erickson [Wed, 29 Aug 2007 16:02:40 +0000 (16:02 +0000)]
Gave the added content module the ability to do book jacket lookups in addition to its current duties
in addition, book jackets are cached in memcache, which means Apache disk/mem caching won't be necessary
also added in an error checking layer where if X consecutive errors occur, all added content lookups,
except for locally cached images, are disabled until the retry timeout has passed. These new
settings go into the added_content section of the opensrf.xml config file.
new settings:
jacket_url
retry_timeout
All that's left is to add the automatic disable capabilities to the other lookups
phasefx [Wed, 29 Aug 2007 16:02:27 +0000 (16:02 +0000)]
tentative fix for the bug in the prevent-redundant-barcode-scan feature during checkout, where a failed attempt at a checkout will not allow a re-attempt
phasefx [Mon, 27 Aug 2007 03:47:48 +0000 (03:47 +0000)]
Fixed a logic error with renew. The patron Items Out interface keeps some state on each row in the list via the id for the circ. But a renewed circ is a new circ, with a new id, and we were still updating a map with the old circ id.
phasefx [Sun, 26 Aug 2007 05:33:11 +0000 (05:33 +0000)]
Bug fix for list class, where refresh-row functionality broke the on_all_fleshed callback handler. Since rows in a list can flesh themselves out asynchonously, and some actions want to act on all rows in a list, we have a function that asks the whole list to flesh itself (because unseen rows normally don't), and a way to pass a callback function for when the list has been completely fleshed. Internally, it keeps a count of total rows and total fleshed rows, and if the two are equal, then the list is considered fleshed. However, the recent refresh-a-single-row functionality was effectively increasing the total fleshed rows tally without first adjusting it, and you would get fleshed rows > total rows, and the callback would never fire. So this fixes all of that.
phasefx [Sun, 26 Aug 2007 02:30:58 +0000 (02:30 +0000)]
fix for row-refresh bug that caused subsequent on_selects to not get data for the affected row. Basically, if you performed an action on an item in the patron Items Out list, you couldn't perform a subsequent action without refreshing the whole list
phasefx [Thu, 23 Aug 2007 17:06:55 +0000 (17:06 +0000)]
This makes the 'twisties' in the holdings maintenance interface default to a closed position (ie. the plus sign), and they'll sync up correctly with rows whether they're expanded or not, but, new bug??, if you toggle these rows open and close rapidly enough with mouseclicks, you can get a 'missed-click' where the toggle does not happen. I tried tracing the behavior to pinpoint where exactly things go awry, but I'm giving up for now.. seems to be a minor bug, and the other display/render bug is fixed
miker [Sat, 18 Aug 2007 09:01:23 +0000 (09:01 +0000)]
* Big messagecatalog-ization for offline mode (Dan Scott)
* Inline comment support for messagecatalog properties fields (miker)
* Attempt at replacing escaped characters in getFormattedString() (miker)
erickson [Wed, 15 Aug 2007 14:39:25 +0000 (14:39 +0000)]
added support for patron search-by-location-and-depth to patron search api
added support for user_org_unit_opt_in handling
- checking to see if it's globally enabled
- checking to see if a patron has opted-in at the current workstation org
- creating an opt-in record for a patron
phasefx [Wed, 15 Aug 2007 13:59:36 +0000 (13:59 +0000)]
Handle a VOLUME_LABEL_EXISTS event. We might want to consider offering a magic item transfer or volume merge when this happens. Dan, I apologize for not using messageCatalog here yet; I'm still tinkering with that.
phasefx [Tue, 14 Aug 2007 14:57:10 +0000 (14:57 +0000)]
Let's inline a sprintf implementation to get around remote xul restrictions.. also trying out a different sprintf implementation, one with a better license
miker [Tue, 7 Aug 2007 15:36:56 +0000 (15:36 +0000)]
Patch from Dan Scott which starts the process of proper staff client internationalization:
As documented at
http://developer.mozilla.org/en/docs/XUL_Tutorial:Property_Files ,
string bundles are the normal method for holding translatable text
from Javascript files in XULRunner applications. The staff client
currently does not use string bundles; it currently uses a mix of
hardcoded strings, with reliance on some of the strings that are
converted from lang.dtd into the massive hash that is lang.js.
This patch hopes to start changing that, by introducing string bundles
for a few small parts of the staff client -- hopefully paving the way
for conversion of more of the client to stringbundles and i18n glory.
1) We create one .properties file per chrome/content/ subdirectory
that we're converting to string bundles, plus a common.properties file
for common strings like error messages. In this patch, I address the
chrome/content/admin/ and chrome/content/cat/ directories; therefore,
I add:
* Open-ILS/xul/staff_client/chrome/locale/en-US/admin.properties
* Open-ILS/xul/staff_client/chrome/locale/en-US/cat.properties
* Open-ILS/xul/staff_client/chrome/locale/en-US/common.properties
2) Note that the files are being added to a directory that does not
currently exist in the Subversion repository. It doesn't make sense to
serve the properties files remotely from the web/locale/en-US/
directory, as their real home is the chrome. So that's where I've
stuck them. Accordingly, the xul/staff_client/Makefile has been
modified to no longer make this directory.
3) chrome/content/cat/opac.xul has been converted to a fully i18n-ized
file, by adding XML entities to web/opac/locale/en-US/lang.dtd and
converting hardcoded strings in the Javascript to stringbundle calls
that pull in the properties defined in cat.properties and
common.properties.
4) chrome/content/admin/survey* have also been given the entity &
stringbundle treatment.
I have tested the opac and survey screens in the client and nothing
seems to be going wrong, so I'm taking that as a good sign :)
erickson [Mon, 6 Aug 2007 22:51:22 +0000 (22:51 +0000)]
when someone wants to reuse a workstation, we no longer delete the
exsiting workstation with the same name. If necessary (and with appropriate
UPDATE_WORKSTATION perms), we change the owning_lib of the workstation. If
not necessary, we just return the ID of the existing workstation.
miker [Wed, 1 Aug 2007 13:56:56 +0000 (13:56 +0000)]
Patch from Dan Scott to entity-ize a chunk of the Staff Client interfaces:
Taking this one step at a time, I'll move the remaining hard-coded
strings from main/*.xul into lang.dtd.
I've also followed the same ${locale} approach for the remote XUL
files rather than having the en-US hard-coded into the XUL files
themselves -- although at the moment we don't really have a good way
of serving up a different locale anyways. Accordingly, I've included a
modification to eg_vhost.conf.
lang.dtd is getting very big. To try to make it a bit easier to work
with, I've sorted the bulk of the UI entity definitions as seemed to
be the general intent (a few had fallen out of order) so the diffs in
this patch are a little more than absolutely necessary; sorry about
that.