dbs [Tue, 22 Mar 2011 01:24:43 +0000 (01:24 +0000)]
Build libjs from source for now until the Perl threading problem can be reliably tested
The previous commit changed JavaScript::SpiderMonkey to be built without
thread safety, but on Fedora we had been installing the js-devel package
(which was threaded). As you can't build a non-threaded
JavaScript::SpiderMonkey on top of a threaded libjs, for now we'll make
Fedora build a non-threaded libjs from source just like the other distros.
Note that the problem with threaded JavaScript::SpiderMonkey surfaced
on Debian Lenny when targeting holds. Lenny has libthreads-perl 1.71, but
a number of fixes have been committed to subsequent versions
(http://cpansearch.perl.org/src/JDHEDDEN/threads-1.82/Changes) - so
threading support might be reliable on more current distros. Once we
have a reliable test, we might be able to go back to the packaged
version of libjs.
dbs [Tue, 22 Mar 2011 00:40:44 +0000 (00:40 +0000)]
Another layer of escaping for the escapes in the DB update script
Apologies for the previous patch that led to breakage; I "tested" in a
script in which I had already hardcoded the answer I wanted after
running into problems with the previous version of the script. Not
good testing practice.
Thanks to Bill Erickson for suggesting the extra escapes. Tested in
a clean environment this time, and it works.
dbs [Mon, 21 Mar 2011 20:12:17 +0000 (20:12 +0000)]
Enable update_db.sh to handle upgrading databases with release version
When '0475' is mixed with '1.6.0.4', the latter wins as far as a plain MAX()
operator goes. Ensuring that the version is all digits should help avoid the
release version numbers in the config.upgrade_log table.
Thanks to Thomas Berezansky for suggesting a SIMILAR TO operator variation
that led to this approach.
dbs [Mon, 21 Mar 2011 14:11:21 +0000 (14:11 +0000)]
Use the human-friendly interval syntax for Apache caching
Apache mod_expires has supported a human-readable syntax for defining cache
expiry directives for a long time. In this commit we simply convert the
existing defaults from the "Anumber_of_seconds" syntax to the equivalent
English syntax documented at
http://httpd.apache.org/docs/2.0/mod/mod_expires.html
Whether CSS _should_ have a cache lifetime of 50 minutes by
default is left to the subject of future commits :)
miker [Sat, 19 Mar 2011 14:33:18 +0000 (14:33 +0000)]
In-database Grace intervals
s it stands the grace period for a transaction is only specified on the fine generator script, but fines can be generated on checkin as well without one. As a result any checkin results in grace periods being negated.
The attached patch adds in-db grace periods for all non-booking transactions, removing the grace period from the fine generator cron job script and placing it on the circulation. Grace periods are set on the recurring fine rule with an override available on the circ matchpoint, identical to how renewal count can be overridden from the duration rule already. These grace periods are respected on checkin fine generation as well as on the fine generator cron job script.
The upgrade script will assume, by default, a grace period of 1, but has a \set line to change this. It will also update all circulations without a checkin by default, but another \set will allow disabling of that.
[ NOTE: for the history of the patch, see: https://bugs.launchpad.net/evergreen/+bug/732679 ]
Signed-off-by: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Mike Rylander <mrylander@gmail.com>
git-svn-id: svn://svn.open-ils.org/ILS/trunk@19823 dcc99617-32d9-48b4-a31d-7c20da2025e4
dbs [Sat, 19 Mar 2011 04:56:06 +0000 (04:56 +0000)]
Remove unused imports, correct some references, and clean up csedit.py
Mostly minor cleanups, some additional docstrings, but a few changes
of concern here (OSRF_APP_CONNECTED_STATUS didn't have the appropriate
osrf.const import, for example).
dbs [Sat, 19 Mar 2011 03:06:58 +0000 (03:06 +0000)]
Parse the field_safe parameter in Python's IDL parser
Also set the correct namespace for the tablename attribute.
The rest of the changes are for PEP8 compliance. We retain the
IDLParser set_IDL() and parse_IDL() methods because things like
constrictor call them directly, but note in the docstrings that
they are deprecated names.
phasefx [Fri, 18 Mar 2011 18:08:51 +0000 (18:08 +0000)]
Switch to AccessControl plugin, more reliable.
Only fetches it when needed.
