| Time |
S |
Nick |
Message |
| 12:27 |
|
nengard |
hey all - sys pref question - shouldn't OPACUserCSS default to '' ? Right now it's defaulting to '0' ... isn't that printing out a 0 somewhere in the code?? |
| 12:39 |
|
nengard |
hey all - sys pref question - shouldn't OPACUserCSS default to '' ? Right now it's defaulting to '0' ... isn't |
| 12:39 |
|
owen |
If so, nengard, then you're right--that's a bug. |
| 12:41 |
|
nengard |
owen - oops - here's my full question |
| 12:41 |
|
nengard |
hey all - sys pref question - shouldn't OPACUserCSS default to '' ? Right now it's defaulting to '0' ... isn't that printing out a 0 somewhere in the code?? |
| 12:41 |
|
nengard |
i am just a bug magnet |
| 12:41 |
|
owen |
I think no, because the template system will interpret "0" the same as "empty." |
| 12:42 |
|
owen |
So it's a minor bug |
| 12:42 |
|
nengard |
ah - but it should still be an empty string? |
| 12:42 |
|
nengard |
ok |
| 12:51 |
|
thd |
kados ping |
| 12:54 |
|
owen |
kados is on the road, thd, so I don't know if you'll be able to catch him |
| 12:55 |
|
thd |
owen: which road is that? |
| 12:55 |
|
owen |
All of them, I think. He had a huge itinerary |
| 12:58 |
|
thd |
owen: will bug #218 be on your itinerary in future? |
| 12:58 |
|
owen |
218? |
| 12:59 |
|
owen |
thd, do you mean 2189? |
| 13:00 |
|
thd |
owen: #2169 I mean. It was not one that I was thinking of my latest koha-devel list posting but kados had asked me to post that bug almost a year ago. |
| 13:01 |
|
owen |
I would like to see Bug 2169 fixed just as much as you, but it's outside of my expertise. |
| 13:01 |
|
owen |
Bug 2169 has *always* existed, as far as I know. |
| 13:01 |
|
thd |
s/#2\d+/#2189/ # tubule tying today |
| 13:02 |
|
owen |
? |
| 13:02 |
|
thd |
another always existed |
| 13:02 |
|
thd |
2189 |
| 13:03 |
|
owen |
thd, is the primary issue of 2189 that of database field sizes? |
| 13:04 |
|
thd |
owen: yes it is not a bug in itself but has been an obstacle to fixing bugs by raising the number of changes required |
| 13:06 |
|
thd |
owen: although one current bug is directly caused by it perhaps |
| 13:07 |
|
owen |
"The only Koha code which restricts the size of organisation codes to my |
| 13:07 |
|
owen |
knowledge is the hard coding in the templates." |
| 13:07 |
|
owen |
The templates are generally set up to match field sizes in the databases |
| 13:07 |
|
owen |
The database is not configured to match form field sizes |
| 13:10 |
|
thd |
owen: I did not look to see where the problem was I only guessed |
| 13:11 |
|
owen |
If there is a problem with the size of database fields they can certainly be changed, but it's necessary to pick a limit |
| 13:12 |
|
thd |
owen: what then restricts the ability to enter more than a given set of characters in a Koha for field? |
| 13:12 |
|
owen |
It's true that the templates may need to be changed to reflect the database field limits, but it's the database that sets the limit, not the templates |
| 13:13 |
|
owen |
The templates just have a maxlength attribute applied to the input tag |
| 13:14 |
|
thd |
owen: exactly it is the maxlength value which causes the problem |
| 13:14 |
|
thd |
owen: the maxlength value is hardcoded when it need not be |
| 13:15 |
|
owen |
The maxlength attribute should reflect the maximum number of characters allowed by the database field |
| 13:15 |
|
owen |
Unless you're finding that the database field allows enough characters but the form field maxlength is smaller |
| 13:17 |
|
thd |
owen: are you saying that maxlenghth is currently read from the database limit for that field in real time? |
| 13:17 |
|
owen |
No |
| 13:18 |
|
thd |
owen: then it is hard coded as I had presumed? |
| 13:18 |
|
owen |
Yes |
| 13:19 |
|
thd |
owen: my suggestion is to calculate it in real time from the database maximum for the particular field |
| 13:19 |
|
hdl |
thd: hi |
| 13:19 |
|
hdl |
nice to see you |
| 13:19 |
|
atz |
thd: that does not sound very reasonable |
| 13:19 |
|
owen |
Why? Because you suspect libraries will be changing the size of their database fields? |
| 13:20 |
|
thd |
hello hdl |
| 13:20 |
|
hdl |
this would really be MUCH proc-intensive. |
| 13:20 |
|
thd |
owen: because it is self validating and allows changes when necessary |
| 13:20 |
|
atz |
