/vim.html
http://refdb.sourceforge.net/emacs.html

Both editors can do syntax coloring which makes it easy to spot typos like
these.

BTW this error can also explain the JF/JO problem that you reported in a
separate mail. If you have two consecutive references, one with JO, the other
with JF, separated by an invalid end-of-reference tag, then RefDB will make the
JO and JF contents synonyms of the same periodical.

regards,
Markus


-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

[Refdb-users] RIS import errors

[Z F <mai...@ya...>]

Hello,

I have another question/problem.

I have created a RIS file manually.

Major points are that each reference starts from an empty line
followed by TY  - field and ends by ER  -

In the middle, there are some other fields.
The problem I have is that when I use addref to load the data into
the database, all the references are recored as on, i.e. only one ID
is issued. If I split the file into many so that each file contains 
a single reference without modification of the reference content,
the addref operation is successful on every file.

Did anybody experience this?

Thanks

Lazar

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

Re: [Refdb-users] mandatory RIS fields

[Z F <mai...@ya...>]

> 
> JF  - Journal of Irreproducible Results
> JO  - J.Irrep.Res.
> 
> then the first one (and all other datasets that specify either of
> these names) will show both the JF and the JO entry when retrieving
> it
> again. RefDB keeps the journal name synonyms in a separate table and
> retrieves whichever synonyms are available for a particular
> periodical. This is a feature, not a bug. There is actually a command
> available (refdba:updatejo) to edit the synonyms.

Oh, so you are saying I made a mistake somewhere and one of the
references contained both fields:

JF  - Journal of Irreproducible Results
JO  - I.Can't.Do.It

and therefore, all future references to "I.Can't.Do.It" were
accossiated
with the "Journal of Irreproducible Results". Ok, I will check if
this is what had happened.

Thanks

Lazar

PS in refdb 0.9.6-pre3, I do not see updatejo neither in refdba nor in
refdbc. Was it introduced later? It is a useful command.

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

[Refdb-users] mandatory RIS fields

[Markus H. <mar...@mh...>]

Hi,

Z F writes:
 > There seems to be a little problem with the addref command.
 > I have noticed that if a reference specifies a JF field, but some other
 > reference in the same ris file does not specify JF (and specifies JO),
 > the value JF field is copied from reference to reference.
 > This happens if a ris file contains several references. Can anyone
 > reproduce this?
 > 

I can't reproduce this here. If the first dataset specifies

JF  - Journal of Irreproducible Results

and the second has

JO  - J.Biol.Chem.

then each dataset shows "his" journal name when retrieving it from
the database.

However, if the first dataset has

JF  - Journal of Irreproducible Results

and the second one has

JF  - Journal of Irreproducible Results
JO  - J.Irrep.Res.

then the first one (and all other datasets that specify either of
these names) will show both the JF and the JO entry when retrieving it
again. RefDB keeps the journal name synonyms in a separate table and
retrieves whichever synonyms are available for a particular
periodical. This is a feature, not a bug. There is actually a command
available (refdba:updatejo) to edit the synonyms.

If this does not explain the JF and JO entries in your case, please
send me a simple RIS file which reproduces the problem, along with the
output that you get back again with the getref command.

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

Re: [Refdb-users] /usr/local/bin/runbib: line 319: 12645 Bus error

[Jeremy M. <Je...@Ma...>]

Markus Hoenicka wrote:
> Jeremy Malcolm <Je...@Ma...> was heard to say:
> 
>>No joy on URL wrapping yet using your tip, but will keep looking at it.
> 
> Come to think of it, my suggestion can't work. The RefDB driver file calls your
> customization layer, therefore it overrides anything with the same precedence.
> It would have worked if the inclusion went the other way round.
> 
> In any case, you may try to edit the RefDB driver files directly and see whether
> URL wrapping works. If it does, I'll include this modification into the driver
> files.

Yes it does!  It works great.  After including the new ulink stuff in my 
local driver file I just commented out this in the RefDB stuff:

(element ulink
     (refdb-process-inline "URLSTYLE"))

-- 
JEREMY MALCOLM <Je...@Ma...> - lawyer, IT consultant and actor.
Internet and Open Source specialist. Web site: http://www.malcolm.id.au.
Disclaimer: http://www.terminus.net.au/disclaimer.html. GPG key: finger.

Re: [Refdb-users] /usr/local/bin/runbib: line 319: 12645 Bus error

[Markus H. <mar...@mh...>]

Jeremy Malcolm <Je...@Ma...> was heard to say:


> No joy on URL wrapping yet using your tip, but will keep looking at it.
>

Come to think of it, my suggestion can't work. The RefDB driver file calls your
customization layer, therefore it overrides anything with the same precedence.
It would have worked if the inclusion went the other way round.

In any case, you may try to edit the RefDB driver files directly and see whether
URL wrapping works. If it does, I'll include this modification into the driver
files.

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

Re: [Refdb-users] The state of my Debian RefDB/PHPweb confusion

[Markus H. <mar...@mh...>]

Hi,

Doug du Boulay <dou...@gm...>:

>
> After some initial confusion on my part in creating the database (following
> the instructions in the tutorial "a day with the refdb clients"
> which made no mention of running the refdba command "createdb DBNAME"),


The tutorial was written for users who are placed in front of a running RefDB
system and who are expected to work with existing databases. The manual has
"Getting Started" sections as well which include the magic incantations to
create reference databases.

