NO JS:
* no comment mod button (for admin, anyway)
* plus/minus
* tag arrow
no JS
degrade
logged in, w/JS
not logged in, pop up modal
no JS, link to FAQ
FAQ entry for no JS
link to it from certain items
triangle
nodnix
custom menu edit button
consistent about FH daybreak links, whether we reload page or not, degrade cleanly
do_index actually GOOD! ;-)
search 30 days stories thing
jamie
search.pl preexisting queries, search logs
ie7 and ie8
-noscript
-index2 failure
slashd scaling
make load faster
remove old JS (perl?) libraries
make perl stuff load sooner, stay cached?
improve perl/JS code performance
HTTP/YSlow
benchmarking
non-logged-in behavior?
keybindings index2
-if collapsed, open on +/-/Z
√"current" on click (how detect?)
on paused?
√if text field has focus!
√make work index2, console
keys
√nod/nix '+' '-'
√ jump to text field for typing?
√hotkey to tag input field
√open/close
√next 'd' 's' 'l' 'j' (unread? 'f')
√previous 'a' 'w' 'h' 'k'
skip daybreaks!!!
-top/bottom 't' 'v'
√signoff
√get more 'g'
-color down/up '[' ']' ',' '.'
update master sandbox for sphinx
revisit auto-updating of SQL
live updating of DB and /usr/local
d2 abbreviated progressive updating:
* comment is sent in small version to save bandwidth:
-√ no commentSub (reply link/parent link/moderation popup/delet checkbox)
-√ no "otherdetails" (user address, comment date, last journal date, IPID, etc.)
-- only first $n "printable" bytes of comment
! here we have a quandry. we can either pass the first $n printable
bytes, or strip out non-printable bytes (esp. <quote>), up to
the first non-<quote> linebreak. the former increases initial
bytes sent, the latter increases bytes fetched (decreases offset,
perhaps to 0)
* truncate like we do with large comments
* memcache the truncated version, and the offset at which original
and truncated version are not same
-x no sig? score? (they are small, might not matter? every little bit helps ... ?)
--DO save SIG and send later
*- in JS data structure in HTML, we include truncated comment offset
*√ add id="..._$cid" to each of above elements that don't have one already,
so we can fill them out later, unless we can prepend/append to existing element
* when comment made full, we fetch missing data and fill it in
-√ straightforward Ajax/JSON fetch
- except "commentSub" stuff which is generated dynamically (just need cid/pid/sid)
* make extra-sure we have security in place: we will be giving people
ability in JS to put mod/delete elements on a comment they should
not have access to, which is no big deal, as long as we don't let them
*use* them
- also except comments: using comment offset in JS data structure, we
fetch just the substr of the comment after that offset, then
strip the existing comment from that offset on, and append the fetched data
-x also except maybe sig/score: if they are already there, just show them
* update "read rest of this comment" to do similar thing
* set longcomment default size to 2048 if it is not already
* make subject : "..." in JS refer to a PID and not include actual text?
- have to recursively grep comment tree in jsSelectComments?
* Fetching hidden comments: get whole thing? (for now, yes)
*** what happened to the call to kidHiddens() ... ?
CSS fixes for MSIE 7 and Opera
can we / should we remove threshold/commentsort/mode from comment links? we could save a lot of bandwidth right there ...
add newlines to QUOTE button results
d2 progressive updating
√skip hiddens
√load on-demand
√show "loading" text
√css
√slashdot css
√test other browsers
√MSIE 7 for Windows
√cleanup
√hard disp stuff for /.
d2 TODO:
√fix abbrev. subjects for & vs. & (put strip_literal version in JS?)
√strip zoo icons from hard dispComment
√only use the comments for this part of the discussion, for the thresholds, and for the comments dump! smarter selectComments?
√d2 hiddens problem
search
√caching search object ... much quicker expiry one minute?
√accepted/rejected
why no Ajax?
"finish"
sorting (index, no tokenizing; make sure numbers can be searched lexically too)
-popularity
-activity
-creation date
filtering (index, no tokenizing, unless "content")
√keyword? ("content"?)
-creation date (before/after?)
