Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
Commits on Feb 5, 2011
  1. @pajamian

    Strip bad characters from affiliate source.

    pajamian authored
    Strip bad characters from affiliate source.  These characters cause problems in
    the usertrack log (and possibly other places) if they are allowed to remain in
    the source.
Commits on Jan 25, 2011
  1. @pajamian

    Allow cutomization of "DO ANOTHER" HTML block in the table editor.

    pajamian authored
    The DO ANOTHER block contains hard-coded text and return-pages used for radio
    buttons which is always displayed for new entries.  This change allows
    customized HTML to replace it for when you want to use the table editor in your
    own custom admin pages.
Commits on Dec 30, 2010
  1. @pajamian

    Don't try to optimize an or search where we don't have all the sql sp…

    pajamian authored
    …ecs.
    
    If we are running an or search and not all of the specs have equivelant sql
    specs (WHERE clauses) then don't try to optimize the query as doing so will
    cause records to not be returned that might match other specs.
Commits on Sep 13, 2010
  1. @pajamian
Commits on Apr 15, 2010
  1. @pajamian @racke

    prevent warning when label is undefined

    pajamian authored racke committed
Commits on Sep 30, 2009
  1. @pajamian
  2. @pajamian

    Get rid of CVS tags.

    pajamian authored
Commits on Sep 15, 2009
  1. @pajamian
  2. @pajamian

    flypage SpecialSub changes.

    pajamian authored
    * Revert behavior in the case where the return value is a hashref.
    
    * A return value of undef explicitly states that we should not process a
      flypage.
    
    On 08/31/2009 01:54 PM, Mike Heins wrote:
    > Quoting Peter (peter@pajamian.dhs.org):
    >> On 08/31/2009 12:01 PM, Mike Heins wrote:
    >>> Quoting Jon Jensen (jon@endpoint.com):
    >>>> It looks more reasonable to me than the old code, but the behavior doesn't
    >>>> quite look compatible. Before, $base could be assigned an arbitrary result
    >>>> hashref and thus could be used for a different kind of on-the-fly product,
    >>>> but now it seems the product code has to actually exist in a real
    >>>> database.
    >>> Good point.
    >>>
    >>>> I am not using that feature and don't object to the change, but I thought
    >>>> I'd point out the different behavior.
    >>> Now that I think about it, it was designed to handle OnFly products so the
    >>> change doesn't make sense.
    >> Ok, well that's why I asked.  How about we keep the old code if the
    >> return is a hashref, but if it's just a text scalar which contains a sku
    >> then get $base from product_code_exists_ref()?
    >
    > I think that works.
    >
    >> Also just to give an explicit way to say, "don't display a flypage for
    >> this" a return of undef could result in the flypage being skipped as
    >> if the sku doesn't exist?
    >
    > Yes.
Commits on Sep 13, 2009
  1. @pajamian
  2. @pajamian

    Merge branch 'master' of git://github.com/interchange/interchange

    pajamian authored
    Conflicts:
    	WHATSNEW-5.7 ... MANUALLY RESOLVED
  3. @pajamian

    Remove bloat that duplicates %z functionality.

    pajamian authored
    On 09/12/2009 09:26 PM, Mike Heins wrote:
    > Quoting Peter (peter@pajamian.dhs.org):
    >> Incidentally why the mess of code in Util.pm to get the same result that
    >> %z would give anyways?
    >
    > Probably because I was stupid when I wrote it, and didn't understand what
    > %z did. Or I did it before %z became standard.
    >
    > It should be changed.
Commits on Aug 28, 2009
  1. @pajamian

    New area_page SpecialSub.

    pajamian authored
    New area_page SpecialSub that allows you to change the page name passed to the
    [area] tag before the tag processes the page name.
  2. @pajamian

    flypage SpecialSub fix for non-existent SKUs.

    pajamian authored
    If the flypge SpecialSub returned a non-existent SKU then a blank flypage would
    be displayed (with no data).  This is now fixed so that there is no difference
    to what happens without a SpecialSub.
Commits on Aug 15, 2009
  1. @pajamian
Commits on Jul 20, 2009
  1. @pajamian

    Do not specify a default charset if none is passed via MV_HTTP_CHARSET.