> (I did experiment with replacing  /lib/tls/libc.so.6 with a symlink
> to /lib/libc.so.6  and while the "mv" command still works, "ls" doesn't,
> the CPU started to overheat, the fan went a little berzerk and bib2ris and
> refdbc both still segfaulted - so don't try that at home folks).
>

I hope you managed to pull the plug in time.

> I downloaded the sources for both libdbi and libdbi-drivers 0.8.1
> Although libdbi seemed to build ok, trying to build the libdbi-sqlite driver
> was a nonevent because ./configure --with-sqlite fails with
> error: Invalid SQLite directory - include files not found
>  - presumably that is because I needed libsqlite0-dev  right?

This is correct. Debian strictly separates run-time libraries from the headers
and libraries required to build stuff. Whenever you build something yourself,
you most likely need -dev packages for a dozen of libraries.

> hint of libdbi-0.8.1.  So I gave up and manually scrounged around in
>
>  http://libdbi.sourceforge.net/debian/pool/l/
>
> there I found
>    libdbd-sqlite_0.8.1-1_i386.deb
>    libdbi0_0.8.1-1_i386.deb
>    libdbi0-dev_0.8.0-1_i386.deb
>
> The latter one had a 8.0!=8.1 version conflict so I ran
>    dpkg -force-depends -i  libdbi0-dev_0.8.0-1_i386.deb

Unfortunately I never got the hang of creating Debian repositories. I've always
uploaded what David sent me, but he got lost somehow, so there's no Debian
maintainer currently. This explains the broken state of the repositories.

>
> Then I could build and install refdb-0.9.7-pre1  along with
> RefDB-perlmod-0.4 and RefDBClient-Client-1.13
> (a README to the effect of running "perl Makefile.PL" ; make ; make install
> could have been useful info for the latter two),

This was suggested before. It is my fault that the READMEs are still missing.

> In my subsequent limited testing, it seems that refdbc had some trouble
> producing debugging info and writing it to the log file. Maybe it was because
> of restrictive file permissions? I dunno.

I'm pretty sure this is the reason. You start refdbc as a regular user, but on
most systems only root has sufficient permissions to create files in /var/log.
Either use your own log files, e.g. in ~/var/log, or create the files as root
and change the permissions appropriately.

> Also both refdbc and refdba produce log file info with time stamps that seem
> to be standard GMT rather than local timezone.

This does make sense in a distributed environment. You could start refdbc and
connect to a refdbd instance running in a different timezone.

> Another logged bogus search command was from the refdbkajquery.html form
>    getau refs :AU:=Wang,J.P.
> where the database "refs" is prepended to the command. Actually
> that database is populated with the refdb/examples/testrefs.ris data.
> Presumably missing a leading "-d" before the database name?

I assume you followed the instructions in the phpweb subdirectory? I also had a
hard time setting up the web interface, but I hoped the instructions contain
the essential tricks.

> A different problem arose with the commandline interface for
> refdbc where I can't seem to get wildcard/regular expression searches
> to work e.g.
> getref :AU:~Wang,J.P. works, but
> getref :AU:~Wang*  does not
> getref :AU:~"Wang*"  doesn't
> getref :AU:~"^Wang*" doesn't
>
> Is there something wrong with my syntax,
> a configuration error, or a bug?

The syntax is wrong. If you use PostgreSQL or MySQL as your backend, then you
have to use Unix regular expressions. The expressions above would find Wang,
Wangg, and Wanggggggggggggg, but this is not what you mean. Searching for
"Wang.*" (i.e. Wang, followed by zero or more arbitrary characters) should give
better results.

>
> The only other problem was that the refdb configure script apparently doesn't
> like my debian libbtparse0 library
> ii  libbtparse0        0.34-1             A C library to parse BibTeX files
> -rw-r--r--  1 root root   65808 Jun 18  2004 /usr/lib/libbtparse.so.0.0.0
> > checking for bt_initialize in -lbtparse... no
> > libbtparse not found; bib2ris import filter will not be built
>
Most likely you lack a libbtparse-dev package, see above.

>
> Thats about it, I think.

Well, that was enough for a single mail, I reckon :-)

regards
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

[Refdb-users] The state of my Debian RefDB/PHPweb confusion

[Doug du B. <dou...@gm...>]

Hi list,

I tried several weeks ago to install the debian-ised packages from 
the /etc/apt/sources.list resource:
     deb http://refdb.sourceforge.net/debian/release unstable main

as per the instructions  on the website:
           http://refdb.sourceforge.net/download

I believe, but am not totally sure that this also gave me 
> dpkg --list | grep libdbi

ii  libdbi0        0.7.2-1        Database Independent Abstraction Layer for C
ii  libdbi0-dev    0.7.2-1        Database Independent Abstraction Layer for C
 - along with 
ii  libdbd-sqlite  0.7.1-3.0.1    SQLite database driver for libdbi
ii  refdb          0.9.6-pre2     Reference database and bibliography tool for

After some initial confusion on my part in creating the database (following 
the instructions in the tutorial "a day with the refdb clients" 
which made no mention of running the refdba command "createdb DBNAME"),
it turned out that certain library incompatabilities were leading to 
segmentation faults.