the cataloging module in particular would drag incredibly |
| 13:21 |
|
atz |
it is perfectly valid for a DB field to have capacity greater than the interface makes use of |
| 13:22 |
|
thd |
hdl: well if excessive CPU would be required then we need your idea of auto-building templates from a configuration file |
| 13:23 |
|
thd |
not that that would happen any time soon |
| 13:23 |
|
thd |
non-real-time auto-building |
| 13:24 |
|
thd |
build them whenever the database changes |
| 13:24 |
|
atz |
what fields, specifically, do we think any library would be interested to enlarge? |
| 13:26 |
|
thd |
atz: I mention 2 in bug #2189 as mere examples. Any where enlarging would suit a local need. |
| 13:26 |
|
thd |
owen: #2095 may be a currently problematic instance. |
| 13:28 |
|
thd |
owen: for #2095 maxlength should be no smaller than the maximum size of a MARC record and should not need to be even that small for MARCXML. |
| 13:29 |
|
owen |
thd, your comments are much better addressed to atz and hdl, since you're not talking about a template issue |
| 13:29 |
|
thd |
owen: is 2095 not a template issue? |
| 13:29 |
|
atz |
thd: as a practical matter, I think it would be far easier to address specific problems than to convert wholesale to adding an extra variable for every input field |
| 13:29 |
|
owen |
thd: no, not as far as I know |
| 13:30 |
|
thd |
owen: OK |
| 13:30 |
|
atz |
thd: as I said earlier, it is perfectly valid for the interface to specify a lower limit than the DB field's capacity |
| 13:30 |
|
nengard |
just want to confirm that this sys pref 'RoutingSerials' is talking about serials routing lists? |
| 13:31 |
|
hdl |
nengard: yes |
| 13:31 |
|
nengard |
thanks hdl |
| 13:31 |
|
thd |
atz: the intention was to make Koha easier to customise but why should the interface limit valid data? |
| 13:32 |
|
atz |
because the *limit* could be valid. |
| 13:32 |
|
atz |
not to mention it keeps your server from getting overrun |
| 13:33 |
|
atz |
how is this problem critical? |
| 13:33 |
|
thd |
atz: the first example I describe in bug #2189 is a case where any smaller limit would not allow standards conforming data. |
| 13:33 |
|
atz |
the examples cited are either problems that are already fixed, or ones that *might* occur later |
| 13:35 |
|
thd |
atz: yes, as I told owen a few minutes ago it is not a bug in itself but has been an obstacle to fixing bugs by raising the number of changes required |
| 13:35 |
|
atz |
of course revising the structure that broadly would require an enormous number of changes |
| 13:38 |
|
atz |
for specific fields that we consider more volatile, we might be able to take this approach relatively soon |
| 13:38 |
|
thd |
atz: yes, I presumed that it was easier to make that change once but if the method is too CPU intensive or would require rewriting how the templates used are built for a batch process then I will withdraw the bug to an enhancement for the distant future |
| 13:38 |
|
thd |
atz: which approach? |
| 13:39 |
|
atz |
the dynamic approach |
| 13:39 |
|
thd |
atz: I see for a small number deemed liable to change |
| 13:39 |
|
atz |
i'm saying we might be able to use this selectively, rather than with every input field |
| 13:39 |
|
atz |
right |
| 13:42 |
|
thd |
atz: I guess my thinking was only concerned with values which should not be so small that they could not be used for giving a semantically meaningful name in coded form. |
| 13:45 |
|
thd |
atz: I think that if the few cases of which I am thinking were hard coded at a maximum of 32 characters in the database and the templates you would never see any problem for decades. |
| 13:47 |
|
atz |
thd: that sounds more feasible |
| 13:47 |
|
thd |
atz hdl: if the templates are not the cause of bug #2095 then what would be? |
| 13:47 |
|
atz |
thd: or in the case of framework code and even organization code, perhaps all that is needed is an additional "description" field to allow the longer te |
| 13:47 |
|
atz |
*text |
| 13:49 |
|
atz |
thd: i don't know about #2095. that seems odd. can you give a URL? |
| 13:50 |
|
atz |
or nengard, perhaps you can elaborate? |
| 13:50 |
|
nengard |
atz - off to read it |
| 13:51 |
|
atz |
134 is a weird number indeed |
| 13:51 |
|
hdl |
thd : Some notes fields in UNIMARC are Textarea notes and not input fields. |
