This has bearing on bugs 2905, 2665, 2514 and other "wide character" crashes
related to diacritics and Unicode. This should help open the door for reliable
input of diacriticals via acquisitions.
MARC21_utf8_flag_fix.pl diagnoses and fixes existing problems with MARC data
affected by the bug.
Adding SetMarcUnicodeFlag to TransformKohaToMarc prevents the bug from corrupting
further data.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
Acquisitions process seems to be adding records with incorrect
representation of the MARC encoding in leader/09. It should be
'a' meaning UTF-8, for all Koha's internalized records, but in
many cases it appears blank (for MARC-8). This script diagnoses
and repairs the value in the leader, depending on runtime options.
The symptom of this problem is that high-value UNICODE characters
in the record will cause Koha to crash whenever it tries to parse
the MARCXML, giving a "Wide character" fatal. While we work on
fixing the input, this script will fix the existing data.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
The description for autoBarcode had an unescaped single-quote.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
This doesn't insert the large and typically unused <ldapserver> block,
just the switch with a comment pointing to C4::Auth_with_ldap. This otherwise
was undocumented requirement, making LDAP config a bit of a shot in the dark.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
- Moved the following system preferences to the Serials tab:
OPACDisplayExtendedSubInfo
OPACSubscriptionDisplay
RenewSerialAddsSuggestion
SubscriptionHistory
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
* Use of uninitialized value in pattern match (m//) at C4/Biblio.pm line 1305
(displaying MARC21 856s that don't have $3)
* Use of uninitialized value in multiplication (*) at catalogue/detail.pl line 243.
(attemping to calculate average Amazon rating if no Amazon info
is available)
* opac-detail.pl: Use of uninitialized value in length at C4/External/Amazon.pm line 90
(attempting to calculate length of $isbn if it happens to be undef)
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
* Use of uninitialized value in pattern match (m//) at C4/Biblio.pm line 1305
(displaying MARC21 856s that don't have $3)
* Use of uninitialized value in multiplication (*) at catalogue/detail.pl line 243.
(attemping to calculate average Amazon rating if no Amazon info
is available)
* opac-detail.pl: Use of uninitialized value in length at C4/External/Amazon.pm line 90
(attempting to calculate length of $isbn if it happens to be undef)
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
The intranet item display provided by moredetail.pl
would display the original due date of the item, and
would thus be incorrect if the item had been renewed
and got a different due date.
The incorrect date was coming from items.onloan,
which AddIssue() sets to the due date. This patch
fixes the bug in two ways:
[1] AddRenewal() now updates items.onloan with the
correct due date.
[2] Two templates were updated to display the
due date from (indirectly) issues.date_due
instead of items.onloan, thus making it less
likely that the wrong value will be displayed.
This patch does *not* update items.onloan to reflect
the correct due date.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
This patch depends on my previous one "Begin cleanup on additem" and
fixes one bug introduced in that patch.
Overhaul of additem sections to factor out common pieces especially attributes.
Plugin js vastly simplified and reimplemented in jquery. Performance improved.
Note that a particularly busy cataloging department should still leave
barcode blank for server-side autopopulation to avoid collision.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
REFACTOR logic out of conditional branches when the assignment ($nextop) is the same.
Be sure to EXIT after printing a redirect, instead of wasting clock filling in the
template for a process the user will never see. Remove the now inapplicable logic for
differentiation between "/cgi-bin" and "non /cgi-bin" installations. There is no
longer any /cgi-bin in the actual directory file path.
This does not fix the value_builder js errors like "Blurbarcode52251 is not defined",
but it does not cause them either (see Bug 2919).
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
Auto truncation now works on LTR languages
For RTL languages(Arab, Hebrew), development should be really improved.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
The re-creation of statement handles for the same queries, repeatedly,
seriously degrades the performance of notices. This introduces new
class-level caching to avoid that problem.
Note: initial provisions
to do the same with caching of COLUMNS information about tables is
also included, but this would probably be most usefully implemented
in C4::Context (or a separate more central module). The benefits of
caching that info would probably be even greater. Note that this
proposed implementation might be obsoleted by ORM integration, where
DB structure info would already be cached.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
GetItemsInfo() returns a count_reserves variable which may be "Waiting" or "Reserved," but opac-detail.pl didn't include a check for these possibilities. This patch has opac-detail.tmpl output "On hold" in either case.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
This is half of the fix for 3090.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
It looks like a lot of change, but it isn't much:
perltidy, removed unused "toggle" code, and corrected comments.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
Convert to GetBranchesLoop, remove unused variables.
