Search Results

Search found 17437 results on 698 pages for 'nick long'.

Page 7/698 | < Previous Page | 3 4 5 6 7 8 9 10 11 12 13 14  | Next Page >

  • Manual Directory Submission Service a Long Way to Go But Effective

    Yes you heard it right Manual Directory Submission Service is really a long way to go but very effective for your business. In fact this is a process of registering or submitting your website with its details to an online web directory. The advantage of this listing is that when a customer needs any service or product, he will search for business on a web directory for assistance and when your expertise meet with the keywords that he typed in search bar, you website link will be visible to him and he will click that link from where he will be redirected to your website.

    Read the article

  • BPMN is dead, long live BPEL!

    - by JuergenKress
    “BPMN is dead, long live BPEL” was the title of our panel discussion during the SOA & BPM Integration Days 2011. At the JAXenter my discussion summery was just published (in German). If you want to learn more about SOA & BPM make sure you register for our up-coming conference October 12th & 13th 2011 in Düsseldorf. The speakers include the top SOA and BPM experts in Germany: Thilo Frotscher & Kornelius Fuhrer & Björn Hardegen & Nicolai Josuttis & Michael Kopp & Dr. Dirk Krafzig & Jürgen Kress & Frank Leymann & Berthold Maier & Hajo Normann & Max J. Pucher & Bernd Rücker & Dr. Gregor Scheithauer & Danilo Schmiedel & Guido Schmutz & Dirk Slama & Heiko Spindler & Volker Stiehl & Bernd Trops & Clemens Utschig-Utschig & Tammo van Lessen & Dr. Hendrik Voigt & Torsten Winterberg  For details please become a member in the SOA Partner Community for registration please visit  www.oracle.com/goto/emea/soa (OPN account required) Blog Twitter LinkedIn Mix Forum Wiki Website

    Read the article

  • SQL SERVER – Database in RESTORING State for Long Time

    - by Pinal Dave
    A very interesting question I received the other day. “Our database has been in restoring stage for a long time. We have already restored all the necessary files there. After restoring the files we are expecting that  the database will be in operational mode, however, it is continuously in the restoring mode. Any suggestion?” The question is very common. I sent user follow up emails to understand what is actually going on with the user. I realized after restoring their bak files and log files their database was in the restoring state because they had not restored the latest log file with RECOVERY options. As they had completed all the database restore sequence (bak and log in order), the real need for them was to recover the database from norecovery state. User can restore log files till the database is no recovery mode. If the database is recovered it will be in operation and it can continue database operation. If the database has another operations we cannot restore further log as the chain of the log file after the database is recovered is meaningless. This is the reason why the database has to be norecovery state when it is restored. There are three different ways to recover the database. 1) Recover the database manually with following command. RESTORE DATABASE database_name WITH RECOVERY 2) Recover the database with the last log file. RESTORE LOG database_name FROM backup_device WITH RECOVERY 3) Recover the database when bak is restored RESTORE DATABASE database_name FROM backup_device WITH RECOVERY To understand how the backup restores timeline works read Backup Timeline and Understanding of Database Restore Process in Full Recovery Model. Reference: Pinal Dave (http://blog.sqlauthority.com) Filed under: PostADay, SQL, SQL Authority, SQL Backup and Restore, SQL Query, SQL Server, SQL Tips and Tricks, T SQL, Technology

    Read the article

  • Installation taking a very long time, hangs at "Configuring bcmwl-kernel-source"

    - by user290522
    I am installing Ubuntu 14.04(32-bit) on my laptop (Compaq Presario V2000), and after about 7 hours, it is still in Configuring bcmwl-kernel-source (i386) mode. The messages I read are as follows: ubuntu kernel: [22814.858163] ACPI: \_SB_.PCI0.LPC0.LPC0.ACAD: ACPI_NOTIFY_BUS_CHECK event: unsupported with the numbers in the square brackets increasing. I have had Windows XP professional on this laptop, and I am erasing it. I am not sure if I should turn off the laptop, and start all over again. About 4 years ago I installed Ubuntu on this laptop, and that was very fast. The only problem I encountered was my wireless, and could not make it to work, and switched back to Windows. I appreciate any comments regarding this installation taking such a long time. After 40 hours the installation was still in configuring mode with the following messages: ubuntu CRON[29329]: (root) CMD ( cd/ && run-part .. report /etc/cron-hourly) I did the following to check for errors: pressed ctrl-alt-f2. This time the system froze. I had no other choice but to turn off the laptop, and start all over again. The exact model of the laptop is "Compaq Presario V2069CL Notebook PC" with AMD processor.