| 13:51 |
|
nengard |
what's the question? when I entered in the 520 field it stopped me from typing after 134 characters - i was in the cataloging module when doing this |
| 13:51 |
|
hdl |
Would that be enough for you ? |
| 13:52 |
|
thd |
atz: there is one description field already for the frameworks but it shows what the librarian wants to see and not what they hope not to no about what the framework. A framework memonic is so the person editing the frameworks can be reminded what he is really editing. |
| 13:52 |
|
atz |
nengard: can you give a URL so I know what page to look at? |
| 13:52 |
|
nengard |
atz koha/cataloguing/addbiblio.pl?frameworkcode=BKS |
| 13:53 |
|
thd |
nengard: I did not see this bug myself. Which Koha editor does it affect? |
| 13:53 |
|
nengard |
thd - maybe it cut it off after i submitted it - or maybe it has been fixed |
| 13:53 |
|
nengard |
it has been a while - let me go try to edit a record .... |
| 13:54 |
|
thd |
nengard: no I mean that I did not test it |
| 13:54 |
|
nengard |
oh |
| 13:55 |
|
thd |
nengard: I see by the URL that you mean the guided forms based record editor not the newer biblios record editor. |
| 13:56 |
|
hdl |
atz : it is quite puzzling for me that C4::Context could use C4::Context->preference in its BEGIN block. |
| 13:56 |
|
hdl |
Is it allowed ?* |
| 13:56 |
|
owen |
There are several instances of 'maxlength="255"' in addbiblio.pl, which I assume correspond to database field limits |
| 13:56 |
|
nengard |
thd hdl atz - i just tested in the newest version i have and it is not limited |
| 13:57 |
|
nengard |
i got over 400 characters into the field when editing |
| 13:57 |
|
atz |
hdl: it is an unusual exceptional case |
| 13:57 |
|
nengard |
here's the edit URL: koha/cataloguing/addbiblio.pl?biblionumber=263&op= |
| 13:57 |
|
atz |
see what is happening is that the sub "handle_errors" is being defined |
| 13:57 |
|
thd |
owen atz: most subfields have no limits defined in the MARC standards and should not be limited in any significant way. |
| 13:57 |
|
nengard |
thd++ |
| 13:58 |
|
atz |
that *sub* uses ->preference, but the sub itself will not be called until after the BEGIN block. |
| 13:59 |
|
thd |
owen: what record editor fields are limited to 255 for records? That is too big for any fixed length fields and too small for anything else by the standard at least. |
| 13:59 |
|
atz |
hdl: does that make sense re: Context? |
| 13:59 |
|
hdl |
atz : asking it because on SUN it threw me out. |
| 14:00 |
|
atz |
hdl: interesting |
| 14:00 |
|
atz |
what version of perl? |
| 14:00 |
|
hdl |
5.8.5 |
| 14:00 |
|
mc |
hello all |
| 14:01 |
|
thd |
hello mc |
| 14:01 |
|
hdl |
thd : mc is a BibLibre man. |
| 14:01 |
|
owen |
thd: view source on the Add MARC page to see |
| 14:02 |
|
hdl |
mc: thd is american and worked on MARC21 frameworks. he is really helpful in general use cases. |
| 14:02 |
|
mc |
thx, hdl |
| 14:03 |
|
thd |
mc: do you live in France? |
| 14:03 |
|
mc |
thd, yep |
| 14:03 |
|
mc |
strasbourg |
| 14:03 |
|
mc |
near germany |
| 14:03 |
|
mc |
(right at the frontier, in fact) |
| 14:04 |
|
mc |
hdl: peraps it's time for me to read the perldoc again but BEGIN block is executed asap |
| 14:04 |
|
mc |
when bytecodinf |
| 14:04 |
|
mc |
bytecoding |
| 14:05 |
|
mc |
so you can't use C4::Context in the C4::Context's BEGIN |
| 14:05 |
|
hdl |
atz ? |
| 14:06 |
|
atz |
sorry, on conf. call..... bbl |
| 14:07 |
|
mc |
hdl: i just verified with a simple code |
| 14:07 |
|
mc |
and it is! |
| 14:08 |
|
hdl |
Can you detail ? |
| 14:08 |
|
hdl |
it is what ? |
| 14:12 |
|
mc |
hdl, are you talking about this line : |
| 14:12 |
|
mc |
30 my $debug_level = C4::Context->preference("DebugLevel"); |
| 14:14 |
|
hdl |
yes |
| 14:17 |
|
thd |
nengard: how did you fit 400 characters into a field defined as maxlength 255? |
| 14:17 |
|
nengard |
LOL |
| 14:17 |
|
nengard |
no idea :) |
| 14:17 |
|
nengard |
but it worked |
| 14:17 |
|
nengard |
..... |
| 14:18 |
|
nengard |
504 -BIBLIOGRAPHY, ETC. NOTE |
| 14:18 |
|
nengard |
a Bibliography, etc Includes bibliographical references and indexes. |
| 14:18 |
|
nengard |
520 -SUMMARY, ETC. |
| 14:18 |
|
nengard |
