| Time |
S |
Nick |
Message |
| 12:24 |
|
kados |
hey indradg |
| 12:28 |
|
indradg |
kados, did u guys hold the bug triage meeting? I was stuck in traffic |
| 12:29 |
|
indradg |
took me 2 hrs to get back home instead of the regular 20 minutes drive |
| 12:31 |
|
indradg |
guess i missed it |
| 12:32 |
|
slef |
yep, I nearly did too |
| 12:44 |
|
kados |
indradg: yep ... that's ok ... bad traffic eh? |
| 12:45 |
|
indradg |
yeah...! |
| 14:10 |
|
kados |
slef: can you attend another bugsquashing session later this week? |
| 14:11 |
|
kados |
indradg: you too? |
| 14:11 |
|
kados |
(I'm thinking we'll move the time this time so that NZ can attend) |
| 14:33 |
|
slef |
kados: probably. Schedules are still unclear. |
| 15:02 |
|
kados |
slef: is there a time you are sure _will_ work? |
| 15:48 |
|
indradg |
kados, was away |
| 15:49 |
|
indradg |
this week is tough for me.. will be travelling... weekends are free though |
| 15:52 |
|
kados |
cool |
| 16:02 |
|
slef |
kados: not at the minute. It's been a bank holiday here, so a bit tricky to schedule. :-/ |
| 17:23 |
|
GenjiZzZz |
morning |
| 17:23 |
|
GenjiZzZz |
kados, you still here? |
| 17:23 |
|
kados |
GenjiZzZz: yep ;-) |
| 17:23 |
|
kados |
Genji: so is 969 fixed? |
| 17:24 |
|
Genji |
Yes, 969 was fixed as soon as i said it was. I haven't tried it yet but i know its fixed. |
| 17:24 |
|
kados |
Genji: cool ... I'll try it in a bit |
| 17:30 |
|
kados |
Genji: you don't need to install head to add stuff -- just check out the HEAD from the CVS repo, make your changes, and commit |
| 17:30 |
|
Genji |
for testing the changes, i do. |
| 17:31 |
|
kados |
Genji: right ... have you seen the 'updating koha ' doc on kohadocs.org? |
| 17:31 |
|
kados |
Genji: http://www.kohadocs.org/Updating_Koha.html |
| 17:33 |
|
Genji |
symlinking.. right.... can't do that with windows. |
| 17:34 |
|
Genji |
or can you.... |
| 17:34 |
|
kados |
Genji: hehe not that I'm aware of ;-) |
| 17:34 |
|
kados |
Genji: don't you have a spare pc hanging around you can install fedora on? |
| 17:34 |
|
kados |
Genji: :-) |
| 17:35 |
|
Genji |
might scrounge up the hdd space to install coLinux sometime. |
| 17:36 |
|
Genji |
whats Fedora? |
| 17:36 |
|
kados |
Genji: it's redhat's free linux version |
| 17:45 |
|
Genji |
loads of windows symlink progs. |
| 18:04 |
|
Genji |
okay. rebooting router. seems the dns caching has stuffed up. |
| 19:01 |
|
Genji |
hiya kados, still here? |
| 19:08 |
|
kados |
Genji: yep ... what's up? |
| 19:13 |
|
Genji |
just symlinking.. or hardlinking... maybe, the kohahead. |
| 19:13 |
|
kados |
sweet |
| 19:14 |
|
kados |
I'd symlink rather than hardlink |
| 19:14 |
|
Genji |
how come? |
| 19:14 |
|
kados |
hardlinks are a bad idea unless you know what you're doing ;-) |
| 19:14 |
|
Genji |
meaning? |
| 19:22 |
|
kados |
Genji: here's a good desciption of links in windows: |
| 19:22 |
|
kados |
Genji: http://shell-shocked.org/article.php?id=284 |
| 19:22 |
|
kados |
Genji: hardlinks and softlinks (symlinks) in Windows are the same as in Linux (basically) |
| 19:23 |
|
kados |
links are a filesystem thing not an operating system |
| 19:24 |
|
kados |
thing |
| 19:24 |
|
kados |
:-) |
| 19:45 |
|
Genji |
ah.. ntfslink creates symlinks. |
| 06:08 |
|
Genji |
whats the matter? |
| 06:08 |
|
paul |
??? |
| 06:08 |
|
paul |
Genji never goes to bed... |
| 06:09 |
|
Genji |
hey, i do. |
| 06:09 |
|
Genji |
went to be from 4am to 10am today. |
| 06:09 |
