*** siesel has joined #gnuenterprise *** Morphous has joined #gnuenterprise *** Amorphous has quit IRC *** wendall911 has quit IRC *** dcmwai has quit IRC *** _florin_ has joined #gnuenterprise *** holycow has joined #gnuenterprise *** dcmwai has joined #gnuenterprise hey dudes :) *** johannesV has joined #gnuenterprise *** kilo has joined #gnuenterprise *** reinhard has joined #gnuenterprise *** btami has joined #gnuenterprise reinhard: good morning hi i tried to run a report that returns non-ascii chars from a database and it failed due to conversions i can hack it using an encode() functions just like wxEncode but as there are several places that use wxEncode (or textEncode in btami's code) wouldnt it be practical to put a textEncode function somewhere in common? so that it could be used anywhere eg in TextUtils.py what do you think sorry was on phone kilo: the problem is we have different encodings wxEncode has to encode into the X font encoding other things (like tracebacks) must encode to your current terminal encoding arg... phone again :( i would think that raw xml reports should be encoded into utf-8 regardless of your locale seems reasonable but if i put non-ascii chars in the .grd file like in the table heading part gnue-reports dumps it correctly into the xml - that is it encodes it correctly so when dealing with headings it encodes ok but when dumping fields it fails it's ok to encode into utf-8 but then the code that encodes headings should also reutrn utf-8 it'd be logical to just encode to utf-8 immediately before writing kilo: if you have non-ascii characters in the grd you either have to define the encoding in the xml header or use utf-8 in either case you should get UnicodeType values from the parser yes in the xml header cp1250 is defined i only tried to suggest that all sections should be dealt with the same way then as it doesnt seems to be so *** dneighbo has joined #gnuenterprise reinhard: finally i'v hunted down my mysterious bug: --------------------------- Runtime Error occured: Traceback (most recent call last): File "", line 4, in _messageBox File "", line 1 SyntaxError: can't assign to operator (, line 1) -------------------------- it's coming from connections.conf if i have a section with "-" in hes name, like [gnue-sb] than in logic/adapter/python.py __namespace will contain it, and ah gotcha exec '%s'... will fail i think i can fix that (later) ok, thx *** derek has quit IRC reinhard: any suggestions? *** dcmwai has quit IRC kilo: i have no idea how the output file is built now but IMHO, the most logical way would be (as always) to process everything in Unicode internally and just in the process of writing the output file make the conversion to utf-8 i agree * reinhard has not even looked at gnue-reports doing a grep on "unicode" and "encode" in the reports source it very much looks like there hasn't been done much in that respect the problem is a bit complicated than with forms, koz in first step reports process the non layout tags+data and produce a temp xml file then filters process this xml (regarding layout tags) yeah and do whatewer they want i think this temp xml should be in utf-8 in any case which it is probably not now me too yes my understanding was always that i could also use the xml as "final" output then in a separate step generate a html *and* a pdf and whatever (for example) or use the xml for archiving the result of a report and when i want to get back to it, render it as pdf etc. however i think i have far too little knowledge about reports now to fix that :( ok and for many companies here it's the first day of a new fiscal year so my telephone is my hell today :( fool's day it is that also :) *** SachaS has joined #gnuenterprise hi SachaS if i have a section with "-" in hes name, like [gnue-sb] than in logic/adapter/python.py btami: this should be fixed now; good morning reinhard :) hi johannesVC hi johannesV hi SachaS any chance of having list properties sooner, than in 0.2.0? I would have some code to test it. hi SachaS hi kilo. SachaS: sure SachaS: if you implement them j/k reinhard: good joke :) SachaS: i'm very sure that bound procedures and calculated fields are much more important and urgent more urgent than multithreading... 