Also make NSIS Less...verbose. Warnings/errors only.
Author: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Jason Etheridge <jason@esilibrary.com>
git-svn-id: svn://svn.open-ils.org/ILS/trunk@19810 dcc99617-32d9-48b4-a31d-7c20da2025e4
dbs [Fri, 18 Mar 2011 17:50:19 +0000 (17:50 +0000)]
Test for an installed libdbi0 package and bail on Ubuntu and Debian
This isn't a problem on Fedora, where we use the packaged version of libdbi,
but the older version of libdbi0 and possibly the compile options on Ubuntu
Lucid has been demonstrated to be a barrier to successfully installing
Evergreen.
berick [Fri, 18 Mar 2011 16:08:06 +0000 (12:08 -0400)]
make all field_safe classes accessible via fetch/cache by default. reduce page churn by creating then caching the fetch/cache and org-tree generator funcs and simply inserting them into the page context on each page load. IOW, no need to redefine the funcs on each page load
miker [Fri, 18 Mar 2011 15:53:04 +0000 (15:53 +0000)]
Intelligent Metarecord Format Selection for Holds
Previous to this commit, Evergreen would show all possible record formats
when promoting a T-type hold to an M-type. This change causes the middle
layer method to respect the Hard Boundary of the pickup library when
considering what constituent records within a metarecord will contribute
formats.
Signed-off-by: Mike Rylander <mrylander@gmail.com>
git-svn-id: svn://svn.open-ils.org/ILS/trunk@19805 dcc99617-32d9-48b4-a31d-7c20da2025e4
berick [Thu, 17 Mar 2011 20:32:06 +0000 (16:32 -0400)]
as was recently done w/ existing kcls skin; hide the open transaction summary info in my-account, citing confusion and unnecessary additional information. note, you can still see total balance owed
phasefx [Thu, 17 Mar 2011 20:16:56 +0000 (20:16 +0000)]
Pre-modified xulrunner-stub files + icons
Icons courtesy of Mike Peters, or so my note alongside them says.
The pre-modified stubs are built on the 1.9.2.15 build of xulrunner. They should continue to work on future 1.9.2 builds.
Author: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Jason Etheridge <jason@esilibrary.com>
git-svn-id: svn://svn.open-ils.org/ILS/trunk@19795 dcc99617-32d9-48b4-a31d-7c20da2025e4
phasefx [Thu, 17 Mar 2011 20:14:41 +0000 (20:14 +0000)]
Copy is not valid. How I missed that before I don't know.
Author: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Jason Etheridge <jason@esilibrary.com>
git-svn-id: svn://svn.open-ils.org/ILS/trunk@19794 dcc99617-32d9-48b4-a31d-7c20da2025e4
phasefx [Thu, 17 Mar 2011 20:14:04 +0000 (20:14 +0000)]
Default to no license page (for the staff client installer)
Author: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Jason Etheridge <jason@esilibrary.com>
git-svn-id: svn://svn.open-ils.org/ILS/trunk@19793 dcc99617-32d9-48b4-a31d-7c20da2025e4
senator [Thu, 17 Mar 2011 16:23:13 +0000 (12:23 -0400)]
make presentation better by hiding as-yet-unready features, etc
also, made links out of the dashboard counts, but oddly the first one
("checked out") only works in IE, not Chrome or Firefox. I suspect that
the complications of all the "position: absolute"-this and "position:
relative"-that make for fluky layout.
senator [Wed, 16 Mar 2011 18:48:52 +0000 (14:48 -0400)]
"Add to my list" becomes "remove from my list", and
it now works on the record detail page as well as the results page.
Also, I fixed a bug by which an HTML form was rendered wrongly, totally
breaking the display of the results page in IE and preventing the first
result on any page of search results from adding to the anonymous list.
dbs [Wed, 16 Mar 2011 03:46:43 +0000 (03:46 +0000)]
Set default db schema search path to public,pg_catalog
With the addition of the 'evergreen' schema, we discovered that
all of the unqualified database objects were being created in
the 'evergreen' schema if the database user name was 'evergreen'
(as the default search path is "$user",public we started matching
"$user").
Thomas Berezansky proposed a variation of the function provided
here for changing database settings. By creating the function in
the first schema file, the subsequent connections pick up the
new default search path and all works blissfully as it once did.