(I did experiment with replacing  /lib/tls/libc.so.6 with a symlink 
to /lib/libc.so.6  and while the "mv" command still works, "ls" doesn't,
the CPU started to overheat, the fan went a little berzerk and bib2ris and 
refdbc both still segfaulted - so don't try that at home folks).


Sometime thereafter I decided to build from the sources and downloaded 
     refdb-0.9.7-pre1 
Trying to build that highlighted incompatabilities with libdbi-0.7.2.1 
(as well as my libreadline libraries and dev files)
and required an update to 0.8.1 as indicated by Markus 
Re: [Refdb-users] 0.9.6 Compile problem

I downloaded the sources for both libdbi and libdbi-drivers 0.8.1
Although libdbi seemed to build ok, trying to build the libdbi-sqlite driver
was a nonevent because ./configure --with-sqlite fails with
error: Invalid SQLite directory - include files not found
 - presumably that is because I needed libsqlite0-dev  right?

(ii  libsqlite0         2.8.16-1           SQLite shared library)

Anyway, I put it on the backburner for a while until I read Per-Anders' 
recent email and decided to add the following to my apt/sources.list

deb http://refdb.sourceforge.net/debian/release unstable main
deb  http://refdb.sourceforge.net/debian/cvs/ testing main
deb  http://refdb.sourceforge.net/debian/cvs/ unstable main
deb http://libdbi.sourceforge.net/debian unstable main

I dont know what the deal is with the cvs bit, but it added an extra
3.5 megabytes to the packagelist update process and didn't seem to provide any 
hint of libdbi-0.8.1.  So I gave up and manually scrounged around in 

 http://libdbi.sourceforge.net/debian/pool/l/

there I found 
   libdbd-sqlite_0.8.1-1_i386.deb
   libdbi0_0.8.1-1_i386.deb
   libdbi0-dev_0.8.0-1_i386.deb 

The latter one had a 8.0!=8.1 version conflict so I ran
   dpkg -force-depends -i  libdbi0-dev_0.8.0-1_i386.deb 

Then I could build and install refdb-0.9.7-pre1  along with 
RefDB-perlmod-0.4 and RefDBClient-Client-1.13
(a README to the effect of running "perl Makefile.PL" ; make ; make install 
could have been useful info for the latter two),
and now things seem to work, mostly, I think...
Also during ./configure of refdb, I had
a spot of trouble installing from CPAN the MARC-Charset-0.95.tar.gz perl 
module. Fortunately, I was able to install MARC-Charset-0.6.tar.gz from an 
out of date CPAN mirror. 
Are there debian packages for such things or does everyone just use 
perl -MCPAN ... ?

In my subsequent limited testing, it seems that refdbc had some trouble 
producing debugging info and writing it to the log file. Maybe it was because 
of restrictive file permissions? I dunno.
Also both refdbc and refdba produce log file info with time stamps that seem 
to be standard GMT rather than local timezone. 

The reason I was interested in the log file was because I was having 
trouble with the phpweb interface, to the effect that no search results 
were ever forthcoming.

The only debugging info I was getting was in the Apache2/error.log which said
sh: line 1: -c: command not found

I guess this was because the session info wasn't saved because I didn't 
know about the php.ini file settings (so the /usr/local/bin/refdbc command 
wasn't retained and prepended to the command). I tried 
coppying /etc/php4/apache2/php.ini  to the /var/local/share/refdb/www/
directory, and amending the relevant values with those listed in the
refdb/phpweb/INSTALL file. 
I tried to create a /var/local/share/refdb/www/tmp/ directory, in case
the session info needs to be saved in /tmp relative to the 
http://localhost/refdb/ directory. that didn't seem to work.
So I coppied the amended php.ini file back to /etc/php4/apache2/php.ini.
That seemed to do the job because it now creates files like
/tmp/sess_5f278997b0d82e590a8706b6c491110b  owned by nobody of nogroup 
and cures the "command not found" error, but it still fails to return any 
references when I do any kind of search.
(I would like to think there is a better way to make that work, but I have no 
idea what it is)

To figure what was going wrong I set
$_SESSION["refdbc"]="/usr/local/bin/refdbc -l 7 -e 2 -L /var/log/refdbc.log";
in login.php. But without making the log file writeable by all, there 
didn't seem to be any logging output.

Thereafter doing a search for author Wang,J.P. the search command logged was 
apparently:
     getref ALL :AU:~Wang,J.P. 
i.e. a seemingly spurious "ALL" arising from 
function setSearchFields() in refdbsearch.php

Another logged bogus search command was from the refdbkajquery.html form
   getau refs :AU:=Wang,J.P. 
where the database "refs" is prepended to the command. Actually 
that database is populated with the refdb/examples/testrefs.ris data.
Presumably missing a leading "-d" before the database name?




