- Boost - lists.boost.org

[Variant]
by Cosimo Calabrese 06 Oct '09

06 Oct '09

Hi to all, I've looked at the following example: http://www.boost.org/doc/libs/1_40_0/doc/html/variant.html#variant.motivati… and I think that miss an include: #include <string> Best regards, Cosimo Calabrese.

2 1

Off-topic question about Boost trac
by Mateusz Loskot 05 Oct '09

05 Oct '09

Hi, First, forgive me I'm sending this off-topic question to the mailing list, but I'm actually not sure where it would belong. Probably, I should direct my question to Boost Trac admins, but I don't know how to reach them. So, here we go: I'm involved in OSGeo Foundation and we maintain number of Trac instances (http://trac.osgeo.org/) for our projects. For longer while, we've been experiencing server load issues related to Trac. I noticed that Boost Trac displays "Server Load Average" information in below the menu bar. I would like to setup the same feature for OSGeo but I can not find any information how to do it. Could you tell me how you've got this for Boost Trac? Perhaps you could also tell me why you've decided to display this "Server Load Average" information in the Boost Trac? Does it mean Boost were experiencing load issues? Thanks for your help! Best regards, -- Mateusz Loskot, http://mateusz.loskot.net Charter Member of OSGeo, http://osgeo.org

1 0

[task] reschedule_until
by Baxter Jeremy 05 Oct '09

05 Oct '09

> Date: Fri, 2 Oct 2009 18:59:10 +0200 > From: k-oli(a)gmx.de > Subject: Re: [boost] [task] > > Second - I'm finishing boost.task-0.4.0 which replaces > this_task::reschedule_until() by this_task::block(). > block() suspends the current task without blocking the > worker-thread. The > blocked task gets rescheduled by the internal scheduler > (returning from > block()): > > while ( ! condition) > this_task::block(); > > Version 0.4.0 allows to synchronize or exchange messages > between the tasks > without deadlocking the threadpool (that means amount of > tasks >> worker- > threads in pool). > > regards, > Oliver Thanks - I'll steer clear of using reschedule_until at the moment. Being able to block a task without locking out the worker thread would be useful for me, but I suspect my requirement may be rather different to yours. I'm pretty much assuming amount of tasks == worker-threads in pool. My main interest in a thread pool is to avoid creating and destroying a new thread each time a task has to run but I suspect I may use task or threadpool it this way initially and see what the overhead is like. Jeremy The information contained in this E-Mail and any subsequent correspondence is private and is intended solely for the intended recipient(s). The information in this communication may be confidential and/or legally privileged. Nothing in this e-mail is intended to conclude a contract on behalf of QinetiQ or make QinetiQ subject to any other legally binding commitments, unless the e-mail contains an express statement to the contrary or incorporates a formal Purchase Order. For those other than the recipient any disclosure, copying, distribution, or any action taken or omitted to be taken in reliance on such information is prohibited and may be unlawful. Emails and other electronic communication with QinetiQ may be monitored and recorded for business purposes including security, audit and archival purposes. Any response to this email indicates consent to this. Telephone calls to QinetiQ may be monitored or recorded for quality control, security and other business purposes. QinetiQ Limited Registered in England & Wales: Company Number:3796233 Registered office: 85 Buckingham Gate, London SW1E 6PD, United Kingdom Trading address: Cody Technology Park, Cody Building, Ively Road, Farnborough, Hampshire, GU14 0LX, United Kingdom http://www.qinetiq.com/home/notices/legal.html

