@# Quotes DB     useful, funny, interesting





Google
 
Web www.quotesdb.info
Undernet  |  EFnet  |  Quakenet  |  Freenode  |  Dalnet  |  Ircnet  |  Galaxynet
Page: 1 2 3 4 5 6 7 8 9 10



Comments:

<0> In all honesty, MySQL doesn't know what marble was put in last. You have to 'tag' it somehow, like an autoinc or timestamp.
<1> the last INSERT INTO tabla (col1, col2, ...) VALUES('valor1', 'valor2', ...);
<1> ok
<2> hmm..
<2> well what you can do is... use search max id
<2> since you have it auto increment
<2> the max id must be your last entry since it goes up when you add a new entry
<0> !man last insert id
<3> (How to Get the Unique ID for the Last Inserted Row) : http://dev.mysql.com/doc/mysql/en/Getting_unique_ID.html
<0> Also look at that.
<1> thanks for all
<4> Hello people
<0> Hello aliens
<4> :-)
<4> I have a p***word question, if you please, regarding the old-vs-new scheme
<0> WHAT ABOUT IT?



<5> your question-withholding abilities frighten and excite me
<6> Hi Darien :)
<0> RUN DARIEN RUN!
<4> I have a database with p***words stored using the p***word() function. most of them were generated under MySQL 4.0 with old-p***words, and some with the new p***word() call
<4> I'm using pam-mysql for authentication, and with some trickery I can get both types to auth correctly
<4> now the problem - how do I do the same in plain SQL ?
<0> oh baby
<6> :O
<6> omfg
<5> wait a minute
<7> quit stalking me, Daveman
<5> do NOT tell me you're using p***word() outside of setting the p***word for MySQL users
<6> You stop stalking me!
<6> plus I don't even like you, infibot :)
<6> heehee
<6> haha Darien :P
<6> they triiiiiiiiiiiiicked youuuuuuuuuuu :P
<4> anybody ?
<6> haha, silly noobs :)
<7> DELETE Daveman FROM #mysql;
<5> Guss77: dude, I just addressed you
<6> infi, p***.
<5> Guss77: are you using the p***word() function outside of the mysql.users table?
<6> infi: and fyi, you need to unlock the row(s) in question first, noob.
<6> ;p
<4> I am. why ?
<7> heh
<5> DO NOT DO THIS
<5> !m Guss77 p***word
<3> Guss77: (Access Control, Stage 1: Connection Verification) : http://dev.mysql.com/doc/mysql/en/Connection_access.html
<4> sorry - I'm terribly lagging.
<4> yea, I know all about that
<6> Darien, how's CPUnerd doing? :)
<5> hmm
<5> that's not what I wanted
<5> anyway, the manual states very explicitly
<5> p***word() is designed for use in the MySQL authentication scheme ONLY
<5> it is NOT AN ACCEPTABLE P***WORD MECHANISM FOR APPLICATIONS
<5> endcaps
<5> for that you should be using MD5 or (preferably) SHA-256
<5> I don't mean to be a ****, but I must stress this very strongly
<5> Guss77: I would strongly suggest switching to using SHA() for p***words
<4> currently its not an option. I have a large userbase to maintain and I can't get them all to log in and change the p***word
<5> you don't necessarily have to
<4> ahmm. please explain ?
<5> you could do it completely transparently
<5> add another p***word column
<5> then, when a user logs in, you check
<4> no I can't because I don't have the cleartext
<5> listen to me
<5> stop talking
<5> add another p***word column
<5> when the user logs in you check to see if their row has an entry in the new SHA1 column
<5> if so, you do a SHA hash and check that
<5> if not, you check the old p***word format(s)
<5> if those match, then the user is authenticated
<4> hmm.. interesting. I actually don't need another column for that. (I currently use both new and old p***words in the same column. I can add sha1 or whatever)
<5> now you have the cleartext, you generate the SHA1 hash, and then remove the old entries from the database
<5> you could do that too



