Added in section about checking the release notes for other upgrade steps to the end.
Added in the step to restart memcached as Galen Charlton mentioned in his 11/5 dev list email
about clearing out old login sessions for security reasons.
Added in Jessie and Trusty with a pointer to the Ruby EDI bug that users of the distributions
will run into.
Signed-off-by: Josh Stompro <stomproj@larl.org> Signed-off-by: Ben Shum <bshum@biblio.org>
Dan Wells [Tue, 8 Apr 2014 20:22:33 +0000 (16:22 -0400)]
LP#1078593 Regenerate summaries when deleting issuances
Both interfaces currently allow you to delete issuances even if they
have items attached and have the delete cascade. While this might be
a dangerous allowance, we should still make sure to keep our summaries
updated when this happens.
This commit adds automatic summary regeneration when deleting issuances
in either serials interface.
Signed-off-by: Dan Wells <dbw2@calvin.edu> Signed-off-by: Kathy Lussier <klussier@masslnc.org> Signed-off-by: Ben Shum <bshum@biblio.org>
Dan Wells [Wed, 10 Apr 2013 22:42:11 +0000 (18:42 -0400)]
LP#1078593 Add method for regenerating serial summaries
Right now, serial summaries only update when receiving or
resetting items. They need to update more often, so lets start
by adding a method just for doing that.
Signed-off-by: Dan Wells <dbw2@calvin.edu> Signed-off-by: Kathy Lussier <klussier@masslnc.org> Signed-off-by: Ben Shum <bshum@biblio.org>
1) fleshed_issuance_alter() changes
- First, if one of the inner updates returns an event, abort early.
Otherwise, we might overwrite the event while looping and lose
the error.
- Second, add authtoken to the created editor. This editor gets
passed around, and other functions might look for it there.
2) pass $type down into _summarize_contents()
- Right now, we have a bug where supplement/index contents in a
merged MFHD record will get added to the basic contents (and
vice-(vice-)versa). By passing in $type, we can assure that
we only return the type of contents we are looking for.
3) _prepare_summaries() should acknowledge empty summaries
- Under normal use, summaries will generally grow, and never shrink
to nothing. However, we must still handle cases where we have
discarded/deleted our last issuance, and in those cases, make sure
we "empty" the generated_coverage field in our summary.
Signed-off-by: Dan Wells <dbw2@calvin.edu> Signed-off-by: Kathy Lussier <klussier@masslnc.org> Signed-off-by: Ben Shum <bshum@biblio.org>
Galen Charlton [Wed, 28 Jan 2015 17:24:02 +0000 (17:24 +0000)]
LP#1415572: ensure correct version of authority.normalize_heading() is in place
Some upgraded Evergreen databases can have the wrong version
of the authority.normalize_heading() function in place because
of a mixup in the 0848 schema upgrade script. The problem can
manifest itself by there being rows in authority.record_entry
where the heading and simple_heading columns have "NOHEADING"
values regardless of the control set:
This patch ensures that the correct version is in place and
updates authority.record_entry to fix up affected rows without
requiring a full authority ingest.
Signed-off-by: Galen Charlton <gmc@esilibrary.com> Signed-off-by: Chris Sharp <csharp@georgialibraries.org> Signed-off-by: Ben Shum <bshum@biblio.org>
Update search metadata without a reingest by identifying records that
use a space in the audience field, and be sure that the coded value map
has a space for "uncoded" audience.
Signed-off-by: Mike Rylander <mrylander@gmail.com> Signed-off-by: Ben Shum <bshum@biblio.org>
Galen Charlton [Mon, 26 Jan 2015 16:05:10 +0000 (16:05 +0000)]
LP#1414112: avoid excluding record attribute values that contain only blanks
Certain record attributes, chiefly the ones whose values comes
from MARC fixed fields and which have a coded value map associated
with them, can have a string consisting of one or more blanks as a
valid value. Consequently, this patch ensures that
reingest_record_attributes() no longer excludes all attributes that
have blank values from the final attribute list.
