Jason Etheridge [Fri, 14 Oct 2011 16:14:14 +0000 (12:14 -0400)]
LP873614 print_init error printing bill history
xulG.url_prefix needs to be present for anything using util/print.js and the
Mozilla print strategy. Alternately, we could duplicate the logic behind
url_prefix and bundle it into print.js :-/
Signed-off-by: Jason Etheridge <jason@esilibrary.com> Signed-off-by: Jason Stephenson <jstephenson@mvlc.org> Signed-off-by: Dan Scott <dscott@laurentian.ca>
Dan Scott [Fri, 14 Oct 2011 02:10:23 +0000 (22:10 -0400)]
Modernize settings-tester.pl a bit
* Add a section that checks for Evergreen services in opensrf_core.xml
* Ignore the qstore database credentials, as qstore is unused currently
* Split things off into subroutines for easier reading
Dan Scott [Thu, 13 Oct 2011 17:09:38 +0000 (13:09 -0400)]
Fix located URI merging in action.merge_record_assets()
Perhaps due to the unexpected results of the XPath expression, which
returns a count of nodes for subfields, etc, rather than just a count of
the number of 856 datafields, NULL was being concatenated onto a string
- which of course made the string NULL.
Use COALESCE() to prevent the NULLness from spreading and enable
action.merge_record_assets() to do the right thing.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Mike Rylander <mrylander@gmail.com>
Dan Scott [Wed, 12 Oct 2011 18:54:25 +0000 (14:54 -0400)]
Give OrgTree.js a locale, always
We started generating locale-specific OrgTree.js files many moons ago,
and now no longer provide a "default" locale in the root of the
common/js directory. The result was that the index.xhtml page for the
staff client was throwing an "_l is undefined" exception; largely
meaningless, as it seems that the OrgTree isn't even used on that page,
but noise that doesn't need to be in the JavaScript console.
Here we correct that exception by providing a locale, and hopefully
prevent any further copy/paste mistakes.
Ben Shum [Tue, 11 Oct 2011 22:51:32 +0000 (18:51 -0400)]
Change PostgreSQL 9 source for Ubuntu Lucid users.
Currently, Ubuntu Lucid does not have backport options for PostgreSQL 9.
Instead, most community members who use Lucid have been using
Martin Pitt's PPA as an external repository source. Signed-off-by: Ben Shum <bshum@biblio.org> Signed-off-by: Dan Scott <dscott@laurentian.ca>
Bill Erickson [Tue, 27 Sep 2011 20:14:59 +0000 (16:14 -0400)]
Read STDIN up to record separator then stop
The initial (and only known) target for the importer is OCLC Connexion
exports, which export 1 record at a time, with each record ending with
the standard record separator. There is no end-of-file /
end-of-transmission indicator, so the while() loop always blocks and
fails on the timeout alarm after the first record is read.
Until a more flexible solution is in place, roll back support for
reading multiple records via STDIN. The original commit was
Note, the feature being rolled back was not actually necessary to
implement the feature in the commit (file-based imports). It was just
an added bonus.
Signed-off-by: Bill Erickson <berick@esilibrary.com> Signed-off-by: Galen Charlton <gmc@esilibrary.com>
Dan Scott [Mon, 10 Oct 2011 20:46:48 +0000 (16:46 -0400)]
Give CentOS / RHEL a chance at installing PostgreSQL 9.0
Evergreen 2.1 requires PostgreSQL 9.0, and we accordingly updated the
package names back in the summer to point at PostgreSQL 9.0, but nobody
running CentOS / RHEL brought to our attention that the repo target
needed to be updated as well. This, although untested, at least gives
CentOS / RHEL users a fighting chance of installing Evergreen master /
2.1 on their system.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Galen Charlton <gmc@esilibrary.com>
Dan Scott [Mon, 10 Oct 2011 20:15:04 +0000 (16:15 -0400)]
Fix broken Net::Z3950::SimpleServer 1.14 install
Pin the install of Net::Z3950::SimpleServer to the
known-to-work-with yaz 3.0.x version, 1.12. 1.14 introduced dependencies
on yaz 4.1.x that can't be easily worked around.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Galen Charlton <gmc@esilibrary.com>
Galen Charlton [Thu, 6 Oct 2011 17:55:43 +0000 (13:55 -0400)]
fix authentication failure
Ensure that the auth_limit values are read when open-ils.auth
initializes, not during oilsAuthInit(). This solves a problem
observed on brick-based setups where authentication could
incorrectly fail because the drone that processes
open-ils.auth.authenticate.complete didn't have the the block
count value initialized yet.