a Summary, etc Learning to Teach in an Age of Accountability, by Arthur Costigan and Margaret Smith Crocco, with Karen Kepler Zumwalt, documents the voices of many new teachers -- important voices, articulate voices, emotional voices. In an age of accountability, these voices bring to light many significant struggles, tensions, and conundrums that exist for them as they enter middle and secondary school environments. ?Teaching Educ |
| 14:18 |
|
nengard |
ation |
| 14:18 |
|
nengard |
this - after i hit submit |
| 14:19 |
|
owen |
If it's a textarea, then there isn't a maxlength |
| 14:19 |
|
owen |
maxlength only applies to input type="text" |
| 14:19 |
|
paul |
nengard: in MARC editor, there is an automatic mechanism that transform a input into a textarea when subfield size is >200 |
| 14:19 |
|
thd |
owen: I see maxlength 255 everywhere |
| 14:19 |
|
paul |
+ all notes fields are automatically set to textarea |
| 14:20 |
|
thd |
paul: I see 255 for some notes fields at least |
| 14:20 |
|
nengard |
paul - i did notice that when adding a new record the field was a text field and when editing it was textarea ... |
| 14:21 |
|
nengard |
is there anyway to make it always a textarea? |
| 14:31 |
|
paul |
nengard: what do you mean by "it" ? |
| 14:32 |
|
nengard |
sorry - 520 $a on addbiblio.pl |
| 14:32 |
|
nengard |
oh - i was editing a biblio |
| 14:32 |
|
nengard |
paul - let me start over |
| 14:32 |
|
owen |
You could even auto-resize your textareas: http://zivotdesign.com/example[…]ery/textarea.html |
| 14:33 |
|
nengard |
when viewing the 520a field when adding a new record on addbiblio it was a text field when editing a biblio it was a textarea |
| 14:35 |
|
hdl |
owen : COOOL ;) |
| 14:36 |
|
owen |
jquery++ :) |
| 14:36 |
|
thd |
owen: those 255 limits to the extent I have seen them would not be enough on some rare occasions like lengthy statements of responsibility or the longest narrative titles given to some 18th century books. |
| 14:37 |
|
paul |
one day, /me will find some time to investigate jquery a lot... |
| 14:38 |
|
thd |
owen: since the cataloguer should know enough not to abuse the system 255 should be something like text but not text area. |
| 14:39 |
|
thd |
s/know/be trusted/ |
| 14:42 |
|
thd |
owen: no arbitrary limit like 255 should ever allow any possible real world record to have truncated subfields even if no one writes long 18th century book titles when writing books today. |
| 14:47 |
|
nengard |
thd - i've seen those long titles today :) hehe |
| 14:48 |
|
thd |
owen: adding a 0 to 255 would be enough but in case some note subfield was accidentally missed 99999 is the maximum size of a MARC record and certainly safe where 255 would not be. We should never have a risk of data loss merely by saving a record into the editor. |
| 14:51 |
|
thd |
nengard: I am waiting for the eighteenth century enlightenment to come back into full fashion so we can get out of the neo dark ages. |
| 14:52 |
|
nengard |
heh |
| 14:57 |
|
owen |
I'm waiting for all new nonfiction titles to stop following the same pattern: "BIG WORD!: how blah blah blah and more blah blah and even more and more buy my book" |
| 14:58 |
|
gmcharlt |
owen: it's a requirement of PhD to sign a blood oath to entitle all monographs that way |
| 14:59 |
|
hdl |
lol |
| 15:03 |
|
nengard |
ehe |
| 15:04 |
|
thd |
The rule for determining which subfields should be notes subfields and therefore textarea is much too simple. A frameworks value would be required to do it correctly. Only the most important ones have been caught. |
| 15:05 |
|
thd |
owen: I propose that 255 be changed to 99999 to avoid a data loss risk. |
| 15:05 |
|
thd |
... for addbiblio.pl . |
| 15:07 |
|
owen |
Since no one would want to edit 99999 characters from a single-line text input field, the correct solution would be to convert all to textareas |
| 15:08 |
|
thd |
owen: 99999 is merely the theoretical maximum for the standard :) |
| 15:09 |
|
owen |
But my point stands: I wouldn't even want to have to edit 255 characters from a single line input field |
| 15:10 |
|
thd |
owen: I guess you are right that if you were cataloguing those very long book titles you would want it to become textarea so you could see what you would be doing. |
| 15:11 |
|
owen |
Which is why it might be good to convert all inputs to single-row textareas and make them automatically resizeable |
| 15:11 |
|
thd |