This fixes a problem where MARC records with the target audience (008/22)
coded as a blank (unknown or not specified) could no longer be retrieved
using an "audience( )" search filter.
After applying this patch, a reingest of record attributes should be
performed, e.g., by doing
select metabib.reingest_record_attributes(id)
from biblio.record_entry
where not deleted;
Signed-off-by: Galen Charlton <gmc@esilibrary.com> Signed-off-by: Mike Rylander <mrylander@gmail.com> Signed-off-by: Chris Sharp <csharp@georgialibraries.org> Signed-off-by: Ben Shum <bshum@biblio.org>
Jason Stephenson [Thu, 29 Jan 2015 13:02:03 +0000 (08:02 -0500)]
LP#1415898 Fix Potential Cstore Starvation in the OPAC
When a user views their current circulations and/or renews them via
the OPAC, a connection is made to a cstore drone so that the user's
circulations can be retrieved from the database, bypassing the cache.
This connection is not closed, and thus a cstore drone is left unable
to respond to new requests for several seconds. On a busy server,
this has actually caused the number of cstore drones to hit the limit
configured in opensrf.xml. If that situation continues for too long,
it leads to the load on the server spiking to the point where services
must be restarted.
This commit changes from xact_rollback to rollback so that the connection
will be closed.
Signed-off-by: Jason Stephenson <jason@sigio.com> Signed-off-by: Bill Erickson <berickxx@gmail.com>
Jason Stephenson [Thu, 29 Jan 2015 12:59:07 +0000 (07:59 -0500)]
LP#1415898 Fix Potential Cstore Starvation in KPAC
The getit_results function in EGKPacLoader.pm does xact_begin, which
opens a connected session to a cstore drone. It later does xact_rollback,
which does not disconnect the session. This leaves a drone hanging for
several seconds unable to responde to new requests. This commit switches
to a rollback call that does disconnect the session.
Signed-off-by: Jason Stephenson <jason@sigio.com> Signed-off-by: Bill Erickson <berickxx@gmail.com>
Ben Shum [Wed, 21 Jan 2015 21:30:53 +0000 (16:30 -0500)]
LP#1412893: Only apply lost and paid status with the proper transactions
As discovered while testing the lost and paid feature, when closing earlier transactions where
the item is currently lost, it could change the item status to lost and paid prematurely.
Change the logic so that it only does this if the current transaction being closed either has
a status of lost or longoverdue.
Signed-off-by: Ben Shum <bshum@biblio.org> Signed-off-by: Kathy Lussier <klussier@masslnc.org>
Dan Scott [Mon, 12 Jan 2015 18:13:03 +0000 (13:13 -0500)]
LP#1409844: Towards more meaningful catalogue <title> elements
Currently the catalogue <title> elements are formed by concatenating the page
title to "Catalog - ". This means that on browsers with more than a few tabs
open, users will only see "Catalog -", which doesn't help them navigate their
tabs. Also, in search results in general search engines, "Catalog - " is
expected to be a significant word as it appears first in the title of the page.
Instead, let's remove the "Catalog - " prefix entirely and add "- <library
name>" as a suffix to the page title. That way, a user will see the page title
in bookmarks, search results, and other contexts with the most important
information up front (whether that's the title of the work in record details,
or the fact that it's a search for some string), followed by the name of the
library to give it a specific context.
In addition, use the name of the library in the OpenSearch definition to be
more meaningful than the current hardcoded "Evergreen OpenSearch" (many users
will not even know what Evergreen is, if they stumble across the little "Add
search engine" widget at all).
And finally, use the HTML and URI filters appropriately to avoid any possible
damage :)
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Ben Shum <bshum@biblio.org>
Remington Steed [Wed, 26 Nov 2014 14:21:52 +0000 (09:21 -0500)]
LP#1396161: Improve public docs and change osrf-config default from script help
The public docs have an "Action Triggers" chapter, so this commit
applies the same improvements there to match the help docs in the
script. Also in the public docs, in two places a filename was split
across lines, causing a gap to appear in the HTML version of the docs.
That is corrected here. This commit also removes the 'osrf-config'
option from the examples in the script help docs, since they are using
the default value.