8-) kilo: :) i remember programming some little apps multithreaded, in C... now THAT was programming... reinhard: its just that in my next sample, i would need a list property ;) but bound procedures and calculated fields are also important. SachaS: you can emulate any list property with a single find() will a list property be a python list? and provide the methods the python list provides? yes well no and yes ok, so one has to know the python list ... it will not be a python list, but it will behave as it was on one all you have to know is that you can do mylist = find (....) for object in mylist: also you can do count = len (mylist) ok. johannesV: thx, it's OK now *** SachaS_ has joined #gnuenterprise *** SachaS has quit IRC *** btami has quit IRC *** chillywilly has quit IRC *** chillywilly has joined #gnuenterprise reinhard: a question: items and suppliers. where do you have the link? suppliers list items it can supply or item lists which supplier provide the item. or both? :) i realise, that if there will be gnue packages, there will be package, business, organisatorial questions and discussions. *** holycow has quit IRC *** lekma has joined #gnuenterprise *** lxf has joined #gnuenterprise *** kilo has quit IRC *** lekma has quit IRC *** lekma has joined #gnuenterprise hello iemand?? ja :) question: is there any mean to define sequences properties in appserver ?? sequences? do you mean lists? currently only 2 types of properties are supported: no, postgresql sequence, mysql autoincrement... 1) normal property (of type number, string, date) 2) reference property (references another business object) lekma, dont understand hi lekma reinhard: hi lekma: ask reinhard ;) lekma: there are no sequence type fields in appserver what would be the reason to use one? SachaS: i mean in the example you give in the appserver howto lekma? :) where is that? the itemId <-- i would like it to be an autoincrement nuber in the range xxxxxx to yyyyyy is this possible?? ah ok. appserver creates this ID for you. ask reinhard, what algorithm he uses to create that id. i don't want to use gnue_id my users won't like it too long and too abstract you can add your own "id" property and define your algorithm which creates you that ID in the range of xxxx to yyyyy. * lekma is puzzled the gnue_id is used behind the scene, sort of the business object references lekma: if i understand you right, you want to use your own id for, for example customer id's ? yes i think in the future, you will not have to deal with the gnue_id at all. will be done behind the scene by appserver. so you can, without any problems, add an id for your customer business object. and if you can search for your customer business object by its id like this: customerList = find('masterdata_customer', {'masterdata_customerID': 12345}, []) where you are searching for the customer with YOUR id 12345. find will return a list, though. if there is only one item in the list you can access it with: myCustomer = customerList[0] *** jamest has joined #gnuenterprise lekma: does this make sense? sorry phone :( (4th call since i last talked) yes, but what i mean is: is there any way to add a property type 'autoincrement' with which one could autogenerate values when creating new (let's say) customer ie the customerId property would be of type 'autoincrement' * lekma thinks he's confused and he's confusing SachaS lekma. yes. you can write your own procedure for the "customer" business class. so you do: customer = new("masterdata_customer") and then you do: customer.customerId = customer.createMyOwnID() back but you have to write the "createMyOwnID()" procedure. you would have to make sure, that you create UNIQUE ids for example. lekma: I think I understand what you are after lekma: thats how i understand it. you have for example part numbers phone :( *** dsmith has joined #gnuenterprise *** Morphous has quit IRC back lekma: you still here? back ok the way i understand your question you want an item number (like a part number or order number) that is *not* a simple rowid but is - yes, well - the item number and you want it to be assigned automatically yes highest previous number + 1 appserver does not support a concept of autoincrement numbers but you could do one of the following: a) write a procedure in appserver "createItemNumber" and call it from a form trigger in the event of your choice b) after the upcoming release, we will as next feature implement bound procedures, which means that appserver would call "createItemNumber" automatically without needing a form trigger c) of course you could do all the logic in forms, but then you don't need appserver the advantage of the createItemNumber (compared to a serial) is that you can do anything you want in this procedure, e.g. * have different number ranges for different product groups * have some logic in the number, e.g. the last digit is a parity code * have not only numbers but anything you like, like 123-ABC-456 did i understand your question correctly? and is my answer understandable? yes definitely for now i'll try solution a) lekma: excellent jamest: you here? *** dimas has joined #gnuenterprise am now hi dimas jamest: started to use our hotline prject in production today guess what i have some questions on forms :) i have a master detail form hello all hi jamest where the primary key of the master record gets assigned on commit (and not earlier) now