Dan Scott [Thu, 11 Sep 2014 01:34:36 +0000 (21:34 -0400)]
At Laurentian, all usernames will be lowercase
This will prevent mismatches like the user entering "dscott" and the database
having the value "DScott" (in that we're going to force all of the Laurentian
user names in the database to lower case).
Art Rhyno [Wed, 23 Oct 2013 15:21:03 +0000 (11:21 -0400)]
Minor changes to HSN Tpac skin
As per ticket 1011, changed links for PubMed and Library Home Page,
as well as modifying links to open in new window. Also removed link
to "old catalogue" since it redirects to "new catalogue" anyway.
Mike Rylander [Thu, 12 Sep 2013 17:13:23 +0000 (13:13 -0400)]
Optimize container filters in QueryParser
When a container filter is used at the top level of a QueryParser
query (that is, in a simple query with no OR-logic branches or explicit
nesting groups) we can use an INNER join instead of LEFT + IS NOT NULL.
On some production datasets this showed an increase in performance from
20+ seconds to 1ms for the core query.
Signed-off-by: Mike Rylander <mrylander@gmail.com> Signed-off-by: Dan Scott <dscott@laurentian.ca>
If an overdue is returned after a closed date, the generate_fines code
will not generate fines for those items because the section of code that
checks for closed dates returns from the function if a closed date is
encountered.
For example, if an item is due on June 29th, but is not returned until
after July 1st (Canada Day), then when fines are generated for the item,
generate_fines creates a fine for the 30th of June, but when it
gets to the 1st of July it encouters a closed date and executes a
'return' statement which exits the generate_fines code causing the
fine from June 30th to be rolled back as well as preventing further
fines from being created.
This fix replaces the 'return' statements inside the 'for' loop
that is nested within the eval with 'next' statements.
Signed-off-by: Liam Whalen <whalen.ld@gmail.com> Signed-off-by: Dan Scott <dscott@laurentian.ca>
Mike Rylander [Tue, 25 Jun 2013 11:44:22 +0000 (07:44 -0400)]
Only attempt to map copies once per hold
The development of Calculated Proximity Adjustment (CPA) for 2.4 added a "prox"
column to action.hold_copy_map (ahcm) to store on variant of copy proximity. That
value is calculated at the same time that certain other data structures are
set up, and is conceptually tied to them, so the resulting combined data
set is used to insert values into the ahcm. However, user configuration
can cause that data set to contain non-unique values, where the data structure
used previous to CPA was guaranteed to be unique. This lack of uniqueness
would then violate a constraint on the achm table. This commit restores the
required uniqueness for the purpose of creating the ahcm rows.
Bill Erickson [Mon, 13 May 2013 17:22:37 +0000 (13:22 -0400)]
LP1179609 ACQ lineitem identifier inline update
When the identifier value for a linetiem is changed, apply the change
and refresh the lineitem in real time instead of requiring users to
click the 'Apply Lineitem Identifier' link.
The 'Apply Lineitem Identifier' link has also been removed, since it is
no longer needed.
Signed-off-by: Bill Erickson <berick@esilibrary.com> Signed-off-by: Kathy Lussier <klussier@masslnc.org>
Dan Scott [Mon, 24 Jun 2013 14:12:50 +0000 (10:12 -0400)]
Avoid Z39.50 search warning for uninit var
The debug log blindly attempts to access list members that might not
exist (if, for example, an event code was returned from the attempt to
run do_service_search()), thus generating spurious log warnings.
Instead, move the debug line to only generate output when we have
created the list item in question.
Dan Scott [Fri, 21 Jun 2013 18:49:27 +0000 (14:49 -0400)]
Prevent JavaScript error on non-Novelist sites
Unless the Evergreen site in question is signed up and configured to
support Novelist added content, every details page will generate a
JavaScript error due to the reference to an undefined JavaScript
function. Guard against that by testing for the presence of the Novelist
URL.
Dan Scott [Fri, 21 Jun 2013 18:49:27 +0000 (14:49 -0400)]
Prevent JavaScript error on non-Novelist sites
Unless the Evergreen site in question is signed up and configured to
support Novelist added content, every details page will generate a
JavaScript error due to the reference to an undefined JavaScript
function. Guard against that by testing for the presence of the Novelist
URL.