Also fixed type nit when compare the fail count to the block
limit.
Signed-off-by: Galen Charlton <gmc@esilibrary.com> Signed-off-by: Bill Erickson <berick@esilibrary.com>
Really add *entire* selection list to a purchase order, when desired,
and generally do what a user would expect when, in Acq, the user
executes an operation on "all items".
This addresses LaunchPad bug #810030.
The fundamental issue is that an AcqLiTable, which is used in several
different ways in different interfaces, may be representing items on a
selection list (for example), but it doesn't necessarily know all of the
items that belong to that selection list. It only knows what it has drawn.
This addresses that deficiency by teaching the getSelected() method
some new tricks so that it can consult the server in the various ways
appropriate to getting a list of /all/ the items on a selection list, a
purchase order, a virtual PO, a set of search results, etc.
Signed-off-by: Lebbeous Fogle-Weekley <lebbeous@esilibrary.com> Signed-off-by: Mike Rylander <mrylander@gmail.com>
Dan Scott [Wed, 5 Oct 2011 17:12:30 +0000 (13:12 -0400)]
README: Explicitly identify accounts for commands
Add a preamble to the beginning listing all of the accounts referenced
in the instructions (Linux accounts, PostgreSQL accounts, Evergreen
administrator accounts). Also move the Developer Instructions to the
start to follow the natural order of operations.
Bill Erickson [Wed, 5 Oct 2011 19:16:17 +0000 (15:16 -0400)]
Avoid collecting facet data for non-facet fields
If a config.metabib_field (Admin -> Server Admin -> MARC Search/Facet
Fields) starts out as a facet field but is later changed, data for that
facet can still bubble up since it has already been extracted. This
patch prevents facets for config.metabib_field entries whose
facet_field=false from getting collected/cached for search results.
Signed-off-by: Bill Erickson <berick@esilibrary.com> Signed-off-by: Lebbeous Fogle-Weekley <lebbeous@esilibrary.com>
Dan Wells [Wed, 7 Sep 2011 15:35:18 +0000 (11:35 -0400)]
Clear org_unit cache fix for autogen.sh
autogen.sh is not properly clearing the org_tree cache. It is
unclear how this hasn't caused a problem before, but this fix
works for me and should be at least harmless in other enviroments.
Signed-off-by: Dan Wells <dbw2@calvin.edu> Signed-off-by: Lebbeous Fogle-Weekley <lebbeous@esilibrary.com>
Dan Scott [Wed, 5 Oct 2011 03:04:15 +0000 (23:04 -0400)]
i18n: Fix es-ES illegal use of '%' in lang.dtd
The '%' character is used to include previously defined blocks in DTD
files; would be nice if translate-toolkit gave a more obvious error to
alert us to this rather than saying the en-US doesn't validate and
dying.
Dan Scott [Sat, 20 Aug 2011 21:51:46 +0000 (17:51 -0400)]
Remove cache-generator.sh from Makefile.am
In a previous commit, we removed cache-generator.sh. Now we remove the
remaining reference to it in the Makefile.am that was generating an
error (reported by Ben Webb).
Dan Scott [Thu, 18 Aug 2011 19:32:53 +0000 (15:32 -0400)]
Update README to reflect the --create-database option
Now that eg_db_config.pl is intelligent enough to create the database,
update the README to free people from the tyranny of having to manually
issue database creation (and contrib-adding) commands.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Mike Rylander <mrylander@gmail.com>
Dan Scott [Tue, 4 Oct 2011 18:40:44 +0000 (14:40 -0400)]
Slightly updated about.html for rel_2_1
Not sure we want to actually keep listing individuals in this file -
there's a lot more credit to spread around - but at least I can bring it
somewhat up to date with current info.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Mike Rylander <mrylander@gmail.com>
1 - You have new DEBUG_CLIENT permission (after login)
This should work even on operator change.
Once a given debug interface is open you can change operator back.
2 - You have a client with the debug preference enabled
This causes the client in all places I touched to think you have DEBUG_CLIENT
This also allows debugging *before* login, including in the Standalone interface.