|
Genji |
~s/be/bed/ |
| 06:13 |
|
Genji |
project of the night, updating my sidebar feature for HEAD and committing it. |
| 06:19 |
|
paul |
great ! |
| 06:26 |
|
Genji |
updating my HEAD. |
| 06:49 |
|
kados |
Genji: let me know when it's updated |
| 06:55 |
|
paul |
hi kados, good morning. |
| 06:55 |
|
paul |
did you test marc_Tword script ? |
| 06:56 |
|
kados |
paul: I tried ;-) |
| 06:56 |
|
paul |
and ? |
| 06:56 |
|
kados |
paul: 2820000 in 17319.9062049389 s |
| 06:57 |
|
kados |
that's as far as it got and for some reason my machine rebooted :-( |
| 06:57 |
|
paul |
the" builder" ? |
| 06:57 |
|
kados |
yep |
| 06:57 |
|
paul |
how many lines do you have in marc_subfield_table ? |
| 06:57 |
|
kados |
hang on a sec |
| 06:58 |
|
kados |
5606920 |
| 06:59 |
|
paul |
mmm... then we have a problem |
| 06:59 |
|
kados |
I don't see a marc_tword table either |
| 06:59 |
|
paul |
because even if it reaches 5 600 000, it will be in a long, long long time |
| 06:59 |
|
paul |
(you have to create it manually) |
| 06:59 |
|
paul |
before running the script. |
| 07:00 |
|
kados |
ahh right |
| 07:00 |
|
kados |
I was looking on the wrong machine anyway ;-) |
| 07:01 |
|
kados |
I think building marcTword is slower than building Plucene index ;-) |
| 07:01 |
|
paul |
even if it's the case, the question is : during search, is plucene or tword faster ? |
| 07:01 |
|
kados |
did you get a chance to read the performance mysql book? |
| 07:02 |
|
kados |
paul: tword for sure! |
| 07:02 |
|
paul |
you also could try to replace line 108 by : |
| 07:02 |
|
kados |
(I suspect) |
| 07:02 |
|
paul |
$largehash{$localkey}.=",".substr($title,0,15)."-$biblionumber"; |
| 07:02 |
|
kados |
ok ... I'll do that |
| 07:02 |
|
paul |
(to index only 15 characters, & lower RAM usage) |
| 07:02 |
|
kados |
right |
| 07:03 |
|
paul |
& you're sure you were in READING part, not in WRITING ? |
| 07:03 |
|
paul |
(if you were writing, you should have some lines in marc_Tword |
| 07:03 |
|
kados |
reading for sure ... |
| 07:03 |
|
kados |
but the machine did reboot so it's not really a fair test |
| 07:04 |
|
kados |
I'll rerun it in a second here |
| 07:04 |
|
paul |
wait a minut. |
| 07:04 |
|
kados |
ok |
| 07:04 |
|
paul |
to be so slow, you may have eat too much RAM |
| 07:04 |
|
paul |
& use disk as ram |
| 07:04 |
|
paul |
how many RAM do you have ? |
| 07:05 |
|
paul |
(i mean free) |
| 07:05 |
|
Genji |
what does "cvs commit: sticky tag 'HEAD' for file '...' is not a branch" mean? |
| 07:05 |
|
paul |
because the reading part is just parsing all the table once, so it should be fast |
| 07:06 |
|
kados |
paul: on my test machine I have 256M |
| 07:06 |
|
paul |
genji : head is not a branch, it's "head" |
| 07:06 |
|
paul |
maybe a little too little. |
| 07:06 |
|
kados |
paul: with a 2 gig swap |
| 07:06 |
|
Genji |
k, how you commit to head, then? |
| 07:06 |
|
paul |
let's try with only 15 chars & look at free ram with a top |
| 07:06 |
|
kados |
ok |
| 07:06 |
|
paul |
just cvs commit genji. |
| 07:07 |
|
Genji |
trying, but wincvs isnt. |
| 07:08 |
|
kados |
Genji: it probably won't commit to HEAD until you've checked HEAD out |
| 07:08 |
|
kados |
Genji: can wincvs maintain multiple CVS repos simaltaniously? |
| 07:08 |
|
kados |
paul: SELECTING ... |
| 07:09 |
|
Genji |
unknown. |
| 07:11 |
|
paul |
how i have multiple CVS repos : |
| 07:11 |
|
kados |
paul: Mem: 249024k total, 244564k used |
| 07:11 |
|