Signed-off-by: Remington Steed <rjs7@calvin.edu> Signed-off-by: Ben Shum <bshum@biblio.org>
- Removed --granularity-only help docs. Ever since the script was changed to
support parallel execution of different granularity levels the granularity-only
setting has always been set when using --granularity.
- Added --verbose help docs. This option wasn't included in the help docs.
- Clarification - added some clarification to a few of the help descriptions.
- Added an additional example for processing hooks and running pending events
with a granularity set.
- Removed two references to the --granularity-only flag in the Telephony docs.
I believe this change is valid for 2.7, 2.6 and 2.5 since the change to the script
was made in 2010.
Ben Shum [Tue, 16 Dec 2014 19:34:12 +0000 (14:34 -0500)]
LP#1373594: Also ignore subfield 7 in get_graphics_880s
In the Czech Republic the subfield 7 of MARC fields 1XX and 7XX are used for
the National authority control number.
The problem is, that in current Evergreen versions, the subfield 7 is displayed
as a part of the link to author seach in search result or record details, but
when using the linkt to search rcords of the same author, no results are
returned, no matter whether the records exist or not.
That is why we nedd to add the subfield 7 into get_graphic_880s.
Signed-off-by: Ben Shum <bshum@biblio.org> Signed-off-by: Kathy Lussier <klussier@masslnc.org>
Chris Sharp [Mon, 21 Apr 2014 13:10:28 +0000 (09:10 -0400)]
LP#1310619: Add links in reporter.classic_item_list to legacy_cat1 and legacy_cat2 views.
Currently, the reporter.classic_item_list view has the Legacy Cat 1 and 2
fields defined as "text", not "link", so all it will report out is the ID
number of the stat cat. We want nullability control over stat cat entries,
so these need to be "link".
Signed-off-by: Chris Sharp <csharp@georgialibraries.org> Signed-off-by: Ben Shum <bshum@biblio.org>
Dan Wells [Thu, 8 Jan 2015 21:30:59 +0000 (16:30 -0500)]
LP#1408803 Add 'persist' login type to auth_proxy sample config
A recent version of EG added the option to "stay logged in", and this
was implemented as a new login type. Since basically everyone will use
this option via the OPAC, it makes sense that this should be part of the
sample config for auth_proxy.
Signed-off-by: Dan Wells <dbw2@calvin.edu> Signed-off-by: Dan Scott <dscott@laurentian.ca>
LP#980296: Update void on claims returned for longoverdue status.
We add settings for voiding longoverdue fees on claims returned.
We modify code in Circ.pm to check those settings and void those
bills when a circulation is claims returned.
We also make some logical repairs to the new void_lost function
in CircCommon.pm and to event handling where that function is
called in set_circ_claims_returned in Circ.pm.
Signed-off-by: Jason Stephenson <jason@sigio.com> Signed-off-by: Kathy Lussier <klussier@masslnc.org> Signed-off-by: Ben Shum <bshum@biblio.org>
Add additional packages to the "packager" target to help support
building Evergreen packages on machines where Evergreen and its prereqs
are not installed.
The "packager" target now depends internally on the "developer" target,
since packagers have to build the browser client dependencies.
Since older OSes have to build the nodejs dependencies, the autoconf,
etc. targets have been moved into the developer target, again to support
"headless" packaging.
Signed-off-by: Bill Erickson <berickxx@gmail.com> Signed-off-by: Ben Shum <bshum@biblio.org>
* Adds <osname>-developer target for installing packages only needed by
developers installing from source. Currently, this is limited to
packages needed for managing browser client dependencies, but others
could be added later.
* Adds <osname>-packager target for installing packages needed to build
Evergreen release bundles.
* Updates README install docs to reference these new targets.
* Updates README install docs to remove requirements for installing
libtool/automake/autoconf, since those packages will have been
installed during the OpenSRF install (which is required).