A different problem arose with the commandline interface for 
refdbc where I can't seem to get wildcard/regular expression searches 
to work e.g.
getref :AU:~Wang,J.P. works, but
getref :AU:~Wang*  does not
getref :AU:~"Wang*"  doesn't
getref :AU:~"^Wang*" doesn't 

Is there something wrong with my syntax, 
a configuration error, or a bug?

The only other problem was that the refdb configure script apparently doesn't 
like my debian libbtparse0 library
ii  libbtparse0        0.34-1             A C library to parse BibTeX files
-rw-r--r--  1 root root   65808 Jun 18  2004 /usr/lib/libbtparse.so.0.0.0
> checking for bt_initialize in -lbtparse... no
> libbtparse not found; bib2ris import filter will not be built


Thats about it, I think.
Advice appreciated.
Many thanks
Doug

Re: [Refdb-users] /usr/local/bin/runbib: line 319: 12645 Bus error

[Jeremy M. <Je...@Ma...>]

Markus Hoenicka wrote:
> Jeremy Malcolm writes:
>  > Yes.  It is a problem that was introduced with pre2.
>  > 
> [...]
>  > I now get /usr/local/bin/runbib: line 319:  6706 Bus error
>  > 
> 
> I think I've fixed this problem, along with a couple of silent
> glitches that I caught while I was it. Please give pre4 (see my
> announcement in a separate mail) a try.

All working! :-)

No joy on URL wrapping yet using your tip, but will keep looking at it.

-- 
JEREMY MALCOLM <Je...@Ma...> - lawyer, IT consultant and actor.
Internet and Open Source specialist. Web site: http://www.malcolm.id.au.
Disclaimer: http://www.terminus.net.au/disclaimer.html. GPG key: finger.

[Refdb-users] mandatory RIS fields

[Z F <mai...@ya...>]

As far as I understood only TY and ER are mandatory.

There seems to be a little problem with the addref command.
I have noticed that if a reference specifies a JF field, but some other
reference in the same ris file does not specify JF (and specifies JO),
the value JF field is copied from reference to reference.
This happens if a ris file contains several references. Can anyone
reproduce this?

Thanks

Lazar

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

[Refdb-users] ANN: refdb-0.9.7-pre4 available for download

[Markus H. <mar...@mh...>]

Hi,

after a not-so-shining pre3 (which had to be pulled before it was
officially released) I'm presenting 0.9.7-pre4. I hope that the
problems reported by Per-Anders and by Jeremy are fixed
finally. Please follow this link to get the prerelease:

http://refdb.sourceforge.net/pre/refdb-latest.tar.gz

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

Re: [Refdb-users] /usr/local/bin/runbib: line 319: 12645 Bus error

[Markus H. <mar...@mh...>]

Jeremy Malcolm writes:
 > Yes.  It is a problem that was introduced with pre2.
 > 
[...]
 > I now get /usr/local/bin/runbib: line 319:  6706 Bus error
 > 

I think I've fixed this problem, along with a couple of silent
glitches that I caught while I was it. Please give pre4 (see my
announcement in a separate mail) a try.

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

Re: [Refdb-users] Last reference

[Z F <mai...@ya...>]

Hello everybody,

I have retrieved the last reference from the data base in 
ris format and, indeed, it had ER - twice it its body (after the title)
The original ris file had only one ER - at the end is it should.
So, I have removed the ER - in the middle of the reference and used
updateref. Everything seems to be fine now.

It apears that somehow, during the original import of the ris file with
addref something happened on the last reference. Can not reproduce it
anymore. :-)

Lazar

--- Markus Hoenicka <mar...@mh...> wrote:

> Hi,
> 
> Z F writes:
>  > When I retrieve citations, the citation with the highest ID number
>  > always has "ER -" twice in the body of the reference.
>  > 
>  > Did anybody notice a similar behavior? I am using refdb V.
> 0.9.6-pre3
>  >
> 
> Not here on FreeBSD 5.4. Does anyone else see this behaviour? If so,
> please indicate your platform, maybe there's a pattern that helps to
> find the bug.
> 
> regards,
> Markus
> 
> -- 
> Markus Hoenicka
> mar...@ca...
> (Spam-protected email: replace the quadrupeds with "mhoenicka")
> http://www.mhoenicka.de
> 
> 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

Re: [Refdb-users] Case sensitivity

[Z F <mai...@ya...>]

Hello Markus,

> I wouldn't call it "logic", but there is a historical reason. The
> author field (and keyword, periodical, and a couple of others) are
> fixed-length fields which are stored in VARCHAR columns, as this is
> the most economical type in this case. Title, notes, abstract and
> other
> fields of unlimited length are stored as binary strings. I chose
> binary over text back then because I *wanted* the search in these
> fields to be case-sensitive.

Ok, so the data are stored differently.

> 
> However, with MySQL 4.1 and later binary strings actually have a
> disadvantage as they don't use the character encoding when it comes
> to
> the sorting order. Using the proper character encoding may actually
> outweigh case-sensitive search (which you don't seem to like anyway)
> these days. I'm open for a discussion how this should be treated in
> the future. In any case, MySQL is quite a mess in these issues.