Debug functions I know of:
Debug box on login page (already protected by pref)
Debug items in Admin menu
Venkman
Javascript Shells
Inspector
Chrome List
CTRL+SHIFT+F7 debug box in overlays
Signed-off-by: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Jason Etheridge <jason@esilibrary.com>
Adds support for a print_custom.js file and an org unit specified file.
If either exists and contains a print_custom function it will be called
before printing to allow for DOM manipulation.
If neither exists the print_win.js file will call all functions it has
defined.
In addition, if any code sets do_print to false the window will close
instead of attempting to print.
The print_custom function will receive the type of template used, if any.
This only applies to "Receipt Template" based print jobs, and the type in
question is the "Type" that shows in the template editor.
Functions defined in print_win.js:
print_init
Main function, checks for print_custom and does final printing
print_do_sums
Does summing of values.
print_check_alt
Does swapping out of template blocks with display:none styling.
print_check_noprint
Disables printing under certain circumstances.
For more information about usage of the print_do_sums, print_check_alt,
and print_check_noprint functions check the comments in print_win.js.
Signed-off-by: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Jason Etheridge <jason@esilibrary.com>
Alter backend to check password period, not just for password changes.
Add form elements for asking for current password to JSPac.
Add handling for said form elements where needed.
Signed-off-by: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Bill Erickson <berick@esilibrary.com>
Amount of time seed is valid
Amount of time to keep failure count in memcache since last auth event
Number of failures before locking out auth attempts
Also, remove seed from memcache once it has been used once.
Signed-off-by: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Bill Erickson <berick@esilibrary.com>
link asset.copy_template.status to config.copy_status in the IDL
If nothing else, this provides a dropdown for status in the copy
template editor (the one for Serials, under the admin menus, not the
one on the copy editing interface).
Signed-off-by: Lebbeous Fogle-Weekley <lebbeous@esilibrary.com> Signed-off-by: Dan Wells <dbw2@calvin.edu>
Dan Scott [Tue, 16 Aug 2011 03:14:01 +0000 (23:14 -0400)]
Further autogen.sh cleanup
* Roll cache-generator.sh functionality into autogen.sh
* No longer install deleted scripts
* Teach Perl build infrastructure about OpenILS::Utils::Configure
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Thomas Berezansky <tsbere@mvlc.org>
Dan Scott [Mon, 15 Aug 2011 22:03:37 +0000 (18:03 -0400)]
Revamp autogen.sh to call functions from a Perl module
This cut moves all of the logic from the individual scripts that
autogen.sh used to call into a self-contained Perl module. We use
OpenILS::Utils::Cronscript to avoid much of the init dupe code.
Note that calling Perl functions from a bash script is not crazy pretty,
but perhaps in the long run autogen.sh becomes something else. This is
just a step towards consolidating those functions in a single module
while maintaining the same interface as before.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Thomas Berezansky <tsbere@mvlc.org>
Mike Rylander [Mon, 3 Oct 2011 19:01:26 +0000 (15:01 -0400)]
Correct Located URI visiblity
Located URIs should be visible when they exist at the search context org unit
or an ancestor thereof. However, before the change embodied in this commit,
they effectively acted exactly like copies, making records visible anywhere
within the context+depth range.
An example, using the stock example OU hierarchy:
* Before this change, a Located URI with an owner of BR1 would cause its
record to show up in a search at (or scoped to via a depth parameter) BR1,
SYS1 or CONS.
* Before this change, a Located URI with an owner of SYS1 would NOT cause its
record to show up in a search at (or scoped to via a depth parameter) BR1!
* After this change, a Located URI with an owner of BR1 will cause its record
to show up only in a search with a context OU of BR1, irrespective of depth
scoping.
* After this change, a Located URI with an owner of SYS1 WILL cause its
record to show up in a search with a context OU of BR1 or SYS1, regardless of
depth scoping.
Signed-off-by: Mike Rylander <mrylander@gmail.com> Signed-off-by: Dan Scott <dscott@laurentian.ca>
Dan Wells posted a fix to the DEV list for receipt printing that allowed the
receipt printer to print UTF-8 characters. I used the ideas in that fix to
modify spine_labels.js. In spine_labels.js, the url passed to the preview in
preview_xul_192 was encoding the url with the javascript encode() function.
This function does not preserve UTF-8 encodings. I changed the function to
encodeURIComponent, and UTF-8 characters are now displayed properly by the
spine label printer code.