Dan Scott [Fri, 21 Jun 2013 03:39:21 +0000 (23:39 -0400)]
return, not next, from eval BLOCK
Per perldoc eval, exiting an eval BLOCK is not allowed. And it generates
warnings - 37,000 of them in one day - on a system on which the fine
generator runs every 15 minutes. Simply returning from the block
achieves the same goal without generating scads of warnings.
Dan Scott [Fri, 21 Jun 2013 03:11:20 +0000 (23:11 -0400)]
Silence uninit var warnings from query normalizer
$field_info can return undef from the search for a matching ID from the
list of normalizers, which in turn generates several uninit var warnings
when attempts are made to access members of the undef variable. We saw
6700 occurrences of these warnings during a typical day of operations.
Dan Scott [Thu, 20 Jun 2013 16:55:30 +0000 (12:55 -0400)]
Quick delete UI
For mass deletion efforts, go to http://hostname/eg/cat/moveto/delete in
a web browser and start scanning barcodes. Each barcode will be deleted
immediately. Note that call numbers and bib records will not be deleted
if the copy was the last one, but the records will no longer show up in
search results for regular patrons... and one could always follow up
with a clean up SQL job to delete call numbers and records where
appropriate.
Dan Scott [Wed, 19 Jun 2013 20:49:14 +0000 (16:49 -0400)]
Silence QP warning due to inappropriate cmp op
In QueryParser, we're comparing two strings, but using the numeric !=
operator. Switching to the 'ne' operator silences a warning that came up
7,000 today in our logs so far...
Dan Scott [Wed, 19 Jun 2013 19:45:36 +0000 (15:45 -0400)]
Fix logic for testing for OpenURL URIs
The TT2 logic to test for the existence of OpenURL URIs was different
in two different places, and the result was that non-OpenURL electronic
resources were not being displayed.
Dan Scott [Wed, 19 Jun 2013 15:48:40 +0000 (11:48 -0400)]
659 stands for "garbage", not "genre", in our MARC
Hide the display of 659 fields in Laurentian records. They appear to
have been used for some sort of identifiers, not genres, and are
meaningless to users.
Dan Scott [Tue, 18 Jun 2013 16:17:40 +0000 (12:17 -0400)]
Prevent uninit var warnings in Search/Biblio.pm
The logger messages in Search/Biblio.pm throw _many_ uninitialized
variable warnings. With a little care, we can prevent those warnings and
return our attention to more important log messages.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Ben Shum <bshum@biblio.org>
Dan Scott [Mon, 17 Jun 2013 04:38:43 +0000 (00:38 -0400)]
Support script-based circ in nearest_hold()
The nearest_hold() implementation recently changed to pull in the circ
library ID; however, if circulation runs through scripts, the circ
library on the copy object has been fleshed, and the nearest hold
routine gets very confused.
Check to see if we do, in fact, have an object instead of a plain
scalar, and assign the value to a placeholder variable accordingly.
Thanks to Jeff Godin for his great assistance on debugging this!
Dan Scott [Sun, 16 Jun 2013 13:19:09 +0000 (09:19 -0400)]
Prevent undefined var warning for username
EGCatLoader could generate undefined var warnings if submitted without a
username. Probably should just bail early rather than trying to log the
user in.
Dan Scott [Mon, 17 Jun 2013 04:38:43 +0000 (00:38 -0400)]
Support script-based circ in nearest_hold()
The nearest_hold() implementation recently changed to pull in the circ
library ID; however, if circulation runs through scripts, the circ
library on the copy object has been fleshed, and the nearest hold
routine gets very confused.
Check to see if we do, in fact, have an object instead of a plain
scalar, and assign the value to a placeholder variable accordingly.
Thanks to Jeff Godin for his great assistance on debugging this!
Dan Scott [Sun, 16 Jun 2013 13:19:09 +0000 (09:19 -0400)]
Prevent undefined var warning for username
EGCatLoader could generate undefined var warnings if submitted without a
username. Probably should just bail early rather than trying to log the
user in.
Dan Scott [Thu, 13 Jun 2013 05:16:23 +0000 (01:16 -0400)]
Retain index granularity with minimal bloat
The previous approach to reducing index bloat arguably went too far, in
that analytics such as separately catalogued songs for an album were all
simply aggregated together in a single metabib.title_field_entry row,
rather than being added as separated metabib.title_field_entry rows for
each unique value.
To avoid the original problem of exact duplicate rows being inserted, we
now check for an existing matching row before inserting into the index.
A good test record is title "Cello concerto." which results in 1 title
proper row and 4 added entry title rows in metabib.title_field_entry
after this change.