    pajamian authored
    Do not specify a default charset if none is passed via MV_HTTP_CHARSET.
    Thanks to Raymond Cheng <rayonnet@hotmail.com> for pointing out the regression
    caused by this.
Commits on Jul 1, 2009
  1. @pajamian
Commits on Jun 24, 2009
  1. @pajamian

    Fix pod

    pajamian authored
    Fix pod in BusinessOnlinePayment.pm.  Thanks to Rene Hertell for pointing out
    the problem and Gert van der Spoel for providing the fix.
Commits on May 20, 2009
  1. @pajamian

    MaxQuantity fixes:

    pajamian authored
    * A null or non-numerical entry in the table(s) for the maxquantity field should
      result in no maximum quantity enforcement for that product.
    
    * Remove redundant code for fetching the quantity from the table(s).
    
    * Fix instance where two DB hits are used to fetch one value.
    
    * Get rid of goto.
    
    * Basically a rewrite of the correspondign block of code.
Commits on May 1, 2009
  1. @pajamian

    * New adjust_time() function allows time adjustment based on secs, mi…

    pajamian authored
    …ns, hours,
    
      days, weeks, months and years.  Also can string multiple adjustments together
      and compensate for daylight savings time changes over the adjustment period.
    
      This sub (in Util.pm) accepts up to three arguments.  The first (required)
      argument is the adjustment to be made and is a signed number followed by one
      of seconds, minutes, hours, days, weeks, months or years (these can be
      abbreviated).  The second optional argument is a unixtime to be adjusted and
      defaults to the current time.  If the third argument is true then the
      resulting time will be compensated for any daylight savings time differences.
      The sub uses POSIX::mktime() to do the conversions and so does not come with a
      new module requirement for Interchange.
    
    * _set_acl() in UserDB.pm now uses adjust_time() instead of time_to_seconds().
    
    * [pay-cert] tag now uses the new adjust_time() function instead of the older
      time_to_seconds().
    
    * [component], [convert-date] and [css] now use the new adjust_time() function
      instead of the older time_to_seconds().
    
    * [convert-date] now has new compensate_dst argument that when set to 1 will
       compensate the adjusted time for daylight savings time changes (it basically
       passes 1 as the third argument to adjust_time()).
    
    * fixed bug in [convert-date] that skewed the time by one hour if the starting
      date was during daylight savings time and an adjustment was made.
    
    * [time] now uses adjust_time() instead of the older time_to_seconds().  There
      is also a new compensate_dst attribute that when set to 1 will compensate the
      adjusted time for daylight savings time changes (it basically passes 1 as the
      third argument to adjust_time()).
    
    Following are examples that help to illustrate all of the above new features.
    Note that many of the examples may require adjustment depending on your local
    time zone and locale settings:
    
    Code:
    <pre>
    Current time: [time]%c[/time]
    +5 hours: [time adjust="+5 hours"]%c[/time]
    +5 years: [time adjust="+5 years"]%c[/time]
    -8.3 months: [time adjust="-8.3 months"]%c[/time]
    +2.6 days: [time adjust="+2.6 days"]%c[/time]
    +7.7 months: [time adjust="+7.7 months"]%c[/time]
    +7 months: [time adjust="+7 months"]%c[/time]
    +7 months (compensate_dst): [time adjust="+7 months" compensate_dst=1]%c[/time]
    +5: [time adjust="+5"]%c[/time]
    -500 hours=1: [time adjust="-500" hours=1]%c[/time]
    hours=+500: [time hours=+500]%c[/time]
    +500: [time adjust="+500"]%c[/time]
    +5 years, 6 months, 3 days, -4 hours, 7 minutes (compensate_dst): [time adjust="+5 years, 6 months, 3 days, -4 hours, 7 minutes" compensate_dst=1]%c[/time]
    
    duration filter:
    [cgi name=start_date set=200502120800 hide=1]
    [cgi name=offset set="12 hours 10 mins" hide=1]
    20050212201000: [filter duration.start_date.offset][/filter]
    20050212200500: [filter duration.-dummy.12.hours.5.mins]200502120800[/filter]
    
    convert-date tag:
    200807151600: [convert-date adjust="+6 months" format="%c"]200807151600[/convert-date]
    200807151600 (compensate_dst): [convert-date adjust="+6 months" format="%c" compensate_dst=1]200807151600[/convert-date]
    </pre>
    