Signed-off-by: Liam Whalen <lwhalen@evergreen-dev.catalogue.nrcan.gc.ca> Signed-off-by: Dan Scott <dscott@laurentian.ca>
Bill Erickson [Thu, 29 Sep 2011 15:57:46 +0000 (11:57 -0400)]
Added missing Penalty module import for AssetCommon
The lack of import primarily affected the Action/Trigger MarkItemLost
reactor, which does not already import OpenILS::Utils::Penalty directly
or indirectly.
Signed-off-by: Bill Erickson <berick@esilibrary.com>
Dan Scott [Thu, 29 Sep 2011 03:06:19 +0000 (23:06 -0400)]
Remove an installation step: Listen 443 uncommenting
Every distribution known to man ships Apache listening on port 443, so
installers always have to comment out the "Listen 443" directive - so
why not just comment it out ourselves and save a step in the
installation instructions?
Also, merge Apache configuration instructions from master.
Mike Rylander [Wed, 28 Sep 2011 19:45:36 +0000 (15:45 -0400)]
Provide proper hrefs for links back from report debugging pages
Reimplementation of a fix provided by Michael Peters. The originally submitted
patch included several fixes provided by an unrelated patch, and a regression
in build-time embedded tokens.
Original branch at: http://git.evergreen-ils.org/?p=working/Evergreen.git;a=shortlog;h=refs/heads/user/mrpeters-isl/clark_kent_back_links
Signed-off-by: Mike Rylander <mrylander@gmail.com>
Dan Scott [Thu, 22 Sep 2011 16:51:22 +0000 (12:51 -0400)]
LP856567: Mark HTML report output as Unicode-encoded
Non-ASCII characters in report ouput that has been generated in HTML
format appears to be corrupted when viewed in a browser. This can be
manually corrected by switching the "Encoding" browser option, but it
would be better to include <meta charset='utf-8'> in the output HTML to
tell the browser directly.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Mike Rylander <mrylander@gmail.com>
Dan Wells [Tue, 27 Sep 2011 16:22:38 +0000 (12:22 -0400)]
MFHD compression fails with pattern-less captions
The current code assumes that if you try to compress a range of
MFHD holdings, you actually have the necessary pattern information.
This commit introduces a very basic check to avoid the most common
failure case.
Signed-off-by: Dan Wells <dbw2@calvin.edu> Signed-off-by: Mike Rylander <mrylander@gmail.com>
Dan Scott [Fri, 23 Sep 2011 14:55:28 +0000 (10:55 -0400)]
LP857460: Get the generic class scheme when adding vols
In a default Evergreen instance, in the "Add Volumes" interface for a
bib record with both LC and Dewey call numbers, the list of call numbers
from which you can select appears to be limited to Dewey, even though
the "Bib Call #" field is populated with an LC call number.
The problem is that here we're grabbing the first entry from the
asset.call_number_classification table, and you might expect that to be
the "Generic" entry - however, it turns out to be the "Dewey" entry
because the entries are ordered alphabetically by .name(), not by
database ID. (You can test this be changing the name of "Dewey" to
"Vicious" and see that the "Generic" entry is chosen instead).
This commit changes the process to first search the acnc array for a
name() = "Generic"; and then if we don't find that, fall back to the
first entry in the classification scheme.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Dan Wells <dbw2@calvin.edu>
Some EDI translation fixes for troublesome input characters
This attempts to address enough of the issues in
https://bugs.launchpad.net/evergreen/+bug/812593 to move forward.
When the ? character appears in lineitem attributes for a PO that we
send off for EDI translation, the translator fails if that ? is the last
character of the string. This commit avoids that situation.
This commit also removes any Unicode characters from the JEDI (the input
to the translator process), as they're not going to make it to the
EDIFACT output anyhow, and will just confound the translator.
This commit also pre-escapes quotation marks in lineitem attributes so
that we don't produce invalid JSON for our JEDI.
This commit also destroys the death star.
Seriously, the right way to fix some of the issues people have had is
for somebody to learn Ruby and to address them either in edi_webrick.rb, in
edi4r, or in openils-mapper (wherever the issues may be). Working on
that.
Signed-off-by: Lebbeous Fogle-Weekley <lebbeous@esilibrary.com> Signed-off-by: Mike Rylander <mrylander@gmail.com>
lp859190: make marc_stream_import always import non-matches
Fixes regression of (implicit, undocumented) functionality
introduced in commit da3a45344205. Current use case for
the streaming MARC importer for OCLC Connexion exports assumes
that non-matching records should always be added.