paul |
* getting a head copy. appears in Koha directory |
| 07:12 |
|
paul |
* renaming directory from "koha" to "koha_head" |
| 07:12 |
|
paul |
* getting a 2.2 copy. appears in koha directory |
| 07:12 |
|
paul |
* renaming directory from "koha" to "koha22" |
| 07:12 |
|
paul |
now, you can update with |
| 07:12 |
|
paul |
cvs update -d when in directory koha_head or koha22 |
| 07:13 |
|
paul |
and don't bother with branches, as they are fixed in directory name ;-) |
| 07:13 |
|
paul |
kados : that's probably the problem : no more RAM, so swapping... |
| 07:14 |
|
paul |
(I could try to download you database & test the script if you give me an address where I could download a sql.gz) |
| 07:15 |
|
Genji |
okay, its 12:15am. im going to bed, ill worry about this problem in the morning. |
| 07:16 |
|
kados |
paul: do you just need marc_subfield_table? |
| 07:16 |
|
paul |
nope, all your database (except marc_word) |
| 07:16 |
|
kados |
Genji: :-( /me was hoping to get the sidebar working on his installation ;-) |
| 07:17 |
|
paul |
mmm... kados |
| 07:17 |
|
kados |
paul: now we are here: 1800000 in 588 |
| 07:17 |
|
Genji |
kados: if i can't get wincvs doing what i want, ill probably have to install coLinux so i have a better support base, amoungst #koha |
| 07:17 |
|
paul |
marc_biblio & marc_subfield_table should be enough for this script. |
| 07:17 |
|
kados |
lots of this: |
| 07:17 |
|
paul |
great. seems you still haven't reach your RAM limit |
| 07:17 |
|
kados |
Use of uninitialized value in substr at build_marc_Tword.pl line 108. |
| 07:17 |
|
paul |
I know. don't bother, it's just a POC |
| 07:18 |
|
paul |
(you will have mysql error too, because of accented chars, if you have) |
| 07:18 |
|
kados |
also this: |
| 07:18 |
|
kados |
Use of uninitialized value in substitution (s///) at build_marc_Tword.pl line 102 |
| 07:18 |
|
kados |
FYI |
| 07:18 |
|
kados |
(I know it's not important now ;-)) |
| 07:18 |
|
kados |
1980000 in 678.928534030914 s |
| 07:19 |
|
kados |
should take quite a long time |
| 07:19 |
|
paul |
and there is still no disk activity |
| 07:19 |
|
paul |
? |
| 07:20 |
|
paul |
I have an other idea to avoid swaping, if needed. |
| 07:20 |
|
paul |
doing 1xx, then 2xx, then 3xx... marc tags. |
| 07:20 |
|
paul |
should lower the hash size. |
| 07:20 |
|
kados |
right ... |
| 07:21 |
|
kados |
select count(*) from marc_Tword; = 0 |
| 07:21 |
|
paul |
normal : everything is written after the read |
| 07:21 |
|
paul |
so, no write until 5 600 000... |
| 07:21 |
|
paul |
then only write & no read |
| 07:22 |
|
paul |
(a little brute force I agree...) |
| 07:22 |
|
kados |
ok ... so it will take a while |
| 07:22 |
|
paul |
yes. You have my timing in koha-devel mail, for 1 900 000 lines DB |
| 07:22 |
|
kados |
right |
| 07:22 |
|
paul |
(did you also complete the %ignore_list to fit your needs ?) |
| 07:23 |
|
paul |
(you may, for this poc, remove isbn / issn that needs a lot of hash place, for only 1 biblio for each entry. |
| 07:23 |
|
kados |
paul: yep |
| 07:23 |
|
paul |
ok, so, let's wait until it's done... |
| 07:24 |
|
kados |
sounds good |
| 08:06 |
|
paul |
salut jean, et bienvenue |
| 08:06 |
|
jean |
bonjour paul |
| 09:26 |
|
tim |
Hi everyone. |
| 09:26 |
|
tim |
Is there a way to reserve by title? |
| 09:27 |
|
tim |
Looking at the reserve table I see an itemnumber, but our reserves are by title and I'm not sure how I'd move that to Koha. |
| 09:32 |
|
kados |