2 1

[task] reschedule_until
by Baxter Jeremy 05 Oct '09

05 Oct '09

- my example and some of Vicente's comment snipped - > > I think that it will not to much complex to provide a > function that submit a task that must run on the context of a > different thread. Would this kind of submit function help you? > I'm not sure that it would help, I think that I could construct a more complex example with more tasks and threads that would still deadlock. In general I think it ought to be a requirement that the threadpool works if it has only a single thread (in which case the above would not work). > > I'm not sure that there is a general solution to this > problem without > > using more threads so that any potentially runnable task will get a > > chance. > > You are right this should be a deadlock problem with the > current implementation. > I don't think however that having more threads would solve > the issue, but just reduce its frequency, as the issue > appears when tasks T1 and T2 are handled on the same thread. > I agree, that is why I was not keen on reschedule-until. If blocking is used I think an unexpected deadlock could occur unless all blocked tasks can be resumed in any order, not strictly in reverse order of their reschedule calls as in the current implementation. > In order to avoid the issue we need to avoid two blocking > tasks to be executed on the same thread. The user should state > blocking task on the submit function. The scheduler will not > take a blocking task while in the reschedule_until function. > Of course, in order to be able to avoid any deadlock > associated to the fact we are using a thread pool, the number > of threads must grow, if there is no free worker threads able > to handle the new blocking task. The reschedule_until > function should throw an exception when the task is not a > blocking task. > > This should solve the issue, but introduce a burden at the > user level, i.e. specify if the task can block or not. > > Jeremy, Oliver, what do you think? > > Best, > Vicente > > I think Oliver's intention to use a fiber style scheduler is the right one. To avoid problems all possible mechanisms of blocking would need to be considered and protected against and the temptation would be to declare most tasks as blocking. Even if tasks don't block they may run on the thread for a long time before returning. A dynamic thread pool would also be a solution, starting more threads as tasks block and cleaning them up when the queue is empty. This is the approach used in the Java Executors class (http://java.sun.com/j2se/1.5.0/docs/api/java/util/concurrent/Executors. html). I'm not sure if there is a way to detect the difference between threads being busy and being blocked as in general you would like to keep the number of unblocked threads down to the number of cores. Jeremy Jeremy Baxter Lead Researcher, UAVs & Autonomy Tel: 01684 894801 email: jwbaxter(a)QinetiQ.com www.QinetiQ.com QinetiQ - Delivering customer focused solutions Please consider the environment before printing this email. The information contained in this E-Mail and any subsequent correspondence is private and is intended solely for the intended recipient(s). The information in this communication may be confidential and/or legally privileged. Nothing in this e-mail is intended to conclude a contract on behalf of QinetiQ or make QinetiQ subject to any other legally binding commitments, unless the e-mail contains an express statement to the contrary or incorporates a formal Purchase Order. For those other than the recipient any disclosure, copying, distribution, or any action taken or omitted to be taken in reliance on such information is prohibited and may be unlawful. Emails and other electronic communication with QinetiQ may be monitored and recorded for business purposes including security, audit and archival purposes. Any response to this email indicates consent to this. Telephone calls to QinetiQ may be monitored or recorded for quality control, security and other business purposes. QinetiQ Limited Registered in England & Wales: Company Number:3796233 Registered office: 85 Buckingham Gate, London SW1E 6PD, United Kingdom Trading address: Cody Technology Park, Cody Building, Ively Road, Farnborough, Hampshire, GU14 0LX, United Kingdom http://www.qinetiq.com/home/notices/legal.html

2 1

Re: [boost] [rdb] rdb/persistent integration
by strasser＠uni-bremen.de 05 Oct '09

05 Oct '09

>> also, you could use more than one resource managers at a time, so you could >> have distributed transactions between two databases, between a file and a >> database, etc., and can have references across resources. (a locator can >> reference an object in any resource, and can be stored as part of an object) >> > Very interesting project. At this point my feeling is that there is room > to slip not one but two of my libs under yours ! > what`s the other library? btw, I wasn`t suggesting we should merge our libraries. a stand-alone sql db layer definitely makes sense. even a standalone o-r mapper might make sense, depending on what kind of user interface you planned for it. it might e.g. be better for some use cases to give the user more control over when and how actual SQL is produced and objects are stored than my library does. much like my resource manager uses Boost.Serialization, but doesn`t replace Boost.Serialization. remember that my intent was a user interface as close as possible to using regular objects in memory, so there is no "store this object now" function. the resource manager is free to cache and postpone accesses to the storage engine, as long as it guarantees that the object is stored on transaction::commit the latest. but we should work together on "linking" the libraries, i.e. implementing a boost.persistent resource manager that uses your o-r mapper, in the current design. or merge your o-r mapper library with my library, but leave the sql db layer a seperate library. depending on the o-r mapper interface you planned.