    
    Results:
    Current time: Fri May  1 05:40:47 2009
    +5 hours: Fri May  1 10:40:47 2009
    +5 years: Thu May  1 05:40:47 2014
    -8.3 months: Fri Aug 22 22:28:47 2008
    +2.6 days: Sun May  3 20:04:47 2009
    +7.7 months: Tue Dec 22 21:28:46 2009
    +7 months: Tue Dec  1 04:40:47 2009
    +7 months (compensate_dst): Tue Dec  1 05:40:47 2009
    +5: Fri May  1 10:40:47 2009
    -500 hours=1: Fri Apr 10 09:40:47 2009
    hours=+500: Fri May 22 01:40:47 2009
    +500: Fri May  1 10:40:47 2009
    +5 years, 6 months, 3 days, -4 hours, 7 minutes (compensate_dst): Tue Nov  4 01:33:47 2014
    
    duration filter:
    
    
    20050212201000: 20050212201000
    20050212200500: 20050212200500
    
    convert-date tag:
    200807151600 +6 months: Thu 15 Jan 2009 03:00:00 PM PST
    (compensate_dst): Thu 15 Jan 2009 04:00:00 PM PST
Commits on Mar 16, 2009
  1. @pajamian

    Added SourcePriority catalog directive:

    pajamian authored
    SourcePriority <source_list>
    
    <source_list> is a prioritized list of cgi variables to get the source
    (affiliate) name from.  Can also include the following:
    
    mv_pc - has the current special casing of mv_pc, (ie RESET is special as
    are values that contain only digits).
    
    cookie-foo check the cookie with the foo label.
    
    session - stop here if session already exists, do not check any further
    variables.
    
    session-foo - stop here if foo session variable is set.
    
    Default: SourcePriority mv_pc mv_source
    
    Examples:
    
    Check the MV_SOURCE cookie for an affiliate name as well as the other defaults:
    SourcePriority mv_pc mv_source cookie-MV_SOURCE
    
    ...as above, but you don't want your affiliates using mv_pc:
    SourcePriority mv_source cookie-MV_SOURCE
    
    Check the cgi variable affid instead:
    SourcePriority affid
    
    Say you send affiliate traffic to other sites, and you don't want
    those sites to get credit for sales if a customer follows a banner from
    them back to your site:
    SourcePriority session mv_pc mv_source
    
    If you want affiliates who use the specialsource cgi variable instead of
    mv_source to get special treatment and can override customers who
    already have sessions:
    SourcePriority specialsource session mv_pc mv_source
    
    If you want to allow affiliates to get credit if there is a
    session but only if no other affiliate is already set:
    SourcePriority session-source mv_pc mv_source
Commits on Oct 24, 2008
  1. @pajamian

    Fix vulnerability where a string passed in the mv_order_item CGI vari…

    pajamian authored
    …able is
    
    displayed verbatim without any input sanitation if there is a valid sku in
    mv_sku.  Thanks to Mat from Bibliopolis for discovering and reporting the
    vulnerability.
Commits on Apr 23, 2008
  1. @pajamian

    Fixed bug where previous rows values were left in $Row if [loop] is i…

    pajamian authored
    …terated with [PREFIX-next].
Commits on Dec 19, 2007
  1. @pajamian

    New universal "hide" attribute. You can now set the hide=1 attrbute f…

    pajamian authored
    …or any tag
    
    to supress its output.
    
