Search Results

Search found 1261 results on 51 pages for 'trivial'.

Page 44/51 | < Previous Page | 40 41 42 43 44 45 46 47 48 49 50 51  | Next Page >

  • With sqlalchemy how to dynamically bind to database engine on a per-request basis

    - by Peter Hansen
    I have a Pylons-based web application which connects via Sqlalchemy (v0.5) to a Postgres database. For security, rather than follow the typical pattern of simple web apps (as seen in just about all tutorials), I'm not using a generic Postgres user (e.g. "webapp") but am requiring that users enter their own Postgres userid and password, and am using that to establish the connection. That means we get the full benefit of Postgres security. Complicating things still further, there are two separate databases to connect to. Although they're currently in the same Postgres cluster, they need to be able to move to separate hosts at a later date. We're using sqlalchemy's declarative package, though I can't see that this has any bearing on the matter. Most examples of sqlalchemy show trivial approaches such as setting up the Metadata once, at application startup, with a generic database userid and password, which is used through the web application. This is usually done with Metadata.bind = create_engine(), sometimes even at module-level in the database model files. My question is, how can we defer establishing the connections until the user has logged in, and then (of course) re-use those connections, or re-establish them using the same credentials, for each subsequent request. We have this working -- we think -- but I'm not only not certain of the safety of it, I also think it looks incredibly heavy-weight for the situation. Inside the __call__ method of the BaseController we retrieve the userid and password from the web session, call sqlalchemy create_engine() once for each database, then call a routine which calls Session.bind_mapper() repeatedly, once for each table that may be referenced on each of those connections, even though any given request usually references only one or two tables. It looks something like this: # in lib/base.py on the BaseController class def __call__(self, environ, start_response): # note: web session contains {'username': XXX, 'password': YYY} url1 = 'postgres://%(username)s:%(password)s@server1/finance' % session url2 = 'postgres://%(username)s:%(password)s@server2/staff' % session finance = create_engine(url1) staff = create_engine(url2) db_configure(staff, finance) # see below ... etc # in another file Session = scoped_session(sessionmaker()) def db_configure(staff, finance): s = Session() from db.finance import Employee, Customer, Invoice for c in [ Employee, Customer, Invoice, ]: s.bind_mapper(c, finance) from db.staff import Project, Hour for c in [ Project, Hour, ]: s.bind_mapper(c, staff) s.close() # prevents leaking connections between sessions? So the create_engine() calls occur on every request... I can see that being needed, and the Connection Pool probably caches them and does things sensibly. But calling Session.bind_mapper() once for each table, on every request? Seems like there has to be a better way. Obviously, since a desire for strong security underlies all this, we don't want any chance that a connection established for a high-security user will inadvertently be used in a later request by a low-security user.

    Read the article

  • MBR Booting from DOS

    - by eflukx
    For a project I would like to invoke the MBR on the first harddisk directly from DOS. I've written a small assembler program that loads the MBR in memory at 0:7c00h an does a far jump to it. I've put my util on a bootable floppy. The disk (HD0, 0x80) i'm trying to boot has a TrueCrypt boot loader on it. It shows up the TrueCrypt screen, but after typing in the password it crashes the system. When I run my little utlility (w00t.com) on a normal WinXP machine it seams to crash immedealty. Apparently I'm forgetting some crucial stuff the BIOS normally does, my guess is it's something trivial. Can someone with better bare-metal DOS and BIOS experience help me out? Heres my code: .MODEL tiny .386 _TEXT SEGMENT USE16 INCLUDE BootDefs.i ORG 100h start: ; http://vxheavens.com/lib/vbw05.html ; Before DOS has booted the BIOS stores the amount of usable lower memory ; in a word located at 0:413h in memory. We going to erase this value because ; we have booted dos before loading the bootsector, and dos is fat (and ugly). ; fake free memory ;push ds ;push 0 ;pop ds ;mov ax, TC_BOOT_LOADER_SEGMENT / 1024 * 16 + TC_BOOT_MEMORY_REQUIRED ;mov word ptr ds:[413h], ax ;ax = memory in K ;pop ds ;lea si, memory_patched_msg ;call print ;mov ax, cs mov ax, 0 mov es, ax ; read first sector to es:7c00h (== cs:7c00) mov dl, 80h mov cl, 1 mov al, 1 mov bx, 7c00h ;load sector to es:bx call read_sectors lea si, mbr_loaded_msg call print lea si, jmp_to_mbr_msg call print ;Set BIOS default values in environment cli mov dl, 80h ;(drive C) xor ax, ax mov ds, ax mov es, ax mov ss, ax mov sp, 0ffffh sti push es push 7c00h retf ;Jump to MBR code at 0:7c00h ; Print string print: xor bx, bx mov ah, 0eh cld @@: lodsb test al, al jz print_end int 10h jmp @B print_end: ret ; Read sectors of the first cylinder read_sectors: mov ch, 0 ; Cylinder mov dh, 0 ; Head ; DL = drive number passed from BIOS mov ah, 2 int 13h jnc read_ok lea si, disk_error_msg call print read_ok: ret memory_patched_msg db 'Memory patched', 13, 10, 7, 0 mbr_loaded_msg db 'MBR loaded', 13, 10, 7, 0 jmp_to_mbr_msg db 'Jumping to MBR code', 13, 10, 7, 0 disk_error_msg db 'Disk error', 13, 10, 7, 0 _TEXT ENDS END start

    Read the article

  • XQuery method question, trying to sum values read from xml

    - by Buck
    I'm pretty new to XQuery and I'm trying to write an example function that I can't get to work. I want to read an xml file, parse out the "time" values, sum them as they're read and return the sum. This is trivial and I'm looking to build more functionality into it but I'd like to get this working first. Also, I know there's a "sum" directive in XQuery that would do just this but I want to add more to it so the built-in sum is insufficient for my needs. Here's my funtion: bool example(Zorba* aZorba) { XQuery_t lQuery = aZorba-compileQuery( "for $i in fn:doc('/tmp/products.xml')//time" "let $sum := xs:integer($i)" " return $sum" ); DynamicContext* lCtx = lQuery-getDynamicContext(); lCtx-setContextItemAsDocument("temp_measurements.xml", lDocStream); try { std::cout << lQuery << std::endl; } catch (DynamicException& e) { std::cerr << e.getDescription() << std::endl; return false; } catch (StaticException& f){ std::cerr << f.getDescription() << f.getErrorCodeAsString(f.getErrorCode()) << std::endl; return false; } } It's called with an appropriate main(). If I comment out the line that starts "let $sum..." then this works in that it returns the time values as a series of integers like this: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3.... Input file looks like this: <?xml version="1.0" encoding="UTF-8"? <temps <temp <time0</time <lat0</lat <long0</long <value0</value </temp <temp <time1</time <lat0</lat <long1</long <value0</value </temp ...