1 0

Looking for a name for a Subjet-Entity-Role library
by vicente.botet 05 Oct '09

05 Oct '09

Hi, I'm developing a library based on the Object Role pattern. The library use the following concepts: Entity, Role, Subject, Cloture. Its main goal is to help on subjetive programming. I was started to name this library SER (for Subject-Entity-Role). Next follows some Entity-Role-Subject declarations class Person : public entity<Person> {...}; template <typename Base> class PersonRoleTmpl : public basic_role<Person, Base> {...}; class PersonRole : public PersonRoleTmpl<> {...};class Employee : public role<Employee,PersonRole> {...}; class Salesman : public Employee {...}; class Developer : public Employee {...}; class Manager : public Employee {...}; class Guarantor : public role<Customer, PersonRole> {...}; class Customer : public role<Customer, PersonRoleTmpl<entity<Customer> > > {...}; class CustomerRole : public basic_role<Customer> {...}; class Borrower : public role<Borrower, CustomerRole> {...}; class Investor : public role<Investor, CustomerRole> {...}; class CustomerAndEmployee : public subject<Person, roles<Customer,Employee> > { ... }; Any sugestions for the name of the library, 'Subjetive', 'SER'? Best, _____________________ Vicente Juan Botet Escribá

5 7

[sandbox] Mysterious (?) commit messages, 'm', 'a', 'd'
by Niels Dekker - address until 2010-10-10 04 Oct '09

04 Oct '09

There appear commit messages of a single letter: 'd', 'm', 'M', 'a'. Or combined: "m/a". It seems like they only appear in the sandbox. https://svn.boost.org/trac/boost/timeline What do they mean? (Maybe "e_r" can tell us.) FWIW, I'd rather see more "human friendly" commit messages. Thanks in advance, Niels -- Niels Dekker http://www.xs4all.nl/~nd/dekkerware Scientific programmer at LKEB, Leiden University Medical Center

2 1

GIL extensions home and guide
by Mateusz Loskot 04 Oct '09

04 Oct '09

Hi, May I ask for update where is most current home for Boost/Adobe GIL library extensions [0]? What's the difference beetween extensions in Boost SVN [1] and those hosted in Google Code [2] ? Second, what document should one refer to learn about developing new I/O extensions for GIL? The GIL Design Guide [3] is quite dated, from 2007 [0] http://opensource.adobe.com/wiki/display/gil/Contribute+to+GIL [1] https://svn.boost.org/trac/boost/browser/trunk/boost/gil/extension/ [2] http://code.google.com/p/gil-contributions/ [3] http://www.boost.org/doc/libs/1_40_0/libs/gil/doc/html/gildesignguide.html Best regards, -- Mateusz Loskot, http://mateusz.loskot.net Charter Member of OSGeo, http://osgeo.org

2 2

[rdb] 0.0.09
by Jean-Louis Leroy 04 Oct '09

04 Oct '09

In this version : * EXISTS * consistency check in INSERT...SELECT * DISTINCT, ALL * transactions The transaction stuff is pellicular. It simply maps to vendor calls. I am hesitating a bit, either make a common interface for the most common operations (like execute) directly in each backend (aside with vendor-specifc calls) or put them in a separate wrapper or even decide that it's not in the scope. There's more redesign in the select from which I think I can get an elegant implementation of vendor-specific features (like LIMIT). The idea is to pass all the data in a fusion map, allowing vendor specific parts of the statement to slip their own data inside it. In fact this also is how DISTINCT works. I dropped the composition of lists via chained calls to operator (). It's all preprocessor-based now. The () chain system was limited by FUSION_MAX_VECTOR_SIZE anyway. I have installed MySQL on my computer. I am still accessing it via ODBC but I'll probably start the native support soon. Yesterday I tried to upload and just when I pushed the upload button the vault server crashed. I suspect that it was my upload that triggered it. J-L

6 15

[trac] Make me property_tree owner
by Sebastian Redl 04 Oct '09

04 Oct '09

Hi, Could somebody who has the authority to do so please make me (cornedbee) the owner of property_tree? Currently, all PTree tickets are still assigned to kaalus, who is the original author. Sebastian

2 1