owen: as long as those are only textarea when the transition point has been passed then making everything textarea should be fine. |
| 15:13 |
|
atz |
lol gmcharlt : you know, sometimes I want to escape from the koha list, too. |
| 15:13 |
|
gmcharlt |
atz: heh - at least we're paid, in effect, to be on it |
| 15:13 |
|
gmcharlt |
;) |
| 15:14 |
|
thd |
being unpaid: I sadly no longer read the koha list |
| 15:15 |
|
owen |
Generally the koha list can be considered the koha-install list |
| 15:16 |
|
paul |
hi gmcharlt. Pls note my comment on bug 2173 |
| 15:16 |
|
gmcharlt |
paul: I've seen the patch - looks OK, but I want to test - hopefully later today |
| 15:17 |
|
paul |
you've seen the patch ? |
| 15:17 |
|
paul |
but hdl send it to me only. hdl, you send it to gmcharlt as well ? |
| 15:17 |
|
thd |
For a period in 2005 tried to give a fulsome answer to every koha list question without being paid but that could not last. |
| 15:18 |
|
gmcharlt |
paul: I thought it was the "restoring startsby patch on authorities" - is that the one? |
| 15:18 |
|
paul |
nope. |
| 15:18 |
|
paul |
it's the "heading-main" search one (search on heading $a) |
| 15:18 |
|
gmcharlt |
thd: flattering Koha users? nice ;) |
| 15:24 |
|
ryan |
gmcharlt: i just noticed yesterday that the '$a only' search does in fact include the entire heading. |
| 15:25 |
|
ryan |
i agree it would be useful for that to work as expected. |
| 15:25 |
|
hdl |
ryan : to reenable it, a new index should be created. |
| 15:26 |
|
hdl |
I reenabled it on UNIMARC. |
| 15:26 |
|
thd |
gmcharlt: how would 2173 be fixed for MARC 21 without searching against special local use field where names were last names would be separated like they are in UNUMARC? |
| 15:26 |
|
hdl |
But would not make it on USMARC since I donot master those DOM files enough. |
| 15:27 |
|
thd |
s/names were// |
| 15:30 |
|
gmcharlt |
thd: simplest would be to make it have a different meaning - search for base heading and ignore subdivisions - although that would not be much different from a left-anchored heading search |
| 15:31 |
|
gmcharlt |
thd: parsing first and last names is not readily supportable in MARC21, of course |
| 15:31 |
|
thd |
gmcharlt: I see yes, left anchored would work for MARC 21. |
| 15:32 |
|
thd |
gmcharlt: The comma is there for MARC 21 so last names could be parsed into a local use subfield for special treatment if needed. |
| 15:33 |
|
gmcharlt |
thd: comma is only a half measure, alas - far too many names, royal titles, and so on, would be edge cases |
| 15:33 |
|
gmcharlt |
thd: on other hand, I'm not sure that a first name search is necessarily needed, at least for staff and cataloging purposes |
| 15:34 |
|
thd |
gmcharlt: Titles etc. already have their own subfields in both MARC 21 and UNIMARC. |
| 15:35 |
|
gmcharlt |
thd: 100 0#$aGustaf$bV,$cKing of Sweden,$d1858-1950 - where's the comma to identify the first name, e.g., |
| 15:36 |
|
thd |
gmcharlt: Is the same code used for finding the authorised form to use in the OPAC when invoking the [...]. |
| 15:36 |
|
ryan |
i was referring, btw, only to the fact that if I search on King in the field marked '$a only' , i'll get gmcharlt's example there in my results. |
| 15:37 |
|
ryan |
i don't see a need to add further indexes on subsets of $a. |
| 15:37 |
|
thd |
gmcharlt: in that case by AACR2 the first name is the most important and the last name is seldom known by searchers. |
| 15:39 |
|
thd |
ryan: If that was the issue then left anchoring on the whole field for names should solve the problem |
| 15:41 |
|
ryan |
Ah, I see. |
| 15:42 |
|
ryan |
The template then needs to be reworded if that's the solution. |
| 15:43 |
|
hdl |
for UNIMARC it is not a solution thogh. |
| 15:43 |
|
hdl |
Our users wants to be able to search on Main Entry. |
| 15:44 |
|
hdl |
But I already posted a patch for that on our sied. |
| 15:44 |
|
hdl |
I can send it to you if you want to have a look at it. |
| 15:45 |
|
gmcharlt |
thd, ryan, hdl: left-anchored is enough for MARC21, so can make searchboxes different for UNIMARC and MARC21 |
| 15:45 |
|
ryan |
yes, I follow now. I haven't looked at the dom indexing definition yet. |
| 15:45 |
|
gmcharlt |
hdl: please send me the patch if you haven't already |
| 15:45 |
|
paul |
gmcharlt: I did it 10mn ago ;-) |