    Read the article

  • Most efficient method to query a Young Tableau

    - by Matthieu M.
    A Young Tableau is a 2D matrix A of dimensions M*N such that: i,j in [0,M)x[0,N): for each p in (i,M), A[i,j] <= A[p,j] for each q in (j,N), A[i,j] <= A[i,q] That is, it's sorted row-wise and column-wise. Since it may contain less than M*N numbers, the bottom-right values might be represented either as missing or using (in algorithm theory) infinity to denote their absence. Now the (elementary) question: how to check if a given number is contained in the Young Tableau ? Well, it's trivial to produce an algorithm in O(M*N) time of course, but what's interesting is that it is very easy to provide an algorithm in O(M+N) time: Bottom-Left search: Let x be the number we look for, initialize i,j as M-1, 0 (bottom left corner) If x == A[i,j], return true If x < A[i,j], then if i is 0, return false else decrement i and go to 2. Else, if j is N-1, return false else increment j This algorithm does not make more than M+N moves. The correctness is left as an exercise. It is possible though to obtain a better asymptotic runtime. Pivot Search: Let x be the number we look for, initialize i,j as floor(M/2), floor(N/2) If x == A[i,j], return true If x < A[i,j], search (recursively) in A[0:i-1, 0:j-1], A[i:M-1, 0:j-1] and A[0:i-1, j:N-1] Else search (recursively) in A[i+1:M-1, 0:j], A[i+1:M-1, j+1:N-1] and A[0:i, j+1:N-1] This algorithm proceed by discarding one of the 4 quadrants at each iteration and running recursively on the 3 left (divide and conquer), the master theorem yields a complexity of O((N+M)**(log 3 / log 4)) which is better asymptotically. However, this is only a big-O estimation... So, here are the questions: Do you know (or can think of) an algorithm with a better asymptotical runtime ? Like introsort prove, sometimes it's worth switching algorithms depending on the input size or input topology... do you think it would be possible here ? For 2., I am notably thinking that for small size inputs, the bottom-left search should be faster because of its O(1) space requirement / lower constant term.

    Read the article

  • realtime diagnostics

    - by Ion Todirel
    I have an application which has a loop, part of a "Scheduler", which runs at all time and is the heart of the application. Pretty much like a game loop, just that my application is a WPF application and it's not a game. Naturally the application does logging at many points, but the Scheduler does some sensitive monitoring, and sometimes it's impossible just from the logs to tell what may have gotten wrong (and by wrong I don't mean exceptions) or the current status. Because Scheduler's inner loop runs at short intervals, you can't do file I/O-based logging (or using the Event Viewer) in there. First, you need to watch it in real-time, and secondly the log file would grow in size very fast. So I was thinking of ways to show this data to the user in the realtime, some things I considered: Display the data in realtime in the UI Use AllocConsole/WriteConsole to display this information in a console Use a different console application which would display this information, communicate between the Scheduler and the console app using pipes or other IPC techniques Use Windows' Performance Monitor and somehow feed it with this information ETW Displaying in the UI would have its issues. First it doesn't integrate with the UI I had in mind for my application, and I don't want to complicate the UI just for this. This diagnostics would only happen rarely. Secondly, there is going to be some non-trivial data protection, as the Scheduler has it's own thread. A separate console window would work probably, but I'm still worried if it's not too much threshold. Allocating my own console, as this is a windows app, would probably be better than a different console application (3), as I don't need to worry about IPC communication, and non-blocking communication. However a user could close the console I allocated, and it would be problematic in that case. With a separate process you don't have to worry about it. Assuming there is an API for Performance Monitor, it wouldn't be integrated too well with my app or apparent to the users. Using ETW also doesn't solve anything, just a random idea, I still need to display this information somehow. What others think, would there be other ways I missed?

    Read the article

  • Yii - Custom GridView with Multiple Tables

    - by savinger
    So, I've extended GridView to include an Advanced Search feature tailored to the needs of my organization. Filter - lets you show/hide columns in the table, and you can also reorder columns by dragging the little drag icon to the left of each item. Sort - Allows for the selection of multiple columns, specify Ascending or Descending. Search - Select your column and insert search parameters. Operators tailored to data type of selected column. Version 1 works, albeit slowly. Basically, I had my hands in the inner workings of CGridView, where I snatch the results from the DataProvider and do the searching and sorting in PHP before rendering the table contents. Now writing Version 2, where I aim to focus on clever CDbCriteria creation, allowing MySQL to do the heavy lifting so it will run quicker. The implementation is trivial when dealing with a single database table. The difficulty arises when I'm dealing with 2 or more tables... For example, if the user intends to search on a field that is a STAT relation, I need that relation to be present in my query. Here's the question. How do I assure that Yii includes all with relations in my query so that I include comparisons? I've included all my relations with my criteria in the model's search function and I've tried CDbCriteria's together ... public function search() { $criteria=new CDbCriteria; $criteria->compare('id', $this->id); $criteria->compare( ... ... $criteria->with = array('relation1','relation2','relation3'); $criteria->together = true; return new CActiveDataProvider( get_class($this), array( 'criteria'=>$criteria, 'pagination' => array('pageSize' => 50) ));} But I still get errors like this... CDbCommand failed to execute the SQL statement: SQLSTATE[42S22]: Column not found: 1054 Unknown column 't.relation3' in 'where clause'. The SQL statement executed was: SELECT COUNT(DISTINCT `t`.`id`) FROM `table` `t` LEFT OUTER JOIN `relation_table` `relation0` ON (`t`.`id`=`relation0`.`id`) LEFT OUTER JOIN `relation_table` `relation1` ON (`t`.`id`=`relation1`.`id`) WHERE (`t`.`relation3` < 1234567890) Where relation0 and relation1 are BELONGS_TO relations, but any STAT relations are missing. Furthermore, why is the query a SELECT COUNT(DISTINCT 't'.'id') ?

