summaryrefslogtreecommitdiff
path: root/TODO
diff options
context:
space:
mode:
authorivan <ivan>1999-01-25 12:10:00 +0000
committerivan <ivan>1999-01-25 12:10:00 +0000
commit6e5eed5d49163fdb06f16c735ab43486c047011d (patch)
treec9c286e4217ae9a3d496502732d6a0ea28e8c1dd /TODO
parent06f4284a7f367b0d70903ab969b305bf4efb6b1b (diff)
yet more mod_perl stuff
Diffstat (limited to 'TODO')
-rw-r--r--TODO48
1 files changed, 36 insertions, 12 deletions
diff --git a/TODO b/TODO
index d53dab078..363c42adb 100644
--- a/TODO
+++ b/TODO
@@ -1,4 +1,4 @@
-$Id: TODO,v 1.19 1999-01-18 23:05:00 ivan Exp $
+$Id: TODO,v 1.20 1999-01-25 12:09:48 ivan Exp $
If you are interested in helping with any of these, please join the mailing
list (send a blank message to ivan-freeside-subscribe@sisd.com) to avoid
@@ -6,18 +6,9 @@ duplication of effort.
-- before 1.2b1 --
-Lay out the forms a bit better.
-
-view/cust_main.cgi should show country
-
-add doc on mod_perl: upgrade to current CGI.pm
-doc Apache::AuthDBI as well
+Can't use [e]idiot with mod_perl.
-FS::cust_pkg _require_'s FS::$svc, but this won't work with %FS::UID::callback
-loading of configuration. (pry need same idea, but will run immediately if
-context allows). Looks like error is masked by 'use FS::cust_svc' which in
-turn 'use's FS::{svc_acct, svc_acct_sm, svc_domain}' which is now explicit
-w/comments in source
+Lay out the forms a bit better.
libwww regression test - since it will create sample data, can also use this
for the web demo
@@ -41,11 +32,24 @@ associated with any agent types so it won't show up for other customers to buy.
(but also... make sure they go away when the customer does! - need this? :
one-off package edits! : need a cust_pkgs or cust_part_pkgs or something table,
with custnum and partpkg (like type_pkgs)
+(what happens if you hit "custom pricing" but the pricing is already custom?)
update site_perl/table_template* (pry out of date)
+update web demo
+
-- before or after ---
+FS::cust_pkg _require_'s FS::$svc, but this won't work with %FS::UID::callback
+loading of configuration. (pry need same idea, but will run immediately if
+context allows). Looks like error is masked by 'use FS::cust_svc' which in
+turn 'use's FS::{svc_acct, svc_acct_sm, svc_domain}' which is now explicit
+w/comments in source
+
+Allow a cancelled/suspended/active status from packages to bubble up to
+the customer lists. Put active, then suspended, then cancelled accounts.
+Similar ordering on the package listing inside a single customer.
+
svc_domain.pm mail sending uses Date::Format which doesn't seem to pick up
correct timezone.
@@ -62,6 +66,19 @@ site_perl/svc_domain.cgi (hmm... or maybe should have a button? or maybe svc_do
-- after --
+bah, table/itable/*table in FS::CGI is silly.
+
+doc Apache::AuthDBI as well
+..
+Provide sample httpd.conf files.
+
+hey look: Tie::DBI! Check that out. Override its commit with something that
+does perl-side caching for ? a performance improvement and as an emulation
+layer to plug in f.ex mysql's atomic transactions
+..
+Record.pm uses does some non-portable DBI things. MySQL and Pg seem fine.
+Fix it anyway unless we migrate to Tie::DBI.
+
faq
cust_bill.pm uses '==' comparison on dates because they're currently ints
@@ -366,6 +383,13 @@ transactions or atomic updates). Or just require a RDBMS that supports
rollback and/or atomic updates and get rid of the work-arounds? The /rdb
interface had this kludge on top of it but is a technical dead-end in most
other ways, unless it can gain an SQL parser and DBD interface.
+...
+if i'm really bored, find the /rdb interface in fsold and port it to NoSQL,
+and while I'm add it add interfaces for AnyDBM_File tied hash.. hmm. Shouldn't
+an FS::Record have something to do with a tied hash? But we don't want
+performance to go gaga... maybe something with commit to help out here?
+...
+Ok: FS::Record gives you a tied hash, and you get methods for commit, etc.
Better automated comparison of our CC records with processors (CyberCash,
at least, has not always had 100% accuracy, though recent versions are