| 15:45 |
|
gmcharlt |
paul: ok, it's *that* one ;) |
| 15:45 |
|
ryan |
gmcharlt: i'm fine with that approach. as long as the interface does what it says i'm happy :) |
| 15:45 |
|
paul |
(although it's a fw: maybe a direct mail would be easier to deal with) |
| 15:46 |
|
gmcharlt |
paul: no, it's fine, git-am is smart enough to deal with it |
| 15:47 |
|
hdl |
ryan : left anchored search would be search starts by |
| 15:47 |
|
hdl |
I sent a patch for it. |
| 15:53 |
|
paul |
tinaburger: are you really here, or is it an auto-connect ? |
| 15:53 |
|
acmoore |
be not afraid, it is her! |
| 15:58 |
|
hdl |
hope you are getting better. |
| 17:40 |
|
thd |
gmcharlt: what is the last message that you have from the koha-devel list? |
| 17:42 |
|
thd |
atz: are you here? |
| 17:53 |
|
danny |
hello #koha |
| 17:58 |
|
atz |
i'm here... had to reset for OS update |
| 18:02 |
|
thd |
atz: what is the last message which you have from the koha-devel list? |
| 18:03 |
|
atz |
Thomas' reply to Koha 3.0 Stable Release Plan |
| 18:04 |
|
thd |
atz: why do I not have it. I checked my subscription on Savannah and it seems fine. |
| 18:05 |
|
atz |
you may have to ask your email provider. |
| 18:06 |
|
thd |
atz: is a LibLime or Katipo mail system involved for koha-devel? |
| 18:09 |
|
thd |
atz: I triggered the alert for the koha users list because I do not subscribe to Koha users with the same address and hope I have not banned myself from receiving on the koha-devel list in consequence. |
| 18:09 |
|
thd |
The Koha users list was in the CC line from Joshua originally. |
| 18:09 |
|
atz |
thd: lists.koha.org => 212.85.152.90 |
| 18:10 |
|
atz |
nslookup 212.85.152.90 |
| 18:10 |
|
atz |
Server: 72.36.190.2 |
| 18:10 |
|
atz |
Address: 72.36.190.2#53 |
| 18:10 |
|
atz |
Non-authoritative answer: |
| 18:10 |
|
atz |
90.152.85.212.in-addr.arpa canonical name = 90.152.85.212.rev-sql.lost-oasis.net. |
| 18:10 |
|
atz |
90.152.85.212.rev-sql.lost-oasis.net name = paulpoulain.com. |
| 18:10 |
|
atz |
so I'm guessing paul is the guy to ask on that |
| 18:11 |
|
thd |
atz: OK so if I have nothing more by Tomorrow morning I will ask paul. |
| 18:24 |
|
hdl |
thd you can ask me. |
| 18:25 |
|
hdl |
thd : I recieved you mails. |
| 18:48 |
|
thd |
hdl: The problem is that I did not receive my usual copy of y own message for the last two in reply to paul and slef |
| 18:48 |
|
thd |
s/ y/ my/ |
| 18:50 |
|
thd |
hdl: mailman had caused problems for me about two years ago |
| 18:51 |
|
thd |
hdl: two years ago mailman said that I was subscribed but had stopped sending me messages. I had to have kados fix my account at the time. |
| 19:16 |
|
dkg |
i'm trying to get koha running against perl 5.10 with CGI::Session 4.30 (debian lenny) |
| 19:16 |
|
dkg |
The installer is giving me this message: |
| 19:16 |
|
dkg |
Can't locate object method "generate_id" via package "CGI::Session::ID::" (perhaps you forgot to load "CGI::Session::ID::"?) |
| 19:16 |
|
dkg |
(i'm using koha via git, up-to-date as of this morning) |
| 19:17 |
|
dkg |
looking through the code, the only invocations of CGI::Session() use serializer:yaml. |
| 19:18 |
|
dkg |
I can coax a test perl file to fail with the same message just by creating a CGI::Session object with serializer:yaml, also, though it doesn't fail if i omit mention of a serializer. |
| 19:18 |
|
dkg |
any thoughts about what next steps to take for the debugging? |
| 19:19 |
|
gmcharlt |
dkg: sounds similar to issue reported for OpenSUSE |
| 19:19 |
|
gmcharlt |
dkg: try add Perl modules CGI::Simple and FreezeThaw |
| 19:19 |
|
gmcharlt |
dkg: also, how did you install CGI::Session - was it from CPAN or from libcgi-session-perl |
| 19:21 |
|
dkg |
libcgi-session-perl |
| 19:21 |
|
dkg |
i'll try those modules, though. |
| 19:23 |
|
dkg |
hrm. i added those modules, but i'm still getting the same error with my test program. |
| 19:23 |
|
dkg |
and the same error with the koha installer, too. |
| 19:24 |
|
atz |
dkg: now go back and make CGI::Session again, iirc |
| 19:25 |
|
dkg |
atz: i installed via aptitude. should i re-install? |
| 19:26 |
|
atz |
hmm... not sure there |
| 19:27 |
|
dkg |
i did "aptitude reinstall libcgi-session-perl", and my test program is still failing, so that's not it. |
| 19:27 |
|
gmcharlt |