when i add a master and already add some details before i commit the master all details have NULL as foreign key value is there some way around this? where is it getting the pk? as a default in the db? so to speak the db is appserver :) but you could think of it like a serial field to be precise it would get the pk upon post() on postgresql we have to use the getSequence to get the value and assign it in the form eeek because currently we don't have a way to get that commited data back in datasources yeah, it's fugly that will become a big problem with appserver phone again... ntryschool_id.autofillBySequence('seq_school_id') ]]> well, the db api doesn't return altered values IIRC nothing would stop us from committing the data, then immediately querying it back based upon the inserted values we know we set i think then forms|appserver|whatever would need tweaked to check master PK fields after they commit and make the appropriate changes in the detail blocks exactly so far I would suggest we hadn't hit anything that made the performance hit worth the implementation however, now appserver is going full steam I guess we have :) it's not the performance hit it's the fact that appserver uses autogenerated pk's in *every* table lol so *every* master detail form would have to contain that code ok, so this is something we need to change i see 2 paths 1. between commit of master and commit of clients, update fk for all clients 2. before insert of first client, post() the master record (which should already make it possible to read the new pk value) 2. would mean that while you are inserting details, you have an uncommitted INSERT for the master pending in the DB not sure if that would be a problem and not sure which would be the better way anyway *** sjc has joined #gnuenterprise hi sjc jamest: hi right now i think forms commits the details per master but i can't look atm maybe it just iterates thru the blocks no, per master forms only commits "master" datasources details are handled by dataobjects #2 is how it works now, btw but we done requery the values to see what changed s/done/dont no that's not realistic to do unfortunately why do you say that? um, very few drivers could support it (and none via the dbsig2) i figured it would have to work similar to updates but we'd not include the pk fields of the master in the where that would quickly fail if there are any backend triggers filling in missing info in tables this wouldn't cover trigger yes bbl must make dinner for kids forms has to be able to grab a pk value from the database backend to be consistent perhaps this could be simplified as an option to the definition instead of a trigger or the appserver dbdriver could handle it internally differently than the dbsig2 drivers, if appserver returns the pk id automatically *** jcater has quit IRC the appserver dbdriver could return the new pk on postChanges() just tell me how it should be returned *** _florin_ has quit IRC *** _florin_ has joined #gnuenterprise <_florin_> hello again :-) hello <_florin_> hey jamest! do you have some time to answer me tons of questions about events? :*) _florin_: kinda iffy, you can ask and i'll answer as I can (real work :) *** Amorphous has joined #gnuenterprise <_florin_> i managed to send some (from object to object) and to receive them. but i don't know i'm using them as i should... <_florin_> ok. same old problem: ui object <-> gf object <_florin_> ui object has EventHandler from its _createWidget()-function (event.eventHandler) -- just like othter ui-objects do. <_florin_> gf object is _not_ EventAware. instead it calls self._form._instance.eventController.registerEventListeners <_florin_> to register its own onItemSelected function to an event: 'treeItemSelected' (which is sent by ui object from time to time...) <_florin_> the good part: it works. <_florin_> the bad part: i simply invented a new event: 'treeItemSelected' and didn't re-use already existent ones (like you actually had told me to) <_florin_> another bad part: i simply used the best eventHandler i found (the one of my form's _instance) instead of creating/registering a new one. <_florin_> is that how it's supposed to work? *** jcater has joined #gnuenterprise i'll have to look at the code but i imagine it's ok for now we don't respect the ._foo == private quite like we should <_florin_> correct me if i'm wrong: for every... "event space" there's exactly one EventController which is responsible for events, right? yes <_florin_> and messages between two different layers (like gf widgets <-> ui widgets) are passed using only events, right? yes <_florin_> the sender needs an eventHandler, the receiver ties an event name to an own member function using an (_the_) EventController, right? <_florin_> (what's an eventHandler? is it another name for eventController? is