Signed-off-by: Galen Charlton <gmc@esilibrary.com> Signed-off-by: Bill Erickson <berick@esilibrary.com>
Jeff Godin [Wed, 21 Sep 2011 20:41:19 +0000 (16:41 -0400)]
Fix: Current Bills thinks it is Bill History
This fixes LP 855894: Bills interface calls itself Bill History,
uses bills_historical receipt template
https://bugs.launchpad.net/evergreen/+bug/855894
At some point in development, the Current Bills and the Bill History
interface were in a unified file, with an xul_param of "current" used
to switch between the two.
Nothing currently sets the xul_param "current", so the Current Bills
interface (bill2.xul and bill2.js) was improperly:
* not hiding the xact_finish column
* using a caption label of "Bill History"
* using the bills_historical receipt template
The Bill History interface was okay -- "not current" was the default.
This commit removes the tests for xul_param('current') and uses the
appropriate code for each interface.
// this code in both bill2.js and bill_history.js:
var template = 'bills_historical';
if (xul_param('current')) template = 'bills_current';
// becomes:
var template = 'bills_current'; // in bill2.js
var template = 'bills_historical'; // in bill_history.js
Bill Erickson [Thu, 22 Sep 2011 15:10:09 +0000 (11:10 -0400)]
Stamped upgrade script for ACQ fund view repairs
At the heart of this change is the need to force acq.fund_debit_total to
return exactly 1 row per fund, instead of 1 per fund + encumbrance
value. However, such a change required rearranging a number of
dependent views.
Also added acq.fund_spent_balance to the set of views that needs
dropping and re-building.
Minor SQL format change to match surrounding code in schema file
See also https://bugs.launchpad.net/evergreen/+bug/800477
Signed-off-by: Bill Erickson <berick@esilibrary.com>
Dan Scott [Sat, 17 Sep 2011 03:44:28 +0000 (23:44 -0400)]
Can't have two identical IDs on the same page
Nobody who cares about the pure Web experience ever noticed that this
page has two identical IDs. Chromium and other browsers complain,
though, so make them happy.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Mike Rylander <mrylander@gmail.com>
Bill Erickson [Thu, 25 Aug 2011 18:20:56 +0000 (14:20 -0400)]
Clear hold shelf [expire] time at checkout to alternate patron
When an item on the holds shelf is checked out to an alternate patron
(i.e. not the hold user), clear the shelf_time and shelf_expire_time
along with capture_time, current_copy, etc. on the hold.
tsbere++ for isolating the problem.
Signed-off-by: Bill Erickson <berick@esilibrary.com> Signed-off-by: Jason Etheridge <jason@esilibrary.com>
Dan Scott [Fri, 16 Sep 2011 13:59:54 +0000 (09:59 -0400)]
LP 851915: Remove references to /openils/lib/perl5
To prevent old versions of the OpenSRF and Evergreen Perl modules from
conflicting with the new versions of these modules (which are now
located in the default system Perl package directories), remove
references to them.
Also, O:W:AddedContent contains an old reference to a distro-specific
Perl directory that we can clean up at the same time.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Galen Charlton <gmc@esilibrary.com>
lp834961: fix authoritative versions of callnumber retrieval calls
Makes open-ils.search.callnumber[.fleshed].retrieve.authoritative
actually be authoritative. Fixes bug where adding a volume
could result in a false ASSET_CALL_NUMBER_NOT_FOUND error if using
database replication and Slony-II.
Dan Scott [Wed, 7 Sep 2011 19:18:07 +0000 (15:18 -0400)]
Display warning message and prompt before PostgreSQL removal
Building on the good work in LP# 801624, use the -P "prompt" flag
instead of -yq for apt(itude) to ensure that the user has the ability to
quit the dependency install and reevaluate whether they want to actually
destroy their existing PostgreSQL databases. Note that this is
specifically in the context of the removal of the database packages,
rather than a global change.
Also, turn the inline comment into a displayed description of what is
happening and why, along with a nice big warning.
Signed-off-by: Dan Scott <dscott@laurentian.ca> Signed-off-by: Bill Erickson <berick@esilibrary.com>
Missing a =s in getopts config means it becomes a 0 or 1, when it should
be a string.
The problem only occurs when you are specifying a particular postgresql
pg_config file (for example, if you have multiple versions of postgresql
installed).
Signed-off-by: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Dan Scott <dscott@laurentian.ca>