hmm - etch has 4.15 of CGI::Session, lenny has 4.30 - perhaps 4.30 has an incompatible API change |
| 19:29 |
|
dkg |
if i modify my test program to omit the ;serializer:yaml, it seems to work fine. |
| 19:29 |
|
hdl |
dkg : I coped with that on SUN. |
| 19:29 |
|
gmcharlt |
dkg: instead of that, try adding ";id:md5" after the serializer:yaml bit |
| 19:29 |
|
gmcharlt |
per patch hdl filed today |
| 19:29 |
|
hdl |
This can be fixed with id:md5 |
| 19:30 |
|
hdl |
Consider adding use CGI::Session::ID::md5 to CGI::Session if it is not enough. |
| 19:30 |
|
atz |
yeah, i saw that patch. seems interesting. |
| 19:31 |
|
hdl |
atz : Indeed, on SUN it was vital. |
| 19:31 |
|
dkg |
if i append that to my test program, i get this error: |
| 19:31 |
|
dkg |
Can't locate object method "generate_id" via package "CGI::Session::ID::md5" (perhaps you forgot to load "CGI::Session::ID::md5"?) at /usr/share/perl5/CGI/Session.pm line 74. |
| 19:32 |
|
dkg |
but if i add "use CGI::Session::ID::md5;" to the top of the file, it completes |
| 19:32 |
|
dkg |
(with a later error message of: |
| 19:33 |
|
dkg |
(in cleanup) Can't locate object method "freeze" via package "CGI::Session::Serialize::yaml" at /usr/share/perl5/CGI/Session.pm line 241 |
| 19:33 |
|
dkg |
) |
| 19:38 |
|
dkg |
hdl: do you think that this represents a bug in CGI::Session? Should we report it to the module authors? |
| 19:38 |
|
gmcharlt |
dkg: if you then add a "use CGI::Session::Serialize::freezethaw" does it help |
| 19:38 |
|
gmcharlt |
dkg: I think it is either an API change, a bug possibly related to Perl 5.10, or perhaps a Debian packaging glitch |
| 19:39 |
|
hdl |
dkg : First, I was surpirsed that CGI::Session on CPAN was an unauthorised version. |
| 19:39 |
|
hdl |
surprised even |
| 19:40 |
|
hdl |
Second, I surely consider those things as bugs. |
| 19:40 |
|
dkg |
gmcharlt: adding "use CGI::Session::Serialize::freezethaw;" did not get rid of the "(in cleanup) Can't locate object..." error message in my test script. |
| 19:40 |
|
hdl |
But I donot know what is the reason for those bugs. |
| 19:40 |
|
gmcharlt |
dkg or hdl: could you file a bug at bugs.koha.org for this |
| 19:41 |
|
hdl |
Look at CGI::Session::Serialize::yaml |
| 19:41 |
|
gmcharlt |
it looks like it will need some time to investigate |
| 19:41 |
|
hdl |
And find if there is freeze method in this Module. |
| 19:42 |
|
hdl |
will do gm |
| 19:44 |
|
hdl |
done. |
| 19:46 |
|
dkg |
http://bugs.koha.org/cgi-bin/b[…]w_bug.cgi?id=2216 |
| 19:46 |
|
dkg |
thanks, hdl. |
| 20:07 |
|
dkg |
I just added a patch against the git head to bug 2216 that works for me to resolve the base issue (i still don't know about the "(in cleanup)" error message) |
| 20:08 |
|
hdl |
dkg : i think that that patch had just been pushed today or will soon. |
| 20:11 |
|
dkg |
hdl: excellent. |
| 20:18 |
|
dkg |
OK, so while the upgrade now works, no user session lasts more than one pageview. I'm assuming this is because the CGI::Session object isn't properly serializing data because of the "Can't locate object..." errors. |
| 20:18 |
|
dkg |
where do i find CGI::Session::Serialize::yaml ? |
| 20:18 |
|
dkg |
it's not in /usr/share/perl5/CGI/Session/Serialize, which is where i'd usually expect it :/ |
| 20:20 |
|
gmcharlt |
it would be /usr/share/perl5/CGI/Session/Serialize/yaml.pm - available in 4.14 (etch package) but not 4.30 (lenny package) |
| 20:22 |
|
gmcharlt |
dkg: ok, from CPAN changelog for CGI::Session, CGI::Session::Serialize::yaml was split out as separate module |
| 20:22 |
|
gmcharlt |
so you'll need to get it from CPAN |
| 20:23 |
|
atz |
gmcharlt: so we'll have to add that as an explicit, separate dependency |
| 20:23 |
|
gmcharlt |
atz: yes, and ask Vincent to package it |
| 21:41 |
|
dkg |
i've filed an RFP for the package in debian, fwiw: |
| 21:41 |
|
dkg |
http://bugs.debian.org/484556 |
| 21:42 |
|
dkg |
i don't know who Vincent is, but if he plans on packaging it, he can close that bug in the first changelog. |
| 21:45 |
|
gmcharlt |
dkg++ |
| 21:48 |
|
dkg |
(using a package built with dh-make-perl of CGI::Session::Serialize::yaml does appear to fix the problem i was having). |
| 21:48 |
|
dkg |
one more annoying question for the day: |
| 21:49 |
|
dkg |
what's the best way to do an import of 10000 marc records with roughly the same number of items? |