* NOTE: this does not include new Makefile.install targets for fedora
Signed-off-by: Bill Erickson <berickxx@gmail.com> Signed-off-by: Ben Shum <bshum@biblio.org>
Per bug # 1046149, we added a script that will generate the contents of the
opac/parts/relators.tt2 file to provide more granular credits for added entries
that have relator codes. However, we haven't run that script since 2013, and
as there are a number of new relator codes, along with some revisions to the
existing relator terms, we should update relators.tt2 accordingly.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Ben Shum <bshum@biblio.org>
Removes the Bib Call Number search as a default numeric search in the
catalog. Release notes include instructions for restoring it for those
sites that wish to restore it.
Dan Wells [Thu, 16 Jan 2014 20:07:40 +0000 (15:07 -0500)]
Changes to smooth out make_release
1) Make the fake 'HeadURL' into a URL which works with our web repo.
Note that this has 'tags/' hardcoded in, so it only works with true
release branches (e.g. rel_x_y_z) not "working" branches
(e.g. rel_x_y).
2) Change range selector when building log to better exclude commits
from previous releases.
Signed-off-by: Dan Wells <dbw2@calvin.edu> Signed-off-by: Bill Erickson <berickxx@gmail.com>
Bill Erickson [Mon, 29 Dec 2014 19:04:45 +0000 (14:04 -0500)]
LP#1406367 Fine generator skips no-fines transactions
Filter out any circs or booking reservations which will never be billed
in the initial overdue transactions query used by the fine generator.
This allows us to avoid loading rows into memory that the fine generator
will ignore anyway.
Signed-off-by: Bill Erickson <berickxx@gmail.com> Signed-off-by: Jason Stephenson <jstephenson@mvlc.org>
Avoid loading the full set of billable transaction objects into
memory when generating fines to reduce the memory footprint of the
open-ils.storage fine generater API. Instead, load transaction IDs
on startup, then fetch individual transactions as needed.
Signed-off-by: Bill Erickson <berickxx@gmail.com> Signed-off-by: Jason Stephenson <jstephenson@mvlc.org>
Michael Peters [Mon, 8 Dec 2014 17:08:21 +0000 (12:08 -0500)]
LP#1319964 "Summaries & More" displays improperly in TPAC
Commit 14eb3d30 added HTML-escaping of OPAC extras display
labels, meaning that if an ampersand in such a label is
desired, it should be entered in the template just as an ampersand.
Signed-off-by: Michael Peters <mpeters@emeralddata.net>
[Amended original commit to tweak message and remove unecessary changes
to .po files, which are rebuilt when a release is rolled.]
Per http://www.loc.gov/marc/bibliographic/bd024.html we
should be checking to see if either the ind1 is "1" or
if there's a subfield 2 with a value of "upc" before
assuming that any "Other standard identifier" 024 field
is a UPC value.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Ben Shum <bshum@biblio.org>
Mike Rylander [Fri, 18 May 2012 12:30:08 +0000 (08:30 -0400)]
LP#921142: Bring back automatic padding of short fixed fields in the GUI MARC editor
Bring back automatic padding of short fixed fields in the GUI MARC
editor. Also, use onblur instead of oninput, so that we've already lost
focus and don't care if that happens again.
Signed-off-by: Mike Rylander <mrylander@gmail.com> Signed-off-by: Remington Steed <rjs7@calvin.edu> Signed-off-by: Ben Shum <bshum@biblio.org>
Dan Wells [Fri, 4 Oct 2013 14:02:39 +0000 (10:02 -0400)]
LP#1389403 Add normalizer detection to call number browse
When finding the pivot for call number browse, we currently rely on
simple label matching (rather than normalized label matching), since
we do not know at the time of the search which normalizer we should
use.
Rather than blindly using the label, let's select the pivot in two
phases. First, we will do what we do know, getting as close as
possible by label alone. If we find an exact match, we are done (same
as the current code). If we don't find an exact match, we take the
normalizer setting from the closest label match, normalize our input,
then search again.
This method *greatly* improves the overall accuracy of call number
searching whenever the call number being browsed for does not exist,
while causing no harm to accuracy when it does exist (since it will
either match the label or normalize to the same call number anyway).