it an Event?) where do you see eventHandler? is that in the UI* stuff? if so, I *think* that's just a handle to the eventController.dispatchEvent method <_florin_> uidrivers/wx/widgets/entryp.py:109 <_florin_> oh, ok :-) <_florin_> so it's actually just the ... ahm... send()-function :-) that's actually something that needs cleaning up i.e., how the eventController gets passed around in the UI drivers iirc, we make it global or something now <_florin_> well, i'm just looking for a way to _send_ events, and it seems like that eventHandler (or evtCtrl.dispatchEvent) gets the job done. if there's no better options now, i'll use that one. yes, that is main option <_florin_> ok, thanks a lot! seems like i'm beginning to figure out the gnue event system:-) <_florin_> bbl, gotta run *** dcmwai has joined #gnuenterprise hi dcmwai Hello jamest derek say that you want to do something with aria-erp.org right? yes i want to try something jamest, what it is about? jamest, Sure... Everything :) i have a person that wants to run quickbooks remove over dialup and. or pass backup files back and forth from home to main office ... they don't know quickbooks and that is a damn near impossible request haha so I'm trying to find web based payroll and GL for them jamest, so you think of arias should be aria reading old mails and stuff it seems there was interest in gnue/arias collaberation or merge Yes there were, but now I'm the only one left :( sorry, i saw it written both ways everyone else quit? Not to say quit.. but that don't know why. Once they are invoived with Aria, Life seem to be busy ... i know how that goes :( That happen to me, jfletch and every developer... * dcmwai is thinking that aria is Not a good name and might not be lucky.... jamest, You can do anything with the Cvs server.. em I say it wrongly... You can have everything from the cvs server, and If you have any problem you can always came to me. don't goto the Smarty branch yet, it is highly unstable... :) well we seem to have a copy of aria in gnue-sb that is postgresql-ized jamest, well that is Years old :) jamest, it is? i'm running it at home w/ postgresql :) jamest, I think I've try to compare the different and it do really kill me jamest, can you do something for me? what? jamest, compare your version of aria to the old version of mysqldb that you build the postgresql. Then Send me the different. i don't understand cause Our CVS version is a lot different in UI then DB code. i really don't know anything about the porting or state of ours i just got it to run the other night and left it at that :) but I can send diffs if that is what you want (not sure how far the gnue-sb people got) jamest, Nevermind. asl long as I can get the Change in DB code then we can tell the different and know the problem so aria is still active project then? *** dimas has quit IRC and what are your long term goals for it? jamest, It is ... and people is actually Bugging me for support :( are you tied to the php front end? jamest, Be frank. I'm thinking of rewrite it with Smarty so that it can be easily debug. is the logic/ui code all mixed together? jamest, Unfornature at the time being. Yes it is.. and I DO REALLY HATE it. after the rewrite, the UI and logic will be split at the same time there will be a debug and test function for all pages. (so that we can really test if there is any problem some where) Without getting in all page and key in data. But that was the long long term goal. i'm not certain what will happen on my end all up to the client of course :) i think in our svn tree someone was converting the aria screens to gnue forms so it would run local vs web jamest, haha that was cool jamest, any news on the new data schema? i'm sorry what? dcmwai, the New Database Schema that will be use for gnue-sb I think it was from dtm i think the aria database structure in gnue-sb is from gnue schema files instead of your sql statements jamest, I know that. however dtm seem to be running a check on all opensource ERP Database Schema and try to mix up one that is 80% usable out of the box ah, i did not know that has he even been in here in a while? jamest, well he have been here all the while... until someday... jamest, I think he is still in freenode, Just in other channel (#ldap mainly) ah jamest, I can't take anymore... I've caugth a cold and Now I must sleep... ok nite jamest, Good Night and If you have anyproblem Just let me know, I will put you in the 1st place. :) for all i know jamest, You knwo my email right? Email me ... they bought quickbooks and will be asking me to make it run remote :) um, let me check jamest, Througth VPN :) haha i have the email on the contact page as aris * dcmwai think that is what I will do if the customer insist :) the sourceforge one i guess it isn't multiuser so that's impossible haha :)