    Read the article

  • Wondering why DisplayName attribute is ignored in LabelFor on an overridden property

    - by Lasse Krantz
    Hi, today I got confused when doing a couple of <%=Html.LabelFor(m=>m.MyProperty)%> in ASP.NET MVC 2 and using the [DisplayName("Show this instead of MyProperty")] attribute from System.ComponentModel. As it turned out, when I put the attribute on an overridden property, LabelFor didn't seem to notice it. However, the [Required] attribute works fine on the overridden property, and the generated errormessage actually uses the DisplayNameAttribute. This is some trivial examplecode, the more realistic scenario is that I have a databasemodel separate from the viewmodel, but for convenience, I'd like to inherit from the databasemodel, add View-only properties and decorating the viewmodel with the attributes for the UI. public class POCOWithoutDataAnnotations { public virtual string PleaseOverrideMe { get; set; } } public class EditModel : POCOWithoutDataAnnotations { [Required] [DisplayName("This should be as label for please override me!")] public override string PleaseOverrideMe { get { return base.PleaseOverrideMe; } set { base.PleaseOverrideMe = value; } } [Required] [DisplayName("This property exists only in EditModel")] public string NonOverriddenProp { get; set; } } The strongly typed ViewPage<EditModel> contains: <div class="editor-label"> <%= Html.LabelFor(model => model.PleaseOverrideMe) %> </div> <div class="editor-field"> <%= Html.TextBoxFor(model => model.PleaseOverrideMe) %> <%= Html.ValidationMessageFor(model => model.PleaseOverrideMe) %> </div> <div class="editor-label"> <%= Html.LabelFor(model => model.NonOverriddenProp) %> </div> <div class="editor-field"> <%= Html.TextBoxFor(model => model.NonOverriddenProp) %> <%= Html.ValidationMessageFor(model => model.NonOverriddenProp) %> </div> The labels are then displayed as "PleaseOverrideMe" (not using the DisplayNameAttribute) and "This property exists only in EditModel" (using the DisplayNameAttribute) when viewing the page. If I post with empty values, triggering the validation with this ActionMethod: [HttpPost] public ActionResult Edit(EditModel model) { if (!ModelState.IsValid) return View(model); return View("Thanks"); } the <%= Html.ValidationMessageFor(model => model.PleaseOverrideMe) %> actually uses [DisplayName("This should be as label for please override me!")] attribute, and produces the default errortext "The This should be as label for please override me! field is required." Would some friendly soul shed some light on this?

    Read the article

  • XML Return from an Oracle Stored Procedure

    - by Tequila Jinx
    Unfortunately most of my DB experience has been with MSSQL which tends to hold your hand a lot more than Oracle. What I'm trying to do is fairly trivial in tSQL, however, pl/sql is giving me a headache. I have the following procedure: CREATE OR REPLACE PROCEDURE USPX_GetUserbyID (USERID USERS.USERID%TYPE, USERRECORD OUT XMLTYPE) AS BEGIN SELECT XMLELEMENT("user" , XMLATTRIBUTES(u.USERID AS "userid", u.companyid as "companyid", u.usertype as "usertype", u.status as "status", u.personid as "personid") , XMLFOREST( p.FIRSTNAME AS "firstname" , p.LASTNAME AS "lastname" , p.EMAIL AS "email" , p.PHONE AS "phone" , p.PHONEEXTENSION AS "extension") , XMLELEMENT("roles", (SELECT XMLAGG(XMLELEMENT("role", r.ROLETYPE)) FROM USER_ROLES r WHERE r.USERID = USERID AND r.ISACTIVE = 1 ) ) , XMLELEMENT("watches", (SELECT XMLAGG( XMLELEMENT("watch", XMLATTRIBUTES(w.WATCHID AS "id", w.TICKETID AS "ticket") ) ) FROM USER_WATCHES w WHERE w.USERID = USERID AND w.ISACTIVE = 1 ) ) ) AS "RESULT" INTO USERRECORD FROM USERS u LEFT JOIN PEOPLE p ON p.PERSONID = u.PERSONID WHERE u.USERID = USERID; END USPX_GetUserbyID; When executed, it should return an XML document with the following structure: <user userid="" companyid="" usertype="" status="" personid=""> <firstname /> <lastname /> <email /> <phone /> <extension /> <roles> <role /> </roles> <watches> <watch id="" ticket="" /> </watches> </user> When I execute the query itself, replacing the USERID parameter with a string and removing the "into" clause, the query runs fine and returns the expected structure. However, when the procedure attempts to execute the query, passing the results of the XMLELEMENT function into the USERRECORD output parameter, I get the following exception: Error report: ORA-01422: exact fetch returns more than requested number of rows ORA-06512: at "USPX_GETUSERBYID", line 4 ORA-06512: at line 3 01422. 00000 - "exact fetch returns more than requested number of rows" *Cause: The number specified in exact fetch is less than the rows returned. *Action: Rewrite the query or change number of rows requested I'm baffled trying to nail this down, and unfortunately my google-fu hasn't helped. I've found plenty of Oracle SQL|XML examples, but none that deal with XML returns from a procedure. Note: I know that an alternate method of retrieving XML using DBMS methods exists, however, it's my understanding that that functionality is deprecated in favor of SQL|XML.

    Read the article

  • Optimizing spacing of mesh containing a given set of points

    - by Feynman
    I tried to summarize the this as best as possible in the title. I am writing an initial value problem solver in the most general way possible. I start with an arbitrary number of initial values at arbitrary locations (inside a boundary.) The first part of my program creates a mesh/grid (I am not sure which is the correct nuance), with N points total, that contains all the initial values. My goal is to optimize the mesh such that the spacing is as uniform as possible. My solver seems to work half decently (it needs some more obscure debugging that is not relevant here.) I am starting with one dimension. I intend to generalize the algorithm to an arbitrary number of dimensions once I get it working consistently. I am writing my code in fortran, but feel free to reply with pseudocode or the language of your choice. Allow me to elaborate with an example: Say I am working on a closed interval [1,10] xmin=1 xmax=10 Say I have 3 initial points: xmin, 5 and xmax num_ivc=3 known(num_ivc)=[xmin,5,xmax] //my arrays start at 1. Assume "known" starts sorted I store my mesh/grid points in an array called coord. Say I want 10 points total in my mesh/grid. N=10 coord(10) Remember, all this is arbitrary--except the variable names of course. The algorithm should set coord to {1,2,3,4,5,6,7,8,9,10} Now for a less trivial example: num_ivc=3 known(num_ivc)=[xmin,5.5,xmax or just num_ivc=1 known(num_ivc)=[5.5] Now, would you have 5 evenly spaced points on the interval [1, 5.5] and 5 evenly spaced points on the interval (5.5, 10]? But there is more space between 1 and 5.5 than between 5.5 and 10. So would you have 6 points on [1, 5.5] followed by 4 on (5.5 to 10]. The key is to minimize the difference in spacing. I have been working on this for 2 days straight and I can assure you it is a lot trickier than it sounds. I have written code that only works if N is large only works if N is small only works if it the known points are close together only works if it the known points are far apart only works if at least one of the known points is near a boundary only works if none of the known points are near a boundary So as you can see, I have coded the gamut of almost-solutions. I cannot figure out a way to get it to perform equally well in all possible scenarios (that is, create the optimum spacing.)