Mike Rylander [Wed, 12 Jun 2013 15:17:08 +0000 (11:17 -0400)]
Bring back "split-brain" indexing normalization
Before 2.4-era changes to indexing and search, normalizers were applied
in a way that resulted in different effects on the value and index_vector
columns of metabib.*field_entry tables. This behavior was lost during
the changes mentioned above, but we need it for several things to work
properly, phrase search and (likely) facets among them. This commit
brings that behavior back.
See here for the original behavioral documentation:
Bob Wicksall [Thu, 13 Jun 2013 05:06:10 +0000 (01:06 -0400)]
Reduce index bloat involving non-search_field values
Rows in metabib.title_field_entry, subject_field_entry,
series_field_entry, and author_field_entry are doubled or tripled due to
bad logic in biblio.extract_metabib_field_entry. This results in these
tables being 2 or more times their correct size.
This was introduced in 2.2.0 when the logic for browse_field and
facet_field were added to biblio.extract_metabib_field_entry. 2.1 is not
affected.
The duplicates are caused when biblio.extract_metabib_field_entry
returns TRUE in the search_field column for all rows even if they should
just be facet_field or browse_field after the first search_field value
is returned.
Signed-off-by: Mike Rylander <mrylander@gmail.com> Signed-off-by: Dan Scott <dscott@laurentian.ca>
Bill Erickson [Mon, 27 Aug 2012 14:38:48 +0000 (10:38 -0400)]
Repair fine generator memory leak
Calling "next" from within a "try" block results in a memory leak,
presumably because "try" is a tangled nest of subs and evals.
Replacing the "try" with a good ol' "eval" avoids the leak.
This particular leak in the fine generator is onerous when the fine
generator is run often (e.g. every 15 mins), which means circs that
have already been processed for the day are re-analzyed over and over,
causing the code to continue early (next) to the next loop iteration
for large numbers of circs. It also happens when a circs are skipped
because they have no fine interval, rate, or max fine.
You know this is happening because you will see something like this in
the storage stderr log:
Exiting eval via next at
/usr/local/share/perl/5.10.1/OpenILS/Application/Storage/Publisher/action.pm
line 820.
Exiting subroutine via next at
/usr/local/share/perl/5.10.1/OpenILS/Application/Storage/Publisher/action.pm
line 820.
This patch does not avoid the "exiting eval via next" warning, since
we're still next'ing out of the eval. It just avoids the memory leak
(and the "Exiting subroutine" warning). More extensive refactoring is
needed to to completely remove the second warning.
Signed-off-by: Bill Erickson <berick@esilibrary.com> Signed-off-by: Mike Rylander <mrylander@gmail.com>
Dan Scott [Thu, 13 Jun 2013 05:16:23 +0000 (01:16 -0400)]
Retain index granularity with minimal bloat
The previous approach to reducing index bloat arguably went too far, in
that analytics such as separately catalogued songs for an album were all
simply aggregated together in a single metabib.title_field_entry row,
rather than being added as separated metabib.title_field_entry rows for
each unique value.
To avoid the original problem of exact duplicate rows being inserted, we
now check for an existing matching row before inserting into the index.
A good test record is title "Cello concerto." which results in 1 title
proper row and 4 added entry title rows in metabib.title_field_entry
after this change.
Mike Rylander [Wed, 12 Jun 2013 15:17:08 +0000 (11:17 -0400)]
Bring back "split-brain" indexing normalization
Before 2.4-era changes to indexing and search, normalizers were applied
in a way that resulted in different effects on the value and index_vector
columns of metabib.*field_entry tables. This behavior was lost during
the changes mentioned above, but we need it for several things to work
properly, phrase search and (likely) facets among them. This commit
brings that behavior back.
See here for the original behavioral documentation:
Bob Wicksall [Thu, 13 Jun 2013 05:06:10 +0000 (01:06 -0400)]
Reduce index bloat involving non-search_field values
Rows in metabib.title_field_entry, subject_field_entry,
series_field_entry, and author_field_entry are doubled or tripled due to
bad logic in biblio.extract_metabib_field_entry. This results in these
tables being 2 or more times their correct size.
This was introduced in 2.2.0 when the logic for browse_field and
facet_field were added to biblio.extract_metabib_field_entry. 2.1 is not
affected.
The duplicates are caused when biblio.extract_metabib_field_entry
returns TRUE in the search_field column for all rows even if they should
just be facet_field or browse_field after the first search_field value
is returned.
