fpdb/docs/known-bugs-and-planned-features.txt
2008-09-22 03:31:33 +01:00

143 lines
7.3 KiB
Plaintext

Everything is subject to change and the order does not indicate priority. Patches for any of these or other features are very welcome, see readme-overview.txt for contacts.
Please also see db-todo.txt
alpha4 (release 25Sep-2Oct)
======
graph: update dependencies.txt, doesnt remove old graph on refresh
print a "press any key" thing after we print the traceback. That way it is easy for them to see the error message.
pgsql recreate doesnt work, and it may not store version field on creation if using sql file with pgadmin.
reading small blind wrong for PS 25/50ct
check we're reading mucked cards from PS
newsletter&mailing list
update requirements to include new pgsql interface lib
ebuild: support pgsql
fix HUD config location and update release script accordingly
windows integrated installer
update install-in-gentoo on website
update ebuild and ubuntu guide for HUD_config.xml
store raw hand in db and write reimport function using the raw hand field
make sure totalProfit shows actual profit rather than winnings.
update abbreviations.txt
export settings[hud-defaultInterval] to conf
fill check-/call-raise cache fields
printhand each and the 2/3 relevant printplayerflags respectively on ps-lhe-ring-successful-steal-by-cutoff.txt and ps-lhe-ring-call-3B-preflop-cb-no2b.txt
change to savannah?
implement steal and positions in stud
anonymiser script to generate testdata without making a dozen find&replace all... remember to replace hand no with running no
Import draw (maybe without HudCache for a start)
table with data for graphs for SD/F, W$wSF, W$@SD
separate db table design version and last bugfix in importer
change tabledesign VALIGN and add link to webpage
finish updating filelist
finish todos in git instructions
debian/ubuntu package http://www.debian.org/doc/maint-guide/ch-start.en.html
howto remote DB
move all user docs to webpage
contributor list on webpage
finish bringing back tourney
No river stats for stud games
hole/board cards are not correctly stored in the db for stud games
HORSE (and presumably other mixed games) hand history files not handled correctly
Some MTTs won't import (rebuys??)
Many STTs won't import
before beta
===========
make linux use /etc/fpdb for config first, then ~/.fpdb.
FTP file with only one partial hand causes error
No Full Tilt support in HUD
HUD stat windows are too big on Windows
HUD task bar entries on Windows won't go away
MTT/STT not tested in HUD
HUD stats not aggregated
Player names with non-Latin chars throw warnings in HUD
HUD doesn't start when fpdb is started from the Windows "Start Menu"
Exiting HUD on Windows doesn't properly clean up stat windows
ebuild: USE gtk, set permissions in it, copy docs to correct place, use games eclass or whatever to get games group notice, git-ebuild, get it into sunrise
make hud display W$SD etc as fraction.
add dedicated update page
update status or make a support matrix table for website
move version into seperate file for fpdb gui and db
SD/F, W$wsF, W$@SD too low as reported by daedal in 2+2 forum on 12/13aug
create little test script for people to run to verify successful installation of pydeps
split hud data generation into separate for loops and make it more efficient
fix bug that sawFlop/Turn/River/CBChance/etc gets miscalculated if someone is allin - might as well add all-in recognition for this
make 3 default HUD configs (easy (4-5 fields), advanced (10ish fields), crazy (20 or so))
make it work with postgres
expand instructions for profile file
maybe remove siteId from gametypes
?change most cache fields to bigint to allow extremely big databases in excess of 2 or 4 million hands per stake and position?
rakeback/frequent player points
skins
separate all gui and all processing into files that are named accordingly
ensure that there is only one db handle flying around and that its state is handled properly, ie. by the GUI. i think that might be why we have to reconnect the DB in tableviewer.
why do we have to reconnect in tv.read_names_clicked?
implement error file in importer
catch index error, type error, file not found error
HUD: use different colours according to classification.
move prepare-git.sh and create-release.sh to utils
offer not storing db password
change definition of bet to exclude bring in
fix GUI's load profile
config wizard
file permission script, use games group
make bulk importer display a grand total in the GUI
change save_to_db into one method and probably move into parse_logic
Any comment or print with "todo" in it in the sourcecode except what is marked todo in the menu
make a quick benchmark of mysql and postgresql: import of my whole db, some tableviewer refreshes with and without updated file
Make tab and enter work as sensible in GUIs and implement Ctrl+Q, Ctrl+X and Alt+F4 for close.
use profile file for bulk import and table viewer settings and pathes
handle errors properly, in particular wrt to SQL rollback.
check that we read sitout correctly in: Full Tilt Poker Game #6150325318: Table Bogside
setup database, database-user and permission from GUI.
update prepare-git to check for license header and copyright.
verify at least 2 or 3 sng hands
no rush but before 1.0RC
========================
make DB version error offer reimport, recreation and continue.
In many places there are unnecessary database accesses or it regenerates information it already had before or just generally does things in obscenely inefficient ways. Optimise this
multi-select in bulk importer
cut down action_types array size to appropriate length
make the gui display errors
log file
move directory import code from gui to backend
convert fpdb_import to not require passing "self", generally clean the parameter passing
Doesn't handle Daylight Saving Time (I don't think at least)
ensure that refresh still takes no more than 10 seks on my P3M-800 (a quick run on git15 indicates this is ok now), or 5 with remote DB
select range of stakes and sng/mtt values and types for hud
recognise somewhere if a file is still active and if so keep it open and only read new hands rather than detecting dupes
return full ftp functionality
can wait till 1.x
=================
in all importer: stop doing if site=="ftp", make class constants for site_id instead
It treats fold due to disconnect as voluntary fold which is not ideal
check for unnecessary db.commit()
aliases
repair hands where the seat lines are missing, happens when observing at FTP
flags for storing the reason for winning (best hi, tie for best low, etc.) to DB. not sure actually if this is such a good idea remember that there can be multiple reasons for the same player in the same hand
benchmark properly on mysql innodb, postgresql, more?
rename things like this: ClassName.methodName and variableName. update codingstyle
CLI (not ncurses, normal CLI) equivalent for fpdb.py
optimise/simplify storing by creating the SQL statements depending on hand rather than calling different methods
make range of activeSeats configurable for tv/hud
License
=======
Trademarks of third parties have been used under Fair Use or similar laws.
Copyright 2008 Steffen Jobbagy-Felso
Permission is granted to copy, distribute and/or modify this
document under the terms of the GNU Free Documentation License,
Version 1.2 as published by the Free Software Foundation; with
no Invariant Sections, no Front-Cover Texts, and with no Back-Cover
Texts. A copy of the license can be found in fdl-1.2.txt
The program itself is licensed under AGPLv3, see agpl-3.0.txt