Pull out params that are assigned the same value in each loop and just assign them once.
Pull out params that are assigned the same value in each half of the conditional, and just assign them unconditionally.
Remove useless call of $template->param($count).
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
Corrects case of "onclick" and creates new style for "send" link rather than re-using inapplicable toolbar ID.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
Per discussion I had with Henri, removing experimental
bulk editing from the staff search results code, as
feature is incomplete and can be dangerous if
one tries to use it on a large search result
set.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
If series title was previously taken as collectiontitle
then you should launch
misc/maintenance/UNIMARC_fix_collectiontitle.pl
misc/batchRebuildBiblioTables.pl
[RM note: applies only to UNIMARC users - MARC21
users should not run the batch jobs.]
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
This fix is in circulation.tmpl instead of the search box include. That eliminates the clash with returns.pl. Focus works correctly when checking out, when circulation.pl is at its "empty" state, and when patron selection form is displayed.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
Previous patch broke this, always providing the
guided report download in CSV format regardless of
whether CSV or tab-delimited was requested.
Also, default name of downloaded file is reportresults.csv
for CSV output and reportresults.tab for tab-delimited.
Also removed a couple subroutine prototypes that
were leading to unnecessary compile-time warnings
under use warnings.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
execute_query now refactored, returns reliable results, does
zero presentation-layer crap. Arguments reduced, client scripts
adapted to new API and performance improved. Text::CSV now used
to generate CSV output, ensuring portability, encoding and accuracy.
Replaced tools/runreport.pl with misc/cronjobs/runreport.pl:
~ security fixed
~ documentation improved
~ TODO: finish sendmail option.
Bug 3077 also fixed.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
Use map to replace multiple temporary hash constructions.
Also a minor POD fix.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
The old location of runreport.pl was under tools, leaving it exposed
to web requests. This is a security flaw since it does NOT check any
Auth and allows the user to request any Saved Report be run. This is
not a problem under misc/crontab/ and it suggests the more appropriate use.
Guided.pm is not fixed here (see bug 3066), but it is prepared to be fixed
and made compatible with runreport as detailed in the perldoc.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
With this patch:
* On Amazon book cover, 'search inside' text is not displayed
anymore because it's not translatable.
* Link the image to local Amazon web site 'search inside' service.
Based on AmazonLocale syspref.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
These changes tidy up ISBN and ISSN indexing, per Michele Maenpaa. It's being
set up manually on many new installations, and probably ought to become part of the default
Koha installation.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
Adds three new switches:
-idmap <filename> - optional output file of
map of source record ID numbers
to Koha biblionumber
-x - if idmap is supplied, MARC tag
to get source record ID from
-y - if idmap is supplied, MARC subfield
to get source record ID from
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
Recommended by Michele Maenpaa, this adds handling colon and end-punctuation stripping
in subroutine _normalize, and fixes length testing in subroutine _get_match_keys
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
Added ISBN tag to RSS results, right after the title.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
this patch add a new column and put the isbn in it.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
This includes a change to Search.pm to add a variable, subtitle_nospan, in order to display subtitles in the feed without search term highlighting HTML. The modified template attempts to display Amazon, Syndetics, or Baker & Taylor book cover images. Google images are not included because they require Javascript.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
This XSL transforms XML UNIMARC biblio record into
its DC representation. It is the equivalent of
MARC21slimOAIDC.xsl.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
This OAI server has those characteristics:
* It doesn't use C4::OAI anymore.
* It uses HTTP::OAI library which is easily expandable.
* XML result is not build anymore directly but using
XML::SAX::Writer. It should solve encoding and XML related issues.
* It returns two formats: oai_dc and marcxml.
* DC is generated from biblioitems.marcxml by applying XSLT
(MARC21slim2OAIDC.xsl or UNIMARCslim2OAIDC.xsl).
* It uses the same tokens as previous OAI server (thanks Philippe
Jaillon).
* There is no more sets (it wasn't operational in previous version).
Add two new dependencies:
* HTTP::OAI
* XML::SAX::Writer
available as a Debian package: libxml-sax-writer-perl
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
This patch modifies z3950 search result page in order to
use jquery to highlight rows rather than an explicit
css class.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>
In UNIMARC, 010 tag contains ISBN. In UNIMARC, ISBN with '-'
are valid. This patch modify ISBN normalization in order to
automaticaly suppress '-' before sending ISBN to Amazon
to get book cover.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Henri-Damien LAURENT <henridamien.laurent@biblibre.com>