Using proper character encoding is a good thing which is an argument
for using a text field. Also, in my opinion, searches should be
case insensitive when "LIKE" is used (or at least there should be a 
way to do case-insensitive searches). The reason for this is simple.
A word can be in the begining of a sentance, which means it starts from
an upper case letter. The word could be an acronym, then all letters
are probably upper case. But when performing a search, the user
should not think weather or not, the word my be in the begining or in 
the middle of a sentence. This is the meaning of "LIKE", in my opinion.
"IS" search should probably be case insensitive (or whatever SQL does)
All of this argues for the use of text fields instead of binary ones.
In any case, you word is the last one. :-)

Lazar

PS
As one of the recent US presidents once said: "It depends what word IS
is"... :-)



__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

Re: [Refdb-users] URL wrapping (off-topic)

[Markus H. <mar...@mh...>]

Jeremy Malcolm <Je...@Ma...> was heard to say:

> * In my driver file (Thesis.dsl) I have this (I didn't write it, I just
>    found it online, so I don't know if it's the only way to do it):
>
[...]
> (element ulink

I didn't get round to run any tests yet, but I believe your driver file does not
succeed in overriding the RefDB ulink formatting because the RefDB driver file
treats this element inside of a mode. Therefore the RefDB driver file
definition is more specific and is applied.

The following definition might work better:

(mode biblioentry-inline-refdb-mode
  (element ulink
      ...your definition goes here...
  )
)

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

Re: [Refdb-users] URL wrapping (off-topic)

[Jeremy M. <Je...@Ma...>]

Markus Hoenicka wrote:
> I'm afraid I've lost you here. The URLs in the bibliography are
> supposed to be ulinks, and they seem to be, according to your own test
> data. The generated bibliography contains e.g.:
> 
> <ulink role="url"
> url="http://homes.eff.org/~barlow/Declaration-Final.html">http://homes.eff.org/~barlow/Declaration-Final.html</ulink> 

Sorry, of course you're right.  Still, the TeX magic doesn't work for 
RefDB ulinks and does work for other ulinks.  Here is what I do:

* In jadetex.cfg, I have \usepackage{url}

* In my driver file (Thesis.dsl) I have this (I didn't write it, I just
   found it online, so I don't know if it's the only way to do it):

--- begins ---

(declare-flow-object-class formatting-instruction
        "UNREGISTERED::James Clark//Flow Object 
Class::formatting-instruction")

(element ulink
   (make sequence
     (if (node-list-empty? (children (current-node)))
         ; ulink url="...", /ulink
         (make formatting-instruction
           data: (string-append "\\url{"
                                (attribute-string (normalize "url"))
                                "}"))
         (if (equal? (attribute-string (normalize "url"))
                     (data-of (current-node)))
         ; ulink url="http://...", http://..., /ulink
             (make formatting-instruction data:
                   (string-append "\\url{"
                                  (attribute-string (normalize "url"))
                                  "}"))
         ; ulink url="http://...", some text, /ulink
             (make sequence
               ($charseq$)
               (literal " (")
               (make formatting-instruction data:
                     (string-append "\\url{"
                                    (attribute-string (normalize "url"))
                                    "}"))
               (literal ")"))))))

--- ends ---

By separate email I will send a small TeX file in which the URL is 
wrapped correctly in the body text but the RefDB URL is not.

-- 
JEREMY MALCOLM <Je...@Ma...> - lawyer, IT consultant and actor.
Internet and Open Source specialist. Web site: http://www.malcolm.id.au.
Disclaimer: http://www.terminus.net.au/disclaimer.html. GPG key: finger.

Re: [Refdb-users] Last reference

[Per-Anders J G. <PA...@lb...>]

Dear Lazar,

The 'hi' installation code means that a package is held, that is, it
won't get upgraded by an 'apt-get dist-upgrade'. The packages are
installed fine and working as they should. (An h in the second position
would mean half-installed.)

Unfortunately, my laptop died yesterday, so I can't dig up more info for
you right now. I supplied the list of installed packages to show what
'works for me' (tm). One reason the installation may look strange is
that the system has gone through quite a few upgrades. For example, the
mysql4 configuration files were left on the system by apt-get when
upgrading to mysql5. As far as I know, they don't interfere with mysql5
at all. I generally tend to trust that the dependencies are set up
correctly for the packages when doing installations and upgrades. If the
system doesn't complain and things seem to work, I'm happy. :) And in
the case of refdb, if it builds and seems to work, I'm happy.

I may look more into this once I get my system up and running again.