<4> problem is - I'd need to patch pam-mysql to do that :-(
<5> why?
<4> I don't think it has that option
<5> to sha hash?
<4> to do the update
<8> is there an efficient way using sql to aggregate datetime values into 'hour' or 'day' bins? i want to aggregate values that may be added during an hour into an hourly total
<4> basically - I have to wait for a user to login using the old p***word - they can only do this using pam-mysql currently.
<4> so I have to "fix" pam-mysql to do the replacement for me.
<5> aha
<5> that's true
<4> hmm.. or easier - I can get it to log the clear text and then have something analyze the logs and do the update myself.
<5> true
<4> heh - that would work :-) thanks !
<5> security implications
<4> of course, always :-)
<5> I recommend storing the data in memory for a while and then writing it to an encrypted file on disk
<5> via gpg
<5> but that's me
<5> realistically, you could log to a line printer and no one would be the wiser
<4> I agree with your security analysis, but that would be to much work to put in a simple hack
<5> true
<4> currently pam-mysql offers logging to a database table, so I can override that to log the clear text into a table only pam-mysql has write access and only me have read access.
<5> nice
<5> hey
<4> cmorgan: you can use temporary tables for that
<4> hey ?
<8> Guss77: sure
<5> you could use DES_ENCRYPT() and DES_DECRYPT()
<8> Guss77: the question is how to get the time ranges
<5> that only works if you have SSL support though
<4> cmorgan: there are some useful date functions in the manuak
<9> cmorgan: select hour(now())
<4> problem is, DES is almost clear text for anyone with a modern computer
<10> cmorgan: group by date(created_date) ?
<8> Xgc: i was planning to run this on a daily basis
<5> Guss77: *almost* :p
<4> although, probably not on texts as short as a p***word.
<5> you could use AES
<8> jpm_: works for grouping by hours?
<10> sure, it's just sql
<10> why wouldn't it
<4> Yes - I can do that :-) thanks a lot !
<5> :)
<5> use a binary p***word too, then they won't even be able to run 'strings' on the binary to get the p***word out to decrypt
<5> unicode ftw
<9> cmorgan: The point is you can group by that type of expression.
<11> hello everyone, I have two tables, each post can have multiple pictures, i wanted to know if theres an easy way with few sql queries to grab the 15 most recent posts from posts and randomly select 1 picture thats ***igned to the same post_id is this possible with one sql query?
<11> Table: posts (post_id, title, description, post_time)
<11> Table: pictures (post_id, picture_id, filename)
<8> Xgc: i'm not seeing it
<9> cmorgan: SELECT ... FROM table1 GROUP BY hour(field1)
<5> bedtime
<5> Guss77: enjoy :)
<4> thanks :-) good night
<8> Xgc: that will group things from the same hour on multiple days right?
<9> cmorgan: Yes. But you can add grouping fields, if you wish.
<8> Xgc: right
<4> Squee: I think so. IIRC Mysql has a rand function
<8> Xgc: how to group by multiple things? group by date(datetimefield), hour(datetimefield) ?
<8> Xgc: in the same query that is(sorry to be confusing)
<9> cmorgan: That's the right form.
<12> )
<8> Xgc: thats pretty neat. let me try that out. i was worried i'd need do issue dozens or hundreds of queries combined with logic on the client side(php server in this case), this looks nice
<9> cmorgan: No need for that.
<11> guss77: yeah there's a random function but how to do a sub query with a random query is where i have no idea.
<4> Squee: heh - I can see your problem :). gimme a minute
<11> sure thing.
<8> Xgc, Guss77, jpm_: thanks for the help guys
<4> Squee: you want a random row. thats the problem
<4> I thought to use limit, but it doesn't accept function output
<4> what I could do - of the top of my head, is make your query (with the right join) and dump it into a temp table with an additional row of random number, then get the highest random value. or something like that
<1> ok i have a way for take only the last entry thank for all :)
<4> but it sounds clunky. I'm sure there's a better way


Name:

Comments:

Please enter the result of the sum 63 + 46 (to avoid spam):






Return to #mysql
or
Go to some related logs:

#perl
opsec fedora
outline dragging ubuntu
Creative VF0040BP Linux
linux restart the process if it dies
#perl
debian+smtpd_check_rules
#math
ubuntu uninstall vnc4server
Gtk-WARNING **: cannot open displa



Home  |  disclaimer  |  contact  |  submit quotes