Long term, we should probably move towards explicitly qualified
database objects.
dbs [Wed, 16 Mar 2011 03:38:26 +0000 (03:38 +0000)]
Make DROP statement match new basetype for agg_tsvector
When the CREATE AGGREGATE function was changed to have a basetype
of pg_catalog.agg_tsvector, the corresponding DROP statement was
not changed in a similar fashion; ergo, on creating the schema
a subsequent time, the CREATE statement would fail because the
aggregate function already existed and the transaction would be
rolled back.
Also - do we need to differentiate between PostgreSQL 8.4 and 9.0
here?
phasefx [Tue, 15 Mar 2011 21:05:49 +0000 (21:05 +0000)]
With some hints from eeevil, tweaked 002.functions.aggregate.sql to work after some search_path mangling. Haven't smoketested yet, just got past the eg_db_config --create-schema bit with this.
senator [Tue, 15 Mar 2011 19:16:23 +0000 (15:16 -0400)]
bib record subject search terms: painstaking translation of BibTemplate to TT
the reference implementation I used is in
Open-ILS/web/opac/skin/default/xml/rdetail/rdetail_summary.xml starting
at line 93 as of this commit.
Only now that I've imitated the default skin's behavior exactly, I note
that the subject search links produced don't actually work on my test server.
Seems to be a local configuration issue though; will investigate.
senator [Tue, 15 Mar 2011 17:47:10 +0000 (13:47 -0400)]
Improve search term propagation from page to page
Also, for things like the author link on the record detail page, build a
link with 'qtype=author&query=JK Rowling' instead of 'query=au:JK Rowling'
The latter works, but causes the search form to look stupid.
Query type dropdown will be set to keyword by default, and the term will
be "au:JK Rowling", which will be a head scratcher for 99% of real
users.
phasefx [Tue, 15 Mar 2011 15:52:12 +0000 (15:52 +0000)]
Misc. changes to installer including:
1 - Fix profile manager links
2 - Add tag at top (currently "Trunk") for easy changing
3 - Add optional extras.nsi for easy additions of extra install/uninstall lines
4 - Run uninstaller if installing to a directory with one (old style OR new style)
5 - Allow license override and only show license page if one defined
Ex. With NSIS_EXTRAOPTS set to "-DPRODUCT_LICENSE=some/path"
Or to disable use "-DPRODUCT_LICENSE"
6 - Similar to devbuild, allow for registration per-machine by default
This includes needing to not wipe out the folder this is stored in
7 - Make "Everyone" able to write to the install folder when:
Autoupdate is an option
Per-machine registration is an option
8 - Make silent uninstall function
Author: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Jason Etheridge <jason@esilibrary.com>
git-svn-id: svn://svn.open-ils.org/ILS/trunk@19750 dcc99617-32d9-48b4-a31d-7c20da2025e4
phasefx [Tue, 15 Mar 2011 11:59:04 +0000 (11:59 +0000)]
Default workstation name to something sensible.
Like the computer's hostname. But not localhost.
Author: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Thomas Berezansky <tsbere@mvlc.org> Signed-off-by: Jason Etheridge <jason@esilibrary.com>
git-svn-id: svn://svn.open-ils.org/ILS/trunk@19747 dcc99617-32d9-48b4-a31d-7c20da2025e4
senator [Mon, 14 Mar 2011 21:48:25 +0000 (17:48 -0400)]
If the TT stuff loads CGI w/o -utf8, that "instance" (not really instance)...
... will compete with use CGI in EGCatloader.pm. Really. Normally I enjoy
perl but the awful, horrendous, unspeakable black magic used in CGI.pm
has given me a great deal of trouble today.
Specifically, I could not really succeed in subclassing it. I had a kind
of half success before, but I could only really replace methods in my
subclass, but not call the methods unchanged from the original base
class (CGI). This manifested in a bug where I had been using
the query_string() method preserve search terms across some pages, but
that method had stopped working since I introduced the CGI subclass.
In an attempt to fix that, I discovered this document
http://stein.cshl.org/WWW/CGI/#subclassing
and tried the advice therein, and several variations thereof, but
everything I produced caused mod_perl to consume memory until it keeled
over from OOM.