    Read the article

  • jquery ajax post callback - manipulation stops after the "third" call

    - by shanyu
    EDIT: The problem is not related to Boxy, I've run into the same issue when I've used JQuery 's load method. EDIT 2: When I take out link.remove() from inside the ajax callback and place it before ajax load, the problem is no more. Are there restrictions for manipulating elements inside an ajax callback function. I am using JQuery with Boxy plugin. When the 'Flag' link on the page is clicked, a Boxy modal pops-up and loads a form via ajax. When the user submits the form, the link (<a> tag) is removed and a new one is created from the ajax response. This mechanism works for, well, 3 times! After the 3rd, the callback function just does not remove/replace/append (tested several variations of manipulation) the element. The only hint I have is that after the 3rd call, the parent of the link becomes non-selectable. However I can't make anything of this. Sorry if this is a very trivial issue, I have no experience in client-side programming. The relevant html is below: <div class="flag-link"> <img class="flag-img" style="width: 16px; visibility: hidden;" src="/static/images/flag.png" alt=""/> <a class="unflagged" href="/i/flag/showform/9/1/?next=/users/1/ozgurisil">Flag</a> </div> Here is the relevant js code: $(document).ready(function() { $('div.flag-link a.unflagged').live('click', function(e){ doFlag(e); return false; }); ... }); function doFlag(e) { var link = $(e.target); var url = link.attr('href'); Boxy.load(url, {title:'Inappropriate Content', unloadOnHide:true, cache:false, behaviours: function(r) { $("#flag-form").live("submit", function(){ var post_url = $("#flag-form").attr('action'); boxy = Boxy.get(this); boxy.hideAndUnload(); $.post(post_url, $("#flag-form").serialize(), function(data){ par = link.parent(); par.append(data); alert (par.attr('class')); //BECOMES UNDEFINED AT THE 3RD CALL!! par.children('img.flag-img').css('visibility', 'visible'); link.remove(); }); return false; }); }}); }

    Read the article

  • Write Scheme data structures so they can be eval-d back in, or alternative

    - by Jesse Millikan
    I'm writing an application (A juggling pattern animator) in PLT Scheme that accepts Scheme expressions as values for some fields. I'm attempting to write a small text editor that will let me "explode" expressions into expressions that can still be eval'd but contain the data as literals for manual tweaking. For example, (4hss->sexp "747") is a function call that generates a legitimate pattern. If I eval and print that, it becomes (((7 3) - - -) (- - (4 2) -) (- (7 2) - -) (- - - (7 1)) ((4 0) - - -) (- - (7 0) -) (- (7 2) - -) (- - - (4 3)) ((7 3) - - -) (- - (7 0) -) (- (4 1) - -) (- - - (7 1))) which can be "read" as a string, but will not "eval" the same as the function. For this statement, of course, what I need would be as simple as (quote (((7 3... but other examples are non-trivial. This one, for example, contains structs which print as vectors: pair-of-jugglers ; --> (#(struct:hand #(struct:position -0.35 2.0 1.0) #(struct:position -0.6 2.05 1.1) 1.832595714594046) #(struct:hand #(struct:position 0.35 2.0 1.0) #(struct:position 0.6 2.0500000000000003 1.1) 1.308996938995747) #(struct:hand #(struct:position 0.35 -2.0 1.0) #(struct:position 0.6 -2.05 1.1) -1.3089969389957472) #(struct:hand #(struct:position -0.35 -2.0 1.0) #(struct:position -0.6 -2.05 1.1) -1.8325957145940461)) I've thought of at least three possible solutions, none of which I like very much. Solution A is to write a recursive eval-able output function myself for a reasonably large subset of the values that I might be using. There (probably...) won't be any circular references by the nature of the data structures used, so that wouldn't be such a long job. The output would end up looking like `(((3 0) (... ; ex 1 `(,(make-hand (make-position ... ; ex 2 Or even worse if I could't figure out how to do it properly with quasiquoting. Solution B would be to write out everything as (read (open-input-string "(big-long-s-expression)")) which, technically, solves the problem I'm bringing up but is... ugly. Solution C might be a different approach of giving up eval and using only read for parsing input, or an uglier approach where the s-expression is used as directly data if eval fails, but those both seem unpleasant compared to using scheme values directly. Undiscovered Solution D would be a PLT Scheme option, function or library I haven't located that would match Solution A. Help me out before I start having bad recursion dreams again.

    Read the article

  • How best to store Subversion version information in EAR's?

    - by Rene
    When receiving a bug report or an it-doesnt-work message one of my initials questions is always what version? With a different builds being at many stages of testing, planning and deploying this is often a non-trivial question. I the case of releasing Java JAR (ear, jar, rar, war) files I would like to be able to look in/at the JAR and switch to the same branch, version or tag that was the source of the released JAR. How can I best adjust the ant build process so that the version information in the svn checkout remains in the created build? I was thinking along the lines of: adding a VERSION file, but with what content? storing information in the META-INF file, but under what property with which content? copying sources into the result archive added svn:properties to all sources with keywords in places the compiler leaves them be I ended up using the svnversion approach (the accepted anwser), because it scans the entire subtree as opposed to svn info which just looks at the current file / directory. For this I defined the SVN task in the ant file to make it more portable. <taskdef name="svn" classname="org.tigris.subversion.svnant.SvnTask"> <classpath> <pathelement location="${dir.lib}/ant/svnant.jar"/> <pathelement location="${dir.lib}/ant/svnClientAdapter.jar"/> <pathelement location="${dir.lib}/ant/svnkit.jar"/> <pathelement location="${dir.lib}/ant/svnjavahl.jar"/> </classpath> </taskdef> Not all builds result in webservices. The ear file before deployment must remain the same name because of updating in the application server. Making the file executable is still an option, but until then I just include a version information file. <target name="version"> <svn><wcVersion path="${dir.source}"/></svn> <echo file="${dir.build}/VERSION">${revision.range}</echo> </target> Refs: svnrevision: http://svnbook.red-bean.com/en/1.1/re57.html svn info http://svnbook.red-bean.com/en/1.1/re13.html subclipse svn task: http://subclipse.tigris.org/svnant/svn.html svn client: http://svnkit.com/