Signed-off-by: Mike Rylander <mrylander@gmail.com> Signed-off-by: Dan Scott <dscott@laurentian.ca>
Bill Erickson [Mon, 27 Aug 2012 14:38:48 +0000 (10:38 -0400)]
Repair fine generator memory leak
Calling "next" from within a "try" block results in a memory leak,
presumably because "try" is a tangled nest of subs and evals.
Replacing the "try" with a good ol' "eval" avoids the leak.
This particular leak in the fine generator is onerous when the fine
generator is run often (e.g. every 15 mins), which means circs that
have already been processed for the day are re-analzyed over and over,
causing the code to continue early (next) to the next loop iteration
for large numbers of circs. It also happens when a circs are skipped
because they have no fine interval, rate, or max fine.
You know this is happening because you will see something like this in
the storage stderr log:
Exiting eval via next at
/usr/local/share/perl/5.10.1/OpenILS/Application/Storage/Publisher/action.pm
line 820.
Exiting subroutine via next at
/usr/local/share/perl/5.10.1/OpenILS/Application/Storage/Publisher/action.pm
line 820.
This patch does not avoid the "exiting eval via next" warning, since
we're still next'ing out of the eval. It just avoids the memory leak
(and the "Exiting subroutine" warning). More extensive refactoring is
needed to to completely remove the second warning.
Signed-off-by: Bill Erickson <berick@esilibrary.com> Signed-off-by: Mike Rylander <mrylander@gmail.com>
Mike Rylander [Tue, 11 Jun 2013 20:44:35 +0000 (16:44 -0400)]
Use the centralized initialization method for QP
QueryParser maintains an effective singleton for internal configuration of
the driver in order to avoid going to the database on every search request
to gather configuration parameters. There is also a centralized sub that
has the task of properly configuring said singleton. Previous to this commit
we were not using that centralized initializer in one case, and that case
happened to be the most important initialization of QP in all of Evergreen.
The fallout from that was that some configuration was not being gathered
and passed to QueryParser, resulting in incorrect search SQL being generated.
This is rectified by using the centralized initializer everywhere that it
is needed and can be used -- and particularly in the effected case.
Signed-off-by: Mike Rylander <mrylander@gmail.com> Signed-off-by: Dan Scott <dscott@laurentian.ca>
Mike Rylander [Tue, 11 Jun 2013 20:44:35 +0000 (16:44 -0400)]
Use the centralized initialization method for QP
QueryParser maintains an effective singleton for internal configuration of
the driver in order to avoid going to the database on every search request
to gather configuration parameters. There is also a centralized sub that
has the task of properly configuring said singleton. Previous to this commit
we were not using that centralized initializer in one case, and that case
happened to be the most important initialization of QP in all of Evergreen.
The fallout from that was that some configuration was not being gathered
and passed to QueryParser, resulting in incorrect search SQL being generated.
This is rectified by using the centralized initializer everywhere that it
is needed and can be used -- and particularly in the effected case.
Signed-off-by: Mike Rylander <mrylander@gmail.com> Signed-off-by: Dan Scott <dscott@laurentian.ca>
Mike Rylander [Tue, 30 Apr 2013 20:45:55 +0000 (16:45 -0400)]
Search clicked /and/ preceding sf values
When looking for authority records to control a bib field, we currently
search using exactly, and only, the subfield that was clicked. This
commit changes that so that the full field up to and including the clicked
subfield is used in the simple authority heading search we perform to find
the best-match (pivot) authority record.
Signed-off-by: Mike Rylander <mrylander@gmail.com> Signed-off-by: Lebbeous Fogle-Weekley <lebbeous@esilibrary.com>
Signed-off-by: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Mike Rylander <mrylander@gmail.com>
Conflicts:
Open-ILS/src/perlmods/lib/OpenILS/Application/Circ/Holds.pm
(Just whitespace conflicts on tab vs space -- repaired by hand.)
Ben Shum [Sat, 2 Mar 2013 07:21:58 +0000 (02:21 -0500)]
TPAC - set autofocus to user input when placing holds in staff client
When placing holds in TPAC via the staff client, the focus had been on the
basic search field above. Change this to have autofocus on the input field
for the hold user by default.
This can potentially save staff from having to perform an extra click to
select the input field. Now they can immediately begin entering the barcode
of the user they plan to create a hold on behalf of.