Regards,

Anders


----- Original Message -----
From: Z F <mai...@ya...>
Date: Saturday, February 11, 2006 1:27 pm
Subject: Re: [Refdb-users] Last reference

> Dear Andres,
> 
> 
> > I installed mysql5 from the official debian repository. The libdbi
> > drivers
> > were installed from the debianised packages on sourceforge. I had 
> > 
> > deb  http://refdb.sourceforge.net/debian/cvs/ testing main
> > deb  http://refdb.sourceforge.net/debian/cvs/ unstable main
> > 
> > in my sources.list and used apt-get. No tweaks were needed, as 
> far as
> > I remember. Though after creating a new reference database with
> > refdba,
> > I had to manually set the user privileges in mysql to be able to
> > access
> > the database from refdbc. 
> > 
> > refdb was configured with ./configure --sysconfdir=/etc
> > --with-db-server=mysql
> > then make and make install.
> > 
> 
> this is what I have too. But you have compiled refdb yourself, you
> did not use the package from refdb.soureforge.net
> 
> Below, in the list of packages, you, indeed, removed mysql4 server,
> but did not remove its configuration files. libdbd-mysql uses
> libmysqlclien12 whcih is compiled  for mysql4. More over, you have
> half-configured libdbd-mysql (I do not know what that means, but it
> isnot installed properly). In addition, you recompiled refdb yourself,
> as you say, and linked it with libmysqlcliet14 (as far as I can tell
> form the list of packages), your setup is very very strange. I do not
> understand how it works, since I could not get the system working
> with mysql5 server and libdbd-mysql linked with libmysqlclient12 (this
> is what you have)
> 
> Lazar
> 
> > My libdbi:
> > 
> > pergl@chipmunk:~$ dpkg --list | grep libdbi
> > hi  libdbd-mysql                     0.8.1-1                      
> > MySQL database server driver for libdbi
> > ii  libdbd-sqlite                    0.8.1-1                      
> > SQLite database server driver for libdbi
> > ii  libdbi-perl                      1.50-2                       
> > Perl5 database interface by Tim Bunce
> > hi  libdbi0                          0.8.1-1                      
> > Database Independent Abstraction Layer for C
> > hi  libdbi0-dev                      0.8.1-1                      
> > Database Independent Abstraction Layer for C
> > 
> > 
> > And my mysql:
> > 
> > pergl@chipmunk:~$ dpkg --list | grep mysql 
> > ii  libapache-mod-auth-mysql         4.3.9-2                      
> > Apache module for MySQL authentication
> > hi  libdbd-mysql                     0.8.1-1                      
> > MySQL database server driver for libdbi
> > ii  libdbd-mysql-perl                3.0002-2                     
> A
> > Perl5 database interface to the MySQL database
> > ii  libmysqlclient12                 4.0.24-10sarge1              
> > mysql database client library
> > ii  libmysqlclient14                 4.1.15-1                     
> > mysql database client library
> > ii  libmysqlclient14-dev             4.1.15-1                     
> > mysql database development files
> > ii  libmysqlclient15                 5.0.18-7                     
> > mysql database client library
> > ii  mysql-admin                      1.1.6-1                      
> > GUI tool for intuitive MySQL administration
> > ii  mysql-admin-common               1.1.6-1                      
> > Architecture independent files for MySQL Administrator
> > ii  mysql-client-5.0                 5.0.18-7                     
> > mysql database client binaries
> > ii  mysql-common                     5.0.18-7                     
> > mysql database common files (e.g. /etc/mysql/my.cnf)
> > rc  mysql-common-4.1                 4.1.11a-4sarge2              
> > mysql database common files (e.g. /etc/mysql/my.cnf)
> > ii  mysql-query-browser              1.1.17-5                     
> > Official GUI tool to query MySQL database
> > ii  mysql-query-browser-common       1.1.17-5                     
> > Architecture independent files for MySQL Query Browser
> > rc  mysql-server-4.1                 4.1.14-6                     
> > mysql database server binaries
> > ii  mysql-server-5.0                 5.0.18-7                     
> > mysql database server binaries
> > ii  php4-mysql                       4.4.2-1                      
> > MySQL module for php4
> > 
> > 
> > myslq4 has been removed ('rc'), while all versions of libmysqlclient
> > seem
> > to happily co-exist.
> > 
> > Regards,
> > 
> > Anders
> > 
> > 
> 
> 
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam protection around 
> http://mail.yahoo.com 
> 

[Refdb-users] URL wrapping (off-topic)

[Markus H. <mar...@mh...>]

Jeremy Malcolm writes:
 > Is there any way to force URLs in the bibliography to wrap in the 
 > printed output, where the URL extends past the end of a line?  Currently 
 > URLs will only wrap if there is a - character in them, they won't wrap 
 > on / characters, which sometimes means they push through the right margin.
 > 

This is a general problem also in preformatted text, like
<programlisting> or <screen> content.

 > I tried to do that by using the TeX 'url' package.  This requires the 
 > URLs to be <ulink>s.  However if I turn my URLs in the bibliography into 
 > <ulinks>, I get errors because the schema does not allow them there.
 > 

I'm afraid I've lost you here. The URLs in the bibliography are
supposed to be ulinks, and they seem to be, according to your own test
data. The generated bibliography contains e.g.:

<ulink role="url"
url="http://homes.eff.org/~barlow/Declaration-Final.html">http://homes.eff.org/~barlow/Declaration-Final.html</ulink> 

Or do you mean the *contents* of the ulink element would have to be an
<ulink> again for the TeX magic to work? If that were the case, I
wonder whether the DSSSL stylesheet could be tweaked to suppress the
contents of the element and to pass the remainder to TeX. ulinks as
the one shown above are pretty ugly in the printed output anyway as
the url is essentially printed twice, the second one being wrapped in
parentheses.