    Read the article

  • a question about rails general practice with REST, json, and ajax

    - by Nik
    Hi all, I have this question concerning REST I think: I have read a few rest tutorials and the feeling I get from them is that each action in a restful controller tends to be lean and almost single purpose: "Index gives off a collection of a model show gives off one model edit/new a prep place for changing/create a model update/create changes and makes new model deletes removes one model" After reading all these tutorials, rest seems to be to be a means to create an interface for a model, much like active resource type of thing. the mantra seems to be "controller provides data and data only and is also pretty convention over configuration, so expect projects_path to return a bunch of projects" I can understand that, and I like the cleanliness. But here's when I run into some trouble in reality in applying these guidelines: say three models, Project with attrib title, User with attrib name, and Location with attrib address. Say in views/users/index.html.erb, I want to use Ajax to fetch and display a project in a div#project_display when the user clicks on a project element, I know that I can use views/projects/show.js.rjs like this: page.replace_html 'project_display' "#{@project.name}" where in the projects_controller.rb def show @project = Project.find(params[:id]) repsond_to do |format| format.js and other formats... end end I have no problem in doing that for a couple of years now. BUT doesn't that mean that my JS response for the project#show action is LOCkED to present data to div#project_display element and show only whatever I that rjs template says it should show? That's very limiting and doesn't sound very "interface" like. I have never used JSON before or much XML, so I thought, maybe the JS response should send back raw stuff, like JSON and somehow the page on which the ajax request was called has the instruction on what do to with these raw data. That sounds a lot more flexible, doesn't it? Because look back at that exmpale, what if in the views/locations/index.html.erb, I want to do the exact same thing except I want to put the response in div#project_goes_here and the response should be #{project.name} I know this is a trivial change but that's the point: the RJS only allows one template at a time. So I think the JSON route is the way to go, but how does the already loaded page, the one that the ajax call came from, know when or how to "look forward" to incoming data? I read that PrototypeJS has this template thing, I wouldn't mind using it with JSON, but if you can demonstrate this or other means for displaying received-from-ajax data, I am all attention. Thank You

    Read the article

  • Job queue manager with RPC interface

    - by admr
    I need a job queue manager that I can control over the Internet. It should be able to execute and stop processes, check on their status (ideally notice and execute some code when a process exits), respond to commands and also be able to report back to a server. Background: I have a GWT application that allows to create jobs to execute on a cloud instance (currently EC2). I want to push a "job packet" (data for a process to operate on etc) to S3, start a Linux EC2 instance (or use one that's already running), and tell a job manager on the instance to execute that job (possibly parallel to other jobs). It should then pull the "job packet" from S3, run a process that operates on that data and report back to the server that is running the server part of my GWT application with some information (e.g. exit code, stdout, stderr). If I have to write e.g. stdour/err to a file from the process and read that file, that's OK too. I would really like the manager to be "close" to the processes it runs, meaning I want to avoid using something like Runtime.exec from the JDK. It seems like I would have to do that if I used Quartz for example. I'm fine with the calls in both directions being asynchronous. I'm fine with any reasonable technology for the calls as long as I can easily build an interface for that in my GWT server side (e.g. HTTP requests to a servlet over SSL would be nice and trivial). The job manager does not need to have a very sophisticated queueing system. Running several processes either sequentially or in parallel should be fine. Determining how much compute time a process received during its lifetime would be nice (AFAIK, this might be challenging). I did not yet find any existing software that does this, including http://java-source.net/open-source/job-schedulers. I suspect I might have to build an RPC interface (with authentication etc, of course) around a job manager; maybe use something like Apache Commons Exec. In that case, I would prefer Java or Python for the job manager part. I would be happy to hear suggestions for either the former or latter scenario!

    Read the article

  • Memory efficient int-int dict in Python

    - by Bolo
    Hi, I need a memory efficient int-int dict in Python that would support the following operations in O(log n) time: d[k] = v # replace if present v = d[k] # None or a negative number if not present I need to hold ~250M pairs, so it really has to be tight. Do you happen to know a suitable implementation (Python 2.7)? EDIT Removed impossible requirement and other nonsense. Thanks, Craig and Kylotan! To rephrase. Here's a trivial int-int dictionary with 1M pairs: >>> import random, sys >>> from guppy import hpy >>> h = hpy() >>> h.setrelheap() >>> d = {} >>> for _ in xrange(1000000): ... d[random.randint(0, sys.maxint)] = random.randint(0, sys.maxint) ... >>> h.heap() Partition of a set of 1999530 objects. Total size = 49161112 bytes. Index Count % Size % Cumulative % Kind (class / dict of class) 0 1 0 25165960 51 25165960 51 dict (no owner) 1 1999521 100 23994252 49 49160212 100 int On average, a pair of integers uses 49 bytes. Here's an array of 2M integers: >>> import array, random, sys >>> from guppy import hpy >>> h = hpy() >>> h.setrelheap() >>> a = array.array('i') >>> for _ in xrange(2000000): ... a.append(random.randint(0, sys.maxint)) ... >>> h.heap() Partition of a set of 14 objects. Total size = 8001108 bytes. Index Count % Size % Cumulative % Kind (class / dict of class) 0 1 7 8000028 100 8000028 100 array.array On average, a pair of integers uses 8 bytes. I accept that 8 bytes/pair in a dictionary is rather hard to achieve in general. Rephrased question: is there a memory-efficient implementation of int-int dictionary that uses considerably less than 49 bytes/pair?