Eventually, we may want to add interface elements for selecting the
call number type, probably defaulting to the default call number type
for the org unit being searched, but even with that in place, this
"auto-detection" logic should still be valuable.
Signed-off-by: Dan Wells <dbw2@calvin.edu> Signed-off-by: Kathy Lussier <klussier@masslnc.org> Signed-off-by: Ben Shum <bshum@biblio.org>
My simplistic approach is to check to see if the link text == the href,
and if so, and we have a note to display, to use the note as the text of
the link instead.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Ben Shum <bshum@biblio.org>
Dan Scott [Sun, 7 Dec 2014 13:51:33 +0000 (08:51 -0500)]
LP#1400082 - avoid uninitialized var warning at TPAC login
Give username a default value of '', which makes the regex happy (avoiding an
uninit warning in Apache logs if no value was entered) and still counts as de
nada for the purposes of the actual login dance.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Ben Shum <bshum@biblio.org>
Jason Stephenson [Sat, 29 Nov 2014 15:34:53 +0000 (10:34 -0500)]
LP#1401271: Add missing dependency for Date::Manip perl module
This commit adds lines to distro-specific Makefiles to install
Perl Date::Manip packages so that Date::Manip::Date is
avaliable for marc_export. Somehow, this got missed or removed.
Signed-off-by: Jason Stephenson <jstephenson@mvlc.org> Signed-off-by: Ben Shum <bshum@biblio.org>
Dan Scott [Sun, 7 Dec 2014 15:17:42 +0000 (10:17 -0500)]
LP#1400100 - Avoid extra </div> when OpenURL is enabled
When the OpenURL resolver is enabled, the record summary.tt2 page always
inserts a </div>, whether or not there actually are any OpenURL results to
point at (which is the condition that triggers the opening <div>).
This has the effect of resulting in the generated RDFa not including anything
below the OpenURL section, effectvely preventing the schema.org/Offer types
from being associated with the record. Which is not good for the semantic web
of library data.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Ben Shum <bshum@biblio.org>
Remington Steed [Fri, 5 Dec 2014 15:29:48 +0000 (10:29 -0500)]
Docs: Web Client preview intro
This commit adds an introduction for the Web Client preview. It explains
what to expect as development continues, and where to give feedback. It
also includes the rationale from the 2.7 release notes.
Revised instructions based on my testing. Swapped around the Monster
and Default skin discriptions since the listed Alternate Skin is the default skin.
Kathy Lussier [Tue, 5 Feb 2013 15:24:17 +0000 (10:24 -0500)]
Add KPAC configuration info to the community docs
Adds the KPAC TechRef docs to root.txt so that this information can be
available via the community docs. Also adds a section to the docs on
configuring the alternate skin and removes a reference to configuration
from the end-user KPAC docs.
Fredric T Parks [Wed, 6 Nov 2013 23:36:32 +0000 (15:36 -0800)]
LP#1246839: marc_stream_importer.pl no longer crashes with vs 0.23 of File::Temp
updated marc_stream_importer.pl to use the 0.23 API
Signed-off-by: Fredric T Parks <fparks@catalystitservices.com> Signed-off-by: Bill Erickson <berickxx@gmail.com> Signed-off-by: Galen Charlton <gmc@esilibrary.com>
Generate overdue fines for lost checkins (if configured) within the main
body of the checkin transaction to avoid complications with negative
balance reconciliation.
Signed-off-by: Bill Erickson <berickxx@gmail.com> Signed-off-by: Dan Wells <dbw2@calvin.edu>
Mike Rylander [Thu, 25 Sep 2014 16:58:45 +0000 (12:58 -0400)]
LP#1198465 Allow fine generator to respect a stop_fines filter
In order to support situations where billings should be generated on
specific circs, even when stop_fines is set, we allow a filter to
be passed. The default of not passing a filter is the same as requiring
that stop_fines be NULL, which is the pre-commit behavior.
We also remove a never-used parameter.
Signed-off-by: Mike Rylander <miker@esilibrary.com> Signed-off-by: Dan Wells <dbw2@calvin.edu>
projects / working / Evergreen.git / log