Could you please send a very simple example of a TeX file with an URL
that is wrapped correctly, including the extra packages that you have
to include? I'll see whether the necessary tweaks could be included
into the RefDB tools.

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

Re: [Refdb-users] Case sensitivity

[Markus H. <mar...@mh...>]

Hi,

Z F writes:
 > This is what I thought. MySQL does the right thing when using
 > = and ~ on the text fields. As far as I understood, author name and
 > title are text fields. Am I wrong? If yes, the comparison should work
 > correctly. However, comparisons on the author field work in
 > case-insensitive way, while comparisons on the title field are case
 > sensitive. So this suggests that author field is text, while title
 > field is somethign else. Is there logic behind this? According to 
 > documentation both fields are alphanumeric.
 > 

I wouldn't call it "logic", but there is a historical reason. The
author field (and keyword, periodical, and a couple of others) are
fixed-length fields which are stored in VARCHAR columns, as this is
the most economical type in this case. Title, notes, abstract and other
fields of unlimited length are stored as binary strings. I chose
binary over text back then because I *wanted* the search in these
fields to be case-sensitive.

However, with MySQL 4.1 and later binary strings actually have a
disadvantage as they don't use the character encoding when it comes to
the sorting order. Using the proper character encoding may actually
outweigh case-sensitive search (which you don't seem to like anyway)
these days. I'm open for a discussion how this should be treated in
the future. In any case, MySQL is quite a mess in these issues.

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de

[Refdb-users] URL wrapping (off-topic)

[Jeremy M. <Je...@Ma...>]

Is there any way to force URLs in the bibliography to wrap in the 
printed output, where the URL extends past the end of a line?  Currently 
URLs will only wrap if there is a - character in them, they won't wrap 
on / characters, which sometimes means they push through the right margin.

I tried to do that by using the TeX 'url' package.  This requires the 
URLs to be <ulink>s.  However if I turn my URLs in the bibliography into 
<ulinks>, I get errors because the schema does not allow them there.

TIA

-- 
JEREMY MALCOLM <Je...@Ma...> - lawyer, IT consultant and actor.
Internet and Open Source specialist. Web site: http://www.malcolm.id.au.
Disclaimer: http://www.terminus.net.au/disclaimer.html. GPG key: finger.

Re: [Refdb-users] Last reference

[Z F <mai...@ya...>]

Dear Andres,


> I installed mysql5 from the official debian repository. The libdbi
> drivers
> were installed from the debianised packages on sourceforge. I had 
> 
> deb  http://refdb.sourceforge.net/debian/cvs/ testing main
> deb  http://refdb.sourceforge.net/debian/cvs/ unstable main
> 
> in my sources.list and used apt-get. No tweaks were needed, as far as
> I remember. Though after creating a new reference database with
> refdba,
> I had to manually set the user privileges in mysql to be able to
> access
> the database from refdbc. 
> 
> refdb was configured with ./configure --sysconfdir=/etc
> --with-db-server=mysql
> then make and make install.
> 

this is what I have too. But you have compiled refdb yourself, you
did not use the package from refdb.soureforge.net

Below, in the list of packages, you, indeed, removed mysql4 server,
but did not remove its configuration files. libdbd-mysql uses
libmysqlclien12 whcih is compiled  for mysql4. More over, you have
half-configured libdbd-mysql (I do not know what that means, but it
isnot installed properly). In addition, you recompiled refdb yourself,
as you say, and linked it with libmysqlcliet14 (as far as I can tell
form the list of packages), your setup is very very strange. I do not
understand how it works, since I could not get the system working
with mysql5 server and libdbd-mysql linked with libmysqlclient12 (this
is what you have)

Lazar

> My libdbi:
> 
> pergl@chipmunk:~$ dpkg --list | grep libdbi
> hi  libdbd-mysql                     0.8.1-1                      
> MySQL database server driver for libdbi
> ii  libdbd-sqlite                    0.8.1-1                      
> SQLite database server driver for libdbi
> ii  libdbi-perl                      1.50-2                       
> Perl5 database interface by Tim Bunce
> hi  libdbi0                          0.8.1-1                      
> Database Independent Abstraction Layer for C
> hi  libdbi0-dev                      0.8.1-1                      
> Database Independent Abstraction Layer for C
> 
> 
> And my mysql:
> 
> pergl@chipmunk:~$ dpkg --list | grep mysql 
> ii  libapache-mod-auth-mysql         4.3.9-2                      
> Apache module for MySQL authentication
> hi  libdbd-mysql                     0.8.1-1                      
> MySQL database server driver for libdbi
> ii  libdbd-mysql-perl                3.0002-2                      A
> Perl5 database interface to the MySQL database
> ii  libmysqlclient12                 4.0.24-10sarge1              
> mysql database client library
> ii  libmysqlclient14                 4.1.15-1                     
> mysql database client library
> ii  libmysqlclient14-dev             4.1.15-1                     
> mysql database development files
> ii  libmysqlclient15                 5.0.18-7                     
> mysql database client library
> ii  mysql-admin                      1.1.6-1                      
> GUI tool for intuitive MySQL administration
> ii  mysql-admin-common               1.1.6-1                      
> Architecture independent files for MySQL Administrator
> ii  mysql-client-5.0                 5.0.18-7                     
> mysql database client binaries
> ii  mysql-common                     5.0.18-7                     
> mysql database common files (e.g. /etc/mysql/my.cnf)
> rc  mysql-common-4.1                 4.1.11a-4sarge2              
> mysql database common files (e.g. /etc/mysql/my.cnf)
> ii  mysql-query-browser              1.1.17-5                     
> Official GUI tool to query MySQL database
> ii  mysql-query-browser-common       1.1.17-5                     
> Architecture independent files for MySQL Query Browser
> rc  mysql-server-4.1                 4.1.14-6                     
> mysql database server binaries
> ii  mysql-server-5.0                 5.0.18-7                     
> mysql database server binaries
> ii  php4-mysql                       4.4.2-1                      
> MySQL module for php4
> 
> 
> myslq4 has been removed ('rc'), while all versions of libmysqlclient
> seem
> to happily co-exist.
> 
> Regards,
> 
> Anders
> 
> 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