    Read the article

  • QValidator for hex input

    - by Evan Teran
    I have a Qt widget which should only accept a hex string as input. It is very simple to restrict the input characters to [0-9A-Fa-f], but I would like to have it display with a delimiter between "bytes" so for example if the delimiter is a space, and the user types 0011223344 I would like the line edit to display 00 11 22 33 44 Now if the user presses the backspace key 3 times, then I want it to display 00 11 22 3. I almost have what i want, so far there is only one subtle bug involving using the delete key to remove a delimiter. Does anyone have a better way to implement this validator? Here's my code so far: class HexStringValidator : public QValidator { public: HexStringValidator(QObject * parent) : QValidator(parent) {} public: virtual void fixup(QString &input) const { QString temp; int index = 0; // every 2 digits insert a space if they didn't explicitly type one Q_FOREACH(QChar ch, input) { if(std::isxdigit(ch.toAscii())) { if(index != 0 && (index & 1) == 0) { temp += ' '; } temp += ch.toUpper(); ++index; } } input = temp; } virtual State validate(QString &input, int &pos) const { if(!input.isEmpty()) { // TODO: can we detect if the char which was JUST deleted // (if any was deleted) was a space? and special case this? // as to not have the bug in this case? const int char_pos = pos - input.left(pos).count(' '); int chars = 0; fixup(input); pos = 0; while(chars != char_pos) { if(input[pos] != ' ') { ++chars; } ++pos; } // favor the right side of a space if(input[pos] == ' ') { ++pos; } } return QValidator::Acceptable; } }; For now this code is functional enough, but I'd love to have it work 100% as expected. Obviously the ideal would be the just separate the display of the hex string from the actual characters stored in the QLineEdit's internal buffer but I have no idea where to start with that and I imagine is a non-trivial undertaking. In essence, I would like to have a Validator which conforms to this regex: "[0-9A-Fa-f]( [0-9A-Fa-f])*" but I don't want the user to ever have to type a space as delimiter. Likewise, when editing what they types, the spaces should be managed implicitly.

    Read the article

  • Repeating fields in similar database tables

    - by user1738833
    I have been tasked with working on a database that I have never seen before and I'm looking at the DB structure. Some of the central and most heavily queried and joined tables look like virtual duplicates of each other. Here's a massively simplified representation of the situation, with business-sensitive information changed, listing hypothetical table names and fields: TopLevelGroup: PK_TLGroupId, DisplaysXOnBill, DisplaysYOnBill, IsInvoicedForJ, IsInvoicedForK SubGroup: PK_SubGroupId, FK_ParentTopLevelGroupId, DisplaysXOnBill, DisplaysYOnBill, IsInvoicedForJ, IsInvoicedForK SubSubGroup: PK_SubSUbGroupId, FK_ParentSubGroupId, DisplaysXOnBill, DisplaysYOnBill, IsInvoicedForJ, IsInvoicedForK I haven't listed the types of the fields as I don't think it's particularly important to the situation. In addition, it's worth saying that rather than four repeated fields as in the example above, I'm looking at 86 repeated fields. For the most part, those fields genuinely do represent "facts" about the primary table entity, so it's not automatically wrong for that reason. In addition, the "groups" represented here have a property inheritance relationship. If DisplaysXOnBill is NULL in the SubSubGroup, it takes the value of DisplaysXOnBillfrom it's parent, the SubGroup, and so-on up to the TopLevelGroup. Further, the requirements will never require that the model extends beyond three levels, so there is no need for flexibility in that area. Is there a design smell from several tables which describe very similar entities having almost identical fields? If so, what might be a better design of the example above? I'm using the phrase "design smell" to indicate a possible problem. Of course, in any given situation, a particular design might well be the best solution. I'm looking for a more general answer - wondering what might be wrong with this design and what might be the better design were that the case. Possibly related, but not primary questions: Is this database schema in a reasonably normal form (e.g. to 3NF), insofar as can be told from the information I've provided. I can't see a problem with the requirements of 2NF and 3NF, except in their inheriting the requirements of 1NF. Is 1NF satisfied though? Are repeating groups allowed in different tables? Is there a best-practice method for implementing the inheritance relationship in a database as I require? The method above feels clunky to me because any query on the SubSubGroup necessarily needs to join onto the SubGroup and the TopLevelGroup tables to collect inherited facts, which can make even trivial joins requiring facts from the SubSubGroup table rather long-winded. There are, of course, political considerations to making a relatively large change like this. For the purpose of this question, I'm happy to ignore that fact in the interests of keeping the answers ring-fenced to the technical problem.

    Read the article

  • Consolidating coding styles: Funcs, private method, single method classes

    - by jdoig
    Hi all, We currently have 3 devs with, some, conflicting styles and I'm looking for a way to bring peace to the kingdom... The Coders: Foo 1: Likes to use Func's & Action's inside public methods. He uses actions to alias off lengthy method calls and Func's to perform simple tasks that can be expressed in 1 or 2 lines and will be used frequently through out the code Pros: The main body of his code is succinct and very readable, often with only one or 2 public methods per class and rarely any private methods. Cons: The start of methods contain blocks of lambda rich code that other developers don't enjoy reading; and, on occasion, can contain higher order functions that other dev's REALLY don't like reading. Foo 2: Likes to create a private method for (almost) everything the public method will have to do . Pros: Public methods remain small and readable (to all developers). Cons: Private methods are numerous. With private methods that call into other private methods, that call into... etc, etc. Making code hard to navigate. Foo 3: Likes to create a public class with a, single, public method for every, non-trivial, task that needs performing, then dependency inject them into other objects. Pros: Easily testable, easy to understand (one object, one responsibility). Cons: project gets littered by classes, opening multiple class files to understand what code does makes navigation awkward. It would be great to take the best of all these techniques... Foo-1 Has really nice, readable (almost dsl-like) code... for the most part, except for all the Action and Func lambda shenanigans bulked together at the start of a method. Foo-3 Has highly testable and extensible code that just feels a bit "belt-&-braces" for some solutions and has some code-navigation niggles (constantly hitting F12 in VS and opening 5 other .cs files to find out what a single method does). And Foo-2... Well I'm not sure I like anything about the one-huge .cs file with 2 public methods and 12 private ones, except for the fact it's easier for juniors to dig into. I admit I grossly over-simplified the explanations of those coding styles; but if any one knows of any patterns, practices or diplomatic-manoeuvres that can help unite our three developers (without just telling any of them to just "stop it!") that would be great. From a feasibility standpoint : Foo-1's style meets with the most resistance due to some developers finding lambda and/or Func's hard to read. Foo-2's style meets with a less resistance as it's just so easy to fall into. Foo-3's style requires the most forward thinking and is difficult to enforce when time is short. Any ideas on some coding styles or conventions that can make this work?