√category
√type
√section
√author
cache filters?
content
√introtext
√bodytext
√title
√author
√section?
√topic?
√note?
√AND not OR!!
reskeys
√static reskey always on page ... ?
√only ONE static reskey for all, since all the same
√template for it
other crap
wide characters in email
rss titles with entities
http://www.intertwingly.net/blog/2006/10/01/Syndication-Interop
http://www.franklinmint.fm/blog/archives/000877.html
d2
base-36 for cids?
threadbar
top bar
also other sidebar fixes?
then:
-expanding/contracting
expand how now?
progressive updating
dynamic updating
new ajaxy comments (reskeys, preview)
"reply-to" "quoting"
bugs:
-hidden comments on same level:
1 hidden comment
4 hidden comments
2 hidden comments
ALL mod dropdowns disappear if last modpoint used?
rss
"fix" non-ascii titles
messaging
jabber!
rss
T
nate, refresh, misc. junk
W
nate, wes, refresh
R
nate
F
M
SEARCH TODO:
modify items
delete items
update/sync index
D2 TODO:
NO SCROLL when page changes
requires identifying where we are in page?
new content:
updated scores
new comments
noscript for moderate button for non-admins
replyto quoting
D2 DONE?:
threshold controls for NOSCRIPT
D2 OTHER:
difference between collapse a thread, and contract a comment?
two adjacent "children" with different parents
global keybindings?
Opera bug
√ parent link, make parent full, and jump to it?
- how to get back?
√ make work without JS
√kidHiddens()
√change threshold
√replyTo()
√setFocusComment()
√readRest()
malda8859: if pid.subject = cid.subject, replace subject with very small icon... thus freeing up more space on the right hand side for text.
malda8859: I think that's the answer.
Pudge: that's a hell of a lot of extra processing. need to figure out where to do that processing.
malda8859: or how about this:
malda8859: no icon...
malda8859: Simply 'Re:'
malda8859: we abbreviate.
malda8859: Thats easy.
malda8859: if the subject is Re:(.*) then trunc the (.*) in one-liner mode.
Pudge: one one of those 1000+ comment pages ...
Pudge: i'll put it on TODO and ponder
ajax reply (and reskeys)
ajax moderation? (and reskeys)
poll for new comments
Pudge: the first is updating the code to handle HumanConf and increasing speed limits etc.
Pudge: the second is coming up with the proper limits.
Pudge: the third, which we have to do one way or another, is stripping out the existing formkeys and replacing with reskeys, doing code reorg, and that is what i think you're thinking of as wasted effort. and i think only a small part of it will be wasted, if any. but i am not sure.
journal:
editor format
journal:
vertical align headings on left
bad logtoken google thing
form -> tied/object?
reskeys!!
searchtoo test
taint
journal/article thing
reskeys : search
**** !!!!NOBR/WBR!!!! *****
SEARCHTOO!!!!
RESKEYS!!!!
√ task to flush em
√ test em
√ error handling
* graduated limits
* if blocked out of one, blocked out of another
* ACLs?
* checks needed
* HumanConf
* need reskeys
√ Poll
√ Journal
* User
√ Submit
* Comments
√ Zoo
journal/article
* task to submit journal articles
* admin.pl to use existing discussion
* notify journal user
* how to handle journal deletions
* update discussion timestamp
* story not posted too early?
* discussion doesn't end too soon?
####
just so I don't forget- currently journals have no concept of
introtext/bodytext... now we don't want that, but we DO need the
concept of a leading sentance. Essentially something between the
headline that we use in Older Stuff, and closer to what we use in the
book review Slashbox.
####
ac banning proxy (requires login.pl)
registry loader on scl/sc/sd
check vhost value
make sure all proper scripts are loaded
also check modules being loaded in slash.conf
wiki
image server docs, startup script
site-control
permissions
scheduling, CVS
dboff support for scl, sc, sd
RSS
subscribers
full text
dynamic
use etag, send 304
respect index prefs
respect http vs. https
dock subscriber points?
show more than 10 stories (limit? pref?)
ban abusers
story sharing with slash sites
stories on multiple sites
Message Bugs
Stats Features
templates
cache to disk?
static strings in separate table?