| 21:49 |
|
dkg |
can i do this server-side, or do i need to do it through the web ui? |
| 21:49 |
|
gmcharlt |
dkg: server-side, you can use misc/migration_tools/bulkmarcimport.pl (to load in one go) |
| 21:50 |
|
gmcharlt |
or misc/stage_biblios_file.pl and misc/commit_biblios_file.pl if you're adding records to a database that already has some and you need to handle duplicate bibs |
| 21:50 |
|
dkg |
ok, i think that makes sense. |
| 21:50 |
|
gmcharlt |
if you're using Zebra as the indexing engine, you'll then need to follow up with a rebuild_zebra.pl -b -z |
| 21:51 |
|
dkg |
is there a way to reject staged records, or do they hang around forever until you commit them? |
| 21:51 |
|
gmcharlt |
the staged records stick around |
| 21:51 |
|
dkg |
(for instance, if you realize something went wrong after staging but before commit) |
| 21:51 |
|
gmcharlt |
a delete button will be added eventually in the web UI |
| 21:51 |
|
gmcharlt |
but you can also do it from using SQL: delete from import_batches where import_batch_id = N |
| 21:52 |
|
dkg |
ah, cool. that's great. and that'll have no nasty side effects down the line? |
| 21:52 |
|
gmcharlt |
no side effects to deleting a staged batch pre-commit |
| 21:52 |
|
dkg |
thank you. |
| 21:52 |
|
gmcharlt |
post-commit, main effect is that you can't run the tool to undo the import |
| 21:53 |
|
dkg |
I see. |
| 21:53 |
|
gmcharlt |
but if you're doing an initial import, that's OK - bulkmarcimport.pl -d deletes all of the bibs and items if you need to start from scratch |
| 21:54 |
|
dkg |
OK, you all were a huge help for me today. Thank you very much. |
| 21:54 |
|
gmcharlt |
you're welcome |
| 21:54 |
|
dkg |
if you're ever in NYC, i'll get you a beer (or other drink if you don't drink beer)! |
| 21:55 |
|
gmcharlt |
heh - thanks! |
| 22:37 |
|
hdl_laptop |
owen around ? |
| 22:48 |
|
hdl_laptop |
owen : when you read this, I was just wondering how I could get a function in Jquery used both on load and on change. |
| 23:18 |
|
atz |
hdl_laptop: $(document).ready(function() { ... your code here ...; }); |
| 23:18 |
|
atz |
then |
| 23:21 |
|
atz |
$(document).ready(function() { |
| 23:21 |
|
atz |
$('.ajax_buttons' ).css({visibility:"visible"}); |
| 23:21 |
|
atz |
$('body').click(function(event) { |
| 23:21 |
|
atz |
if ($(event.target).is('.ok')) { |
| 23:21 |
|
atz |
$.ajax({ |
| 23:21 |
|
atz |
"data": {ok: $(event.target).attr("title"), CGISESSID: readCookie('CGISESSID')}, |
| 23:21 |
|
atz |
"success": count_approve // success_approve |
| 23:21 |
|
atz |
}); |
| 23:21 |
|
atz |
return false; // cancel submit |
| 23:21 |
|
atz |
} |
| 23:21 |
|
atz |
} |
| 23:21 |
|
atz |
so that would set up an action when the page is ready (unhiding js-dependent buttons) |
| 23:22 |
|
atz |
and also establish what would happen when a click happens on a target that has class="ok" |
| 23:22 |
|
atz |
this example is simplified from koha-tmpl/intranet-tmpl/prog/en/modules/tags/review.tmpl |
| 23:23 |
|
atz |
hdl_laptop: does that answer your question? |
| 23:29 |
|
hdl_laptop |
partially. |
| 23:29 |
|
hdl_laptop |
atz: I would have liked to be able to add |
| 23:31 |
|
hdl_laptop |
$(frequency).change(function(){myfunction;} |
| 23:31 |
|
hdl_laptop |
$(frequency).load(function(){myfunction;} |
| 23:32 |
|
hdl_laptop |
add function to both change and load. |
| 23:32 |
|
hdl_laptop |
If there was. |
| 23:33 |
|
hdl_laptop |
seems that what you suggests suits your use. But mine is quite different. |
| 23:34 |
|
hdl_laptop |
I have 2 selection fields : frequency and number patterns. |
| 23:35 |
|
hdl_laptop |
And I would like to load values on load, if there are some. |
| 23:35 |
|
hdl_laptop |
But maybe it is just daydreaming. |
| 23:35 |
|
hdl_laptop |
(it is 1.30 here...) |
| 23:50 |
|
atz |
when you say "load" you mean when the page is first rendered |
| 23:52 |
|
atz |
in jQuery, $(document).ready means when the DOM has been built and is ready to be manipulated |
| 23:52 |
|
atz |
if you have id="frequency" on your input element, you could use: |
| 23:53 |
|
atz |
$("#frequency").... |
| 23:56 |
|
atz |
change docs: http://docs.jquery.com/Events/change |
| 01:41 |
|
owen |
Hi cnighs |
| 01:43 |
|
cnighs |
hi owen |
| 02:09 |
|
masonj_ |
heya chris ;) |