Finally, I discovered that I never really needed the subclass in the
first place, so long as 'use CGI' with no module parameters doesn't
happen in the same process as 'use CGI qw/-utf8/'.
erickson [Mon, 14 Mar 2011 16:34:25 +0000 (16:34 +0000)]
More menu work for record attr/coded map config UIs
menu entry for config.coded_value_maps,
updated menu label and page title for config.record_attr_definition page;
in some cases (that probabl need investigating), widget.validate function is not defined. in those cases, don't attempt to call the function
dbs [Mon, 14 Mar 2011 04:21:20 +0000 (04:21 +0000)]
Apply jshint suggestions for cleanup to dojo.fieldmapper
Mostly consisting of missing semicolons and avoiding the redeclaration
of variables (generally for iterators) within the same function scope.
Some tightening of null and '' checks to avoid implicit type-casting.
dbs [Sun, 13 Mar 2011 22:09:06 +0000 (22:09 +0000)]
Normalize patron profile search function in "Search for Patrons" UI
Use util.widgets.make_menulist for a more consistent approach to
generating the menulist and use a row/label/hbox layout to normalize the
appearance of the patron profile filter in the search menu.
dbs [Sun, 13 Mar 2011 22:02:19 +0000 (22:02 +0000)]
Apply patch to add a patron profile filter to the "Search for Patrons" UI
Thanks to Michael Peters <mrpeters@library.in.gov> and
Thomas Berezansky <tsbere@mvlc.org> for their work on this additional
functionality to the patron search UI, supplied as a patch in LP# 715837.
dbs [Sat, 12 Mar 2011 23:37:45 +0000 (23:37 +0000)]
Clean up build errors when creating/recreating database schema
Dropping the evergreen schema cascaded and wiped out some previously
defined functions, so move it ahead of the other functions. (And maybe
we can start moving those public functions into the evergreen schema?)
DROP AGGREGATE IF EXISTS exists on any version of PostgreSQL that we
care about for Evergreen 2.0, so we can use it to avoid false errors
during schema creation.
FTS config was complaining about public.english_nostop in the ALTER
statement, so remove the schema qualification. This seems to have
been there for a long time?
Use IF EXISTS when dropping the unapi schema to avoid an error.
dbs [Sat, 12 Mar 2011 02:57:32 +0000 (02:57 +0000)]
Resolve inconsistent results when invoking LOWER() in C vs. UTF8 locale databases
As reported by the Fundamental Science Library of Armenia, patron searches
were not retrieving patrons who were known to be part of the patron database.
While the initial approach to solve this problem used an encode_utf8() call
to encode the data before passing it on to the database, this turned out to
break the patron search function in other environments.
Testing by Dan Wells confirmed that the LOWER() function returned different
results when invoked against text in a database created with LC_CTYPE=C vs
a database created with LC_CTYPE=*.UTF-8. As the patron search function
used a Perl lc() function call to convert the incoming data to lowercase,
the success of the call depended on the LC_CTYPE value of the database.
To avoid this problem in the future, we define our own evergreen.lowercase()
function that can reliably produce lowercase text for characters outside
the Latin1 range, and we convert our indexes and function calls to use that
function consistently.
dbs [Fri, 11 Mar 2011 20:37:50 +0000 (20:37 +0000)]
LANG=C results in SQL_ASCII encoding by default for PostgreSQL, which isn't cool
We want LC_CTYPE=C and LC_COLLATE=C, and SERVER_ENCODING=UTF8.
The docs have been adjusted as 8.4 and up now offer database-level
--lc-ctype / --lc-collate options for createdb command.
dbs [Fri, 11 Mar 2011 16:42:01 +0000 (16:42 +0000)]
Ensure incoming query values to patron search are UTF8-encoded
Addresses the problem reported with searching for patrons with
Armenian names, where Unicode query params were getting corrupted
on its way to the database.
senator [Thu, 10 Mar 2011 22:24:08 +0000 (17:24 -0500)]
Foundational work for temporary/anon lists and per-user lists (bookbags)
There were already some features in EGCatLoader for adding and
deleting items from lists, but there's more to do, as that code only
dealt with numeric IDs, and we need records avaiable. Also, to avoid
search bots creating temporary/anon lists all the time, and just because
the following is good practice, our "add to my list" links need to be
forms that POST, not links that GET.
projects / working / Evergreen.git / log