    In order to preserve backwards compatibility with existing individual tags' hide
    attributes tag output is not reparsed if this is set regardless of the setting
    of interpolate or reparse attributes.  The reasoning behind this is existing
    hide=1 attributes clear the output of the tag before the output is reparsed so
    setting hide=1 on those tags which used to support it had the effect of not
    reparsing output (because there was no output left to reparse).  This can be
    tested with the following line of ITL:
    
    [tmp bar]bar[/tmp][calcn hide=1]'foo[tmp bar][area foo][/tmp]'[/calcn]::[scratchd bar]
    
    In my testing the above line of html outputs:
    ::bar
    
    Without the hide=1 attribute to calcn it outputs something like (url changed to
    protect the guilty):
    foo::http://www.example.com/foo.html
    
    So hide=1 is effectively preventing reparsing as well as hiding the output.
Commits on Nov 14, 2007
  1. @pajamian

    * New Multiple CodeDef flag to indicate that a widget can return mult…

    pajamian authored
    …iple
    
      selections.  Usage example:
           CodeDef checkbox Multiple 1
    
    * Set new Multiple flag for checkbox, movecombo, check_nbsp and multiple
      widgets.
    
    * Option editor will now correctly generate variants for options that use
      widgets capable of multiple selections (ie, multiple or checkout widgets).
    
    * New [widget-info] UI tag:
    
    	[widget-info name attribute]
    
    	name: Name of widget to fetch info on, eg: checkbox, if not specified
    	then a hashref of hashrefs will be returned with all attributes for all
    	defined widgets.
    
    	attribute: Info to get for widget, eg: Multiple.  If not specified then
    	all attributes will be returned for the named widget in a hashref.
Commits on Nov 10, 2007
  1. @pajamian

    Adding some simple options code to Matrix.pm so that it can support m…

    pajamian authored
    …ixed Matrix/Simple option products.
Commits on Nov 9, 2007
  1. @pajamian

    * Allow the default display_type for Matrix widgets to be set from ca…

    pajamian authored
    …talog.cfg
    
      as follows:
         Options Matrix display_type separate
    
    * Clean up and simplify code a bit.
  2. @pajamian

    Allow for the proper display of product options that use text or text…

    pajamian authored
    …area
    
    widgets by showing the actual value if a label doesn't exist for the value.
Commits on Aug 13, 2007
  1. @pajamian

    Fixed problem where form values do not always default to the $Values …

    pajamian authored
    …hash
    
    values for corresponding form fields when defaults=1 and wizard=1 are both set
    in the table editor.
Commits on Aug 12, 2007
  1. @pajamian

    New Vend::Server::set_process_name sub which is used to change the st…

    pajamian authored
    …atus of
    
    the process name indicator.  This respects the MV_DOLLAR_ZERO settings.
Commits on Aug 10, 2007
  1. @pajamian

    Applied patch from Mark Johnson <mark@endpoint.com> to fix problem wi…

    pajamian authored
    …th RPC
    
    mode preforking too many children on server startup due to race condition.
Commits on Aug 9, 2007
  1. @pajamian

    * Updated copyright headers to 2007 (closes bug #102)

    pajamian authored
    * Added GPL and copyright headers to a few files that were missing them.
Commits on Aug 5, 2007
  1. @pajamian

    More shipping code cleanup

    pajamian authored
Commits on Aug 4, 2007
  1. @pajamian

    The new Ship.pm seems to have a lot of problems. Here is a patch which

    pajamian authored
    addresses the following issues:
    
    * Fixed problem where only the last shipping policy will get stored if
    the multi-line format is used in shipping.asc.
    
    * Fixed problem where options are not converted and stored properly on
    all shipping policies.
    
    * Moved more code into the new process_new_beginning sub:
    	* By passing a ref to the @shipping array I can now push new records
    onto the array from inside the sub.
    	* The above means that the old @list array need not be returned since
    it's only purpose at that point was to push the record onto @shipping,
    then immediately be overwritten by @new.
    	* That also means that @new can be assigned to @list inside the sub
    instead of outside it.
    	* Since we no longer need to return @new, this frees up the return
    value which I have used to return the $first value instead.
    	* Because $first is returned we no longer need to flag that condition
    by assigning "_newmode" as the shipping mode.
    
    * I also removed an entire redundant elsif branch from read_shipping.
    The regexp that the branch tested for would never cause the branch to be
    executed because the branch above it tests for a regexp that matches the
    same lines.
    
    I don't think it's a big deal to change the process_new_beginning sub
    the way I did because the sub itself is only about 5 weeks old, so the
    likelyhood of someone having implemented code around it is extremely small.
Commits on Jul 5, 2007
  1. @pajamian
Something went wrong with that request. Please try again.