    Read the article

  • Modular GWT design concerns

    - by GlGuru
    Hi, I have a couple of questions regarding a modular GWT based application framework. I have some ideas about them but being new to the field of web development I feel they are far from ideal. I'd appreciate a few comments and suggestions in this regard. Here are my questions: I am developing a framework which will allow third parties to embed GWT applications into our website and do some communication with them using simple iFrame postMessage. All these third party modules are going to use our SDK which is also GWT based. The problem arises that even though all the modules will be using the same codebase there is going to be a massive overheard in the amount of duplicate Javascript code (i.e. our common SDK code base which is quite large) being downloaded on the client's machine. This is highly redundant and problematic, not only due to the sheer size of duplicate code but, also due to the fact that subsequent updates of the SDK would require the modules to be recompiled which is going to create a DLL hell kind of scenario in the long run. What is the best way of doing this kind of thing? Is there a way where I can have some static GWT code (i.e. the SDK) and the dynamic GWT module refers to it (even if lies on a different domain) and it all work happily? The other part of the problem lies in providing robust scripting front end to the SDK. At first it appears to be trivial since Javascript itself is a scripting language. However, I do not know how to load and call a piece of pure Javascript code at runtime? I am willing to put restrictions on the target Javascript (i.e. having a function main and unique namespace or something). Furthermore the Javascript will come as a string from a database and not as a full URL. If its doable in Javascript how does one get this right in GWT i.e. forcing the compiler to emit a certain function in the generated Javascript? This I believe can be lesser of a problem by having a stub Javascript with all the right requirements which just loads up a GWT generated Javascript. Is any of this possible at all? I generally hate to be this verbose but I hope to find a quick solution to the problem as its holding up my progress. I'd highly appreciate any comments, suggestions and experiences.

    Read the article

  • c++ class member functions selected by traits

    - by Jive Dadson
    I am reluctant to say I can't figure this out, but I can't figure this out. I've googled and searched stackoverflow, and come up empty. The abstract, and possibly overly vague form of the question is, how can I use the traits-pattern to instantiate non-virtual member functions? The question came up while modernizing a set of multivariate function optimizers that I wrote more than 10 years ago. The optimizers all operate by selecting a straight-line path through the parameter space away from the current best point (the "update"), then finding a better point on that line (the "line search"), then testing for the "done" condition, and if not done, iterating. There are different methods for doing the update, the line-search, and conceivably for the done test, and other things. Mix and match. Different update formulae require different state-variable data. For example, the LMQN update requires a vector, and the BFGS update requires a matrix. If evaluating gradients is cheap, the line-search should do so. If not, it should use function evaluations only. Some methods require more accurate line-searches than others. Those are just some examples. The original version instatiates several of the combinations by means of virtual functions. Some traits are selected by setting mode bits. Yuck. It would be trivial to define the traits with #define's and the member functions with #ifdef's and macros. But that's so twenty years ago. It bugs me that I cannot figure out a whiz-bang modern way. If there were only one trait that varied, I could use the curiously recurring template pattern. But I see no way to extend that to arbitrary combinations of traits. I tried doing it using boost::enable_if, etc.. The specialized state info was easy. I managed to get the functions done, but only by resorting to non-friend external functions that have the this-pointer as a parameter. I never even figured out how to make the functions friends, much less member functions. Perhaps tag-dispatch is the key. I haven't gotten very deeply into that. Surely it's possible, right? If so, what is best practice?

    Read the article

  • Scalable Database Tagging Schema

    - by Longpoke
    EDIT: To people building tagging systems. Don't read this. It is not what you are looking for. I asked this when I wasn't aware that RDBMS all have their own optimization methods, just use a simple many to many scheme. I have a posting system that has millions of posts. Each post can have an infinite number of tags associated with it. Users can create tags which have notes, date created, owner, etc. A tag is almost like a post itself, because people can post notes about the tag. Each tag association has an owner and date, so we can see who added the tag and when. My question is how can I implement this? It has to be fast searching posts by tag, or tags by post. Also, users can add tags to posts by typing the name into a field, kind of like the google search bar, it has to fill in the rest of the tag name for you. I have 3 solutions at the moment, but not sure which is the best, or if there is a better way. Note that I'm not showing the layout of notes since it will be trivial once I get a proper solution for tags. Method 1. Linked list tagId in post points to a linked list in tag_assoc, the application must traverse the list until flink=0 post: id, content, ownerId, date, tagId, notesId tag_assoc: id, tagId, ownerId, flink tag: id, name, notesId Method 2. Denormalization tags is simply a VARCHAR or TEXT field containing a tab delimited array of tagId:ownerId. It cannot be a fixed size. post: id, content, ownerId, date, tags, notesId tag: id, name, notesId Method 3. Toxi (from: http://www.pui.ch/phred/archives/2005/04/tags-database-schemas.html, also same thing here: http://stackoverflow.com/questions/20856/how-do-you-recommend-implementing-tags-or-tagging) post: id, content, ownerId, date, notesId tag_assoc: ownerId, tagId, postId tag: id, name, notesId Method 3 raises the question, how fast will it be to iterate through every single row in tag_assoc? Methods 1 and 2 should be fast for returning tags by post, but for posts by tag, another lookup table must be made. The last thing I have to worry about is optimizing searching tags by name, I have not worked that out yet. I made an ASCII diagram here: http://pastebin.com/f1c4e0e53

    Read the article

  • Objective-C Basic class related question, retaining the value of a specific object using a class fil

    - by von steiner
    Members, scholars, code gurus. My background is far from any computer programming thus my question may seems basic and somewhat trivial to you. Nevertheless it seems that I can't put my head around it. I have googled and searched for the answer, just to get myself confused even more. With that, I would kindly ask for a simple explanation suitable for a non technical person such as myself and for other alike arriving to this thread. I have left a comment with the text "Here is the issue" below, referring to my question. // character.h #import <Foundation/Foundation.h> @interface character : NSObject { NSString *name; int hitPoints; int armorClass; } @property (nonatomic,retain) NSString *name; @property int hitPoints,armorClass; -(void)giveCharacterInfo; @end // character.m #import "character.h" @implementation character @synthesize name,hitPoints,armorClass; -(void)giveCharacterInfo{ NSLog(@"name:%@ HP:%i AC:%i",name,hitPoints,armorClass); } @end // ClassAtLastViewController.h #import <UIKit/UIKit.h> @interface ClassAtLastViewController : UIViewController { } -(void)callAgain; @end // ClassAtLastViewController.m #import "ClassAtLastViewController.h" #import "character.h" @implementation ClassAtLastViewController - (void)viewDidLoad { //[super viewDidLoad]; character *player = [[character alloc]init]; player.name = @"Minsc"; player.hitPoints = 140; player.armorClass = 10; [player giveCharacterInfo]; [player release]; // Up until here, All peachy! [self performSelector:@selector(callAgain) withObject:nil afterDelay:2.0]; } -(void)callAgain{ // Here is the issue, I assume that since I init the player again I loss everything // Q1. I loss all the data I set above, where is it than? // Q2. What is the proper way to implement this character *player = [[character alloc]init]; [player giveCharacterInfo]; } Many thanks in advance, Kindly remember that my background is more related to Salmons breeding than to computer code, try and lower your answer to my level if it's all the same to you.