Re: [Refdb-users] Last reference

[Per-Anders G. <PA...@lb...>]

Hi Lazar,

On Sat, Feb 11, 2006 at 06:17:41AM -0800, Z F wrote:
> 
> 
> --- Per-Anders Glans <PA...@lb...> wrote:
> 
> > On Fri, Feb 10, 2006 at 11:46:35PM +0100, Markus Hoenicka wrote:
> > > Hi,
> > > 
> > > Z F writes:
> > >  > When I retrieve citations, the citation with the highest ID
> > number
> > >  > always has "ER -" twice in the body of the reference.
> > >  > 
> > >  > Did anybody notice a similar behavior? I am using refdb V.
> > 0.9.6-pre3
> > >  >
> > > 
> > > Not here on FreeBSD 5.4. Does anyone else see this behaviour? If
> > so,
> > > please indicate your platform, maybe there's a pattern that helps
> > to
> > > find the bug.
> > 
> > I don't see it on Debian testing/unstable, RefDB 0.9.6, MySQL 5.0.18,
> > 
> > libdbi 0.8.1.
> 
> I have very similar setup. Debian unstable with the latest
> packages of mysql5. Anders, I am curious how you installed 
> refdb on your system as I could not install it from the available
> packages, I had to recompile libdbd-mysql driver.

I installed mysql5 from the official debian repository. The libdbi drivers
were installed from the debianised packages on sourceforge. I had 

deb  http://refdb.sourceforge.net/debian/cvs/ testing main
deb  http://refdb.sourceforge.net/debian/cvs/ unstable main

in my sources.list and used apt-get. No tweaks were needed, as far as
I remember. Though after creating a new reference database with refdba,
I had to manually set the user privileges in mysql to be able to access
the database from refdbc. 

refdb was configured with ./configure --sysconfdir=/etc --with-db-server=mysql
then make and make install.

My libdbi:

pergl@chipmunk:~$ dpkg --list | grep libdbi
hi  libdbd-mysql                     0.8.1-1                       MySQL database server driver for libdbi
ii  libdbd-sqlite                    0.8.1-1                       SQLite database server driver for libdbi
ii  libdbi-perl                      1.50-2                        Perl5 database interface by Tim Bunce
hi  libdbi0                          0.8.1-1                       Database Independent Abstraction Layer for C
hi  libdbi0-dev                      0.8.1-1                       Database Independent Abstraction Layer for C


And my mysql:

pergl@chipmunk:~$ dpkg --list | grep mysql 
ii  libapache-mod-auth-mysql         4.3.9-2                       Apache module for MySQL authentication
hi  libdbd-mysql                     0.8.1-1                       MySQL database server driver for libdbi
ii  libdbd-mysql-perl                3.0002-2                      A Perl5 database interface to the MySQL database
ii  libmysqlclient12                 4.0.24-10sarge1               mysql database client library
ii  libmysqlclient14                 4.1.15-1                      mysql database client library
ii  libmysqlclient14-dev             4.1.15-1                      mysql database development files
ii  libmysqlclient15                 5.0.18-7                      mysql database client library
ii  mysql-admin                      1.1.6-1                       GUI tool for intuitive MySQL administration
ii  mysql-admin-common               1.1.6-1                       Architecture independent files for MySQL Administrator
ii  mysql-client-5.0                 5.0.18-7                      mysql database client binaries
ii  mysql-common                     5.0.18-7                      mysql database common files (e.g. /etc/mysql/my.cnf)
rc  mysql-common-4.1                 4.1.11a-4sarge2               mysql database common files (e.g. /etc/mysql/my.cnf)
ii  mysql-query-browser              1.1.17-5                      Official GUI tool to query MySQL database
ii  mysql-query-browser-common       1.1.17-5                      Architecture independent files for MySQL Query Browser
rc  mysql-server-4.1                 4.1.14-6                      mysql database server binaries
ii  mysql-server-5.0                 5.0.18-7                      mysql database server binaries
ii  php4-mysql                       4.4.2-1                       MySQL module for php4


myslq4 has been removed ('rc'), while all versions of libmysqlclient seem
to happily co-exist.

Regards,

Anders