paul: 2820000 in 8351.3491 FYI ;-) |
| 09:33 |
|
kados |
total used |
| 09:33 |
|
kados |
Mem: 243 238 |
| 09:33 |
|
kados |
Swap: 1926 786 |
| 09:34 |
|
kados |
(I need to get a test machine with more RAM ;-)) |
| 09:37 |
|
kados |
tim: Koha 1.2 used to reserve by title |
| 09:37 |
|
kados |
tim: we need to revise reservations quite a bit |
| 09:38 |
|
tim |
Anyone have any tips on how reserves are moved to Koha then? |
| 09:38 |
|
kados |
for instance, there is no way to reserve a specific item (for staff to fix damaged items, etc. while still allowing the patron who found the item to check it out) |
| 09:38 |
|
kados |
have you seen the migration docs on kohadocs.org? |
| 09:39 |
|
kados |
that's where the most detail on how to move transaction data is located |
| 09:41 |
|
tim |
I would've found out more if I just turned the page. |
| 09:42 |
|
tim |
Still not much help. |
| 10:45 |
|
paul |
kados/joshua : how to solve this problem ? |
| 10:46 |
|
paul |
do you want that I build a builder 1 hundred by 1 ? |
| 10:46 |
|
paul |
could it solve your problem you think ? |
| 11:04 |
|
slef |
whois btoumi and why do they keep doing cvs imports? |
| 11:08 |
|
paul |
btoumi is one of the dev from SAN ouest provence. |
| 11:08 |
|
paul |
they begin with CVS, so they do strange things... |
| 11:08 |
|
paul |
(can you explain what they try to do to me pls ?) |
| 11:09 |
|
paul |
beeping slef |
| 11:09 |
|
slef |
peux-tu lui LARTer? |
| 11:09 |
|
paul |
??? |
| 11:09 |
|
paul |
LARTer ? |
| 11:09 |
|
slef |
regarde http://sourceforge.net/mailarc[…]692&forum_id=9203 |
| 11:09 |
|
slef |
et plusiers fois encores |
| 11:10 |
|
paul |
je sais, je suis abonné à la liste. |
| 11:10 |
|
paul |
mais je ne sais pas à quelle genre de commande CVS cela correspond |
| 11:10 |
|
slef |
il donne "cvs import ..." et je sais pas quoi |
| 11:11 |
|
slef |
je pense qu'il veut dire "cvs add ..." mais je sais pas |
| 11:12 |
|
slef |
bah, my technical french isn't up to it |
| 11:12 |
|
slef |
I think he's trying to add a file, but import is for new modules, not new files or directories |
| 11:13 |
|
slef |
can you at least persuade btoumi to add a real name to their sourceforge profile, please? |
| 11:18 |
|
paul |
slef, i send him a mail immediatly (note that francoisl, from Ineo company will be their coordinator for devs) |
| 11:22 |
|
paul |
mail sent (cc to you) |
| 11:24 |
|
kados |
paul: which problem do you mean? not able to reserve a specific item? |
| 11:25 |
|
paul |
no, running Tword script. |
| 11:25 |
|
kados |
paul: ahh ... |
| 11:25 |
|
paul |
(as he seems to have died once again isn't it ?) |
| 11:25 |
|
kados |
here's where it is: |
| 11:25 |
|
kados |
3120000 in 14748.058 |
| 11:25 |
|
kados |
no it's not dead yet |
| 11:25 |
|
paul |
ah, ok, so it continues... |
| 11:25 |
|
kados |
yep |
| 11:25 |
|
paul |
let me know if it goes to the end. |
| 11:25 |
|
kados |
(probably not today ;-)) |
| 11:26 |
|
paul |
time to leave for me, so no problem ;-) |
| 11:26 |
|
kados |
it seems to be slowing down as it gets larger |
| 11:26 |
|
kados |
ok ... we'll talk tomorrow ... have a good evening |
| 11:26 |
|
paul |
the side effect of the swap, probably |
| 11:26 |
|
paul |
ok, drop me a mail when you go to bed to tell me where you are. |
| 11:27 |
|
paul |
if not ended, i'll rewritte the script to divide it Xxx |
| 11:27 |
|
paul |
(not tommorrow, as i won't be here) |
| 11:27 |
|
paul |
ok, have a good day. |