    Read the article

  • Precise explanation of JavaScript <-> DOM circular reference issue

    - by Joey Adams
    One of the touted advantages of jQuery.data versus raw expando properties (arbitrary attributes you can assign to DOM nodes) is that jQuery.data is "safe from circular references and therefore free from memory leaks". An article from Google titled "Optimizing JavaScript code" goes into more detail: The most common memory leaks for web applications involve circular references between the JavaScript script engine and the browsers' C++ objects' implementing the DOM (e.g. between the JavaScript script engine and Internet Explorer's COM infrastructure, or between the JavaScript engine and Firefox XPCOM infrastructure). It lists two examples of circular reference patterns: DOM element → event handler → closure scope → DOM DOM element → via expando → intermediary object → DOM element However, if a reference cycle between a DOM node and a JavaScript object produces a memory leak, doesn't this mean that any non-trivial event handler (e.g. onclick) will produce such a leak? I don't see how it's even possible for an event handler to avoid a reference cycle, because the way I see it: The DOM element references the event handler. The event handler references the DOM (either directly or indirectly). In any case, it's almost impossible to avoid referencing window in any interesting event handler, short of writing a setInterval loop that reads actions from a global queue. Can someone provide a precise explanation of the JavaScript ↔ DOM circular reference problem? Things I'd like clarified: What browsers are effected? A comment in the jQuery source specifically mentions IE6-7, but the Google article suggests Firefox is also affected. Are expando properties and event handlers somehow different concerning memory leaks? Or are both of these code snippets susceptible to the same kind of memory leak? // Create an expando that references to its own element. var elem = document.getElementById('foo'); elem.myself = elem; // Create an event handler that references its own element. var elem = document.getElementById('foo'); elem.onclick = function() { elem.style.display = 'none'; }; If a page leaks memory due to a circular reference, does the leak persist until the entire browser application is closed, or is the memory freed when the window/tab is closed?

    Read the article

  • Backtracking infinite loop

    - by Greenhorn
    This is Exercise 28.1.2 from HtDP. I've successfully implemented the neighbors function and all test cases pass. (define Graph (list (list 'A (list 'B 'E)) (list 'B (list 'E 'F)) (list 'C (list 'D)) (list 'D empty) (list 'E (list 'C 'F)) (list 'F (list 'D 'G)) (list 'G empty))) (define (first-line n alist) (cond [(symbol=? (first alist) n) alist] [else empty])) ;; returns empty if node is not in graph (define (neighbors n g) (cond [(empty? g) empty] [(cons? (first g)) (cond [(symbol=? (first (first g)) n) (first-line n (first g))] [else (neighbors n (rest g))])])) ; test cases (equal? (neighbors 'A Graph) (list 'A (list 'B 'E))) (equal? (neighbors 'B Graph) (list 'B (list 'E 'F))) (equal? (neighbors 'C Graph) (list 'C (list 'D))) (equal? (neighbors 'D Graph) (list 'D empty)) (equal? (neighbors 'E Graph) (list 'E (list 'C 'F))) (equal? (neighbors 'F Graph) (list 'F (list 'D 'G))) (equal? (neighbors 'G Graph) (list 'G empty)) (equal? (neighbors 'H Graph) empty) The problem comes when I copy-paste the code from Figure 77 of the text. It is supposed to determine whether a destination node is reachable from an origin node. However it appears that the code goes into an infinite loop except for the most trivial case where the origin and destination nodes are the same. ;; find-route : node node graph -> (listof node) or false ;; to create a path from origination to destination in G ;; if there is no path, the function produces false (define (find-route origination destination G) (cond [(symbol=? origination destination) (list destination)] [else (local ((define possible-route (find-route/list (neighbors origination G) destination G))) (cond [(boolean? possible-route) false] [else (cons origination possible-route)]))])) ;; find-route/list : (listof node) node graph -> (listof node) or false ;; to create a path from some node on lo-Os to D ;; if there is no path, the function produces false (define (find-route/list lo-Os D G) (cond [(empty? lo-Os) false] [else (local ((define possible-route (find-route (first lo-Os) D G))) (cond [(boolean? possible-route) (find-route/list (rest lo-Os) D G)] [else possible-route]))])) Does the problem lie in my code? Thank you.

    Read the article

  • c++ class member functions instatiated by traits

    - by Jive Dadson
    I am reluctant to say I can't figure this out, but I can't figure this out. I've googled and searched stackoverflow, and come up empty. The abstract, and possibly overly vague form of the question is, how can I use the traits-pattern to instantiate non-virtual member functions? The question came up while modernizing a set of multivariate function optimizers that I wrote more than 10 years ago. The optimizers all operate by selecting a straight-line path through the parameter space away from the current best point (the "update"), then finding a better point on that line (the "line search"), then testing for the "done" condition, and if not done, iterating. There are different methods for doing the update, the line-search, and conceivably for the done test, and other things. Mix and match. Different update formulae require different state-variable data. For example, the LMQN update requires a vector, and the BFGS update requires a matrix. If evaluating gradients is cheap, the line-search should do so. If not, it should use function evaluations only. Some methods require more accurate line-searches than others. Those are just some examples. The original version instantiates several of the combinations by means of virtual functions. Some traits are selected by setting mode bits that are tested at runtime. Yuck. It would be trivial to define the traits with #define's and the member functions with #ifdef's and macros. But that's so twenty years ago. It bugs me that I cannot figure out a whiz-bang modern way. If there were only one trait that varied, I could use the curiously recurring template pattern. But I see no way to extend that to arbitrary combinations of traits. I tried doing it using boost::enable_if, etc.. The specialized state info was easy. I managed to get the functions done, but only by resorting to non-friend external functions that have the this-pointer as a parameter. I never even figured out how to make the functions friends, much less member functions. The compiler (vc++ 2008) always complained that things didn't match. I would yell, "SFINAE, you moron!" but the moron is probably me. Perhaps tag-dispatch is the key. I haven't gotten very deeply into that. Surely it's possible, right? If so, what is best practice?

    Read the article

< Previous Page | 40 41 42 43 44 45 46 47 48 49 50 51  | Next Page >