Signed-off-by: Ben Shum <bshum@biblio.org> Signed-off-by: Mike Rylander <mrylander@gmail.com>
Acq: Be fuzzy about case when retrieving EDI documents
The edi_fetcher relies on a routine in OpenILS::Acq::EDI to, among
other things, try to avoid fetching the same EDI document multiple times
when many rows in acq.edi_account refer to the same host and login
credentials.
Since in practice most vendors seem to run FTP servers for EDI on
Windows, not UNIX, and pathnames are therefore case-insensitive, that
test for other occurrences ought also to be case-insensitive.
If I were to look at this as a purist, I could argue that vendor servers
might sometimes by run on UNIX, and that for some reason it is possible
that different vendor-to-buyer EDI documents (order responses or
invoices) could have pathnames that differ only in the case of some
characters. But that seems wildly unlikely. If anyone does take this
possibility seriously, perhaps acq.edi_account needs a Boolean column to
indicate the remote host's case [in]sensitivity.
Signed-off-by: Lebbeous Fogle-Weekley <lebbeous@esilibrary.com> Signed-off-by: Ben Shum <bshum@biblio.org>
Dan Scott [Thu, 6 Jun 2013 13:40:59 +0000 (09:40 -0400)]
Fedora RPM for bzip2 headers is bzip2-devel
There was an attempt to add the right RPM for the bzip2 headers in
Makefile.install, but that failed on Fedora 19. Swap in bzip2-devel
instead to correct that.
Art Rhyno [Wed, 5 Jun 2013 16:24:52 +0000 (12:24 -0400)]
Add new University of Windsor logo to UWIN tpac
The university has unveiled a new logo for the campus and all
Departments are expected to use it. Right now, the "small" and standard
logo are the same but will do some resizing at the same time as changing
the colour combinations to better match the logo.
Signed-off-by: Art Rhyno <art632000@yahoo.ca> Signed-off-by: Dan Scott <dscott@laurentian.ca>
Bill Erickson [Thu, 23 May 2013 13:11:25 +0000 (09:11 -0400)]
LP1183340 selectivly apply editable funds sorting
Avoid sorting editable funds to the front of the fund list when there
are no org units at which the user has the ADMIN_ACQ_FUND permission.
Otherwise, we are sorting on an empty 'IN' list in pcrud, which results
in an error.
Signed-off-by: Bill Erickson <berick@esilibrary.com> Signed-off-by: Dan Wells <dbw2@calvin.edu>
Kathy Lussier [Sat, 25 May 2013 00:52:01 +0000 (20:52 -0400)]
Acquisitions Admin updates
Updates to the chapter on acquisitions administration, including 1) the
reintroduction of the Providers section, which appears to have dropped out
of the docs after 2.1 2) the new default # copies and improvements to
distributions formulas introduced in 2.4 and 3) minor copy editing.
Signed-off-by: Kathy Lussier <klussier@masslnc.org> Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Kathy Lussier <klussier@masslnc.org>
Ben Shum [Thu, 23 May 2013 20:32:49 +0000 (16:32 -0400)]
LP1081576 - fix utf8 characters in searches using portal page
Replace escape() with encodeURIComponent() on the portal page to allow utf8
characters to be passed more appropriately to the subsequent search in catalog.
Thanks Dan Scott for pointing this out.
Signed-off-by: Ben Shum <bshum@biblio.org> Signed-off-by: Pasi Kallinen <pasi.kallinen@pttk.fi>
Bill Erickson [Thu, 23 May 2013 13:43:52 +0000 (09:43 -0400)]
LP1183357 Pass locale to bib search (tpac)
Pass the TPAC locale value, which is collected at each page load, to the
bib search API call. This allows us to take advantage of language-based
relevance bumps.
Signed-off-by: Bill Erickson <berick@esilibrary.com> Signed-off-by: Pasi Kallinen <pasi.kallinen@pttk.fi> Signed-off-by: Dan Wells <dbw2@calvin.edu>
Dan Scott [Fri, 10 May 2013 17:12:55 +0000 (13:12 -0400)]
TPAC: Death date missing from authors in record details
The trailing period prescribed by ISBD punctuation for x00 $d subfields
in certain circumstances confounds the regex that was expected the $d
subfield to match ^\s*\d{4}-\d{4}\s*$. As a result, death dates were
often being omitted from the record detail display. Relaxing the regex
to match against a trailing .* instead of \s* resolves the issue.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Pasi Kallinen <pasi.kallinen@pttk.fi> Signed-off-by: Ben Shum <bshum@biblio.org>