    Read the article

  • Part 3: Customization Strategy or how long does it take

    - by volker.eckardt(at)oracle.com
    The previous part in this blog should have made us aware, that many procedures are required to manage all these steps. To review your status let me ask you a question:What is your Customization Strategy?Your answer might be something like, 'customization strategy, well, we have standards and we let requirement documents approve'.Let me ask you another question:How long does it take to redeploy all your customizations into a fresh installation?In 90% of all installations the answer to this question would be: we can't!Although no one would have to do it (hopefully), just thinking about it and recognizing that we have today too many manual steps involved, different procedures and sometimes (undocumented) manual steps to complete a customization installation. And ... in general too many customizations.Why is working with customizations often so complicated and time consuming?Here are the key reasons as I have identified them in my projects:Customization standards defined, but not maintainedDifferent knowledge on developer side (results getting an individual developer touch)No need to automate deployment (not forced by client)Different documentation styles, not easy to hand over to someone elseDifferent development concepts, difficult for the maintenanceJust the minimum present for testing, often positive testing onlyDeviations from naming conventions accepted, although definedComplicated procedures, therefore sometimes partially ignoredAnd last but not least, hand made version control (still)If you would have to 'redeploy all your customizations' you would have to Follow all your own standards and best practiceTrack deviations and define corrective tasksAutomate as much as possible, minimize manual tasksDo not allow any change coming in without version controlUtilize products to support you in deploymentMinimize hand made scripts and extensive documentationReview regularly used techniques to guarantee that all are in line with the current release and also easy maintainableCreate solution libraries and force the team to contribute and reuseDefine quality activities and execute themDefine a procedure to release customizationsI know, it is easy to write down, but much harder to manage. Will provide some guidelines in my next blog.Volker

    Read the article

  • Craftsmanship Tour Day 1: Didit Long Island

    - by Liam McLennan
    On Monday I was at Didit for my first ever craftsmanship visit. Didit seem to occupy a good part of a non-descript building in Rockville Centre Long Island. Since I had arrived early from Seattle I had some time to kill, so I stopped at the Rockville Diner on the corner of N Park Ave and Sunrise Hwy. I thoroughly enjoyed the pancakes and the friendly service. After walking to the Didit office I met Rik Dryfoos, the Didit Engineering Manager who organised my visit, and got the introduction to Didit and the work they are doing. I spent the morning in the room shared by the Didit developers, who are working on some fascinating deep engineering problems. After lunch at a local Thai place I setup a webcam to record an interview with Rik and Matt Roman (Didit VP of Engineering). I had a lot of trouble with the webcam, including losing several minutes of conversation, but in the end I was very happy the result. Here are the full interviews with Rik and Matt: Interview with Rik Dryfoos Interview with Matt Roman We had a great chat, much of which is captured in the recording. It was such great conversation that I almost missed my train to Manhattan. I’m sure Didit will continue to do well with such a dedicated and enthusiastic team. I sincerely thank them for hosting me for the day. If you are looking for a true agile environment and the opportunity to work with a high quality team then you should talk to Didit.

    Read the article

  • Long term plan of attack to learn math?

    - by zhenka
    I am a web-developer with a desire to expand my skill-set to mathematics relevant to programming. As 2nd career, I am stuck in college doing some of the requirements while working. I was hoping the my education will teach me the needed skills to apply math, however I am quickly finding it to be too much easily-testable breadth-based approach very inefficient for the time invested. For example in my calculus 2 class, the only remotely useful mind expanding experience I had was volumes and areas under the curve. The rest was just monotonous glorified algebra, which while comes easy to me, could be done by software like wolfram alpha within seconds. This is not my idea of learning math. So here I am a frustrated student looking for a way to improve my understanding of math in a way that focuses on application, understanding and maximally removed needless tedium. However I cannot find a good long term study strategy with this approach in mind. So for those of like mind, how would you go about learning the necessary math without worrying too much about stuff a computer can do much better?

    Read the article

  • Exclusive use of a Jini server during long-running call

    - by Matthew Flint
    I'm trying to use Jini, in a "Masters/Workers" arrangement, but the Worker jobs may be long running. In addition, each worker needs to have exclusive access to a singleton resource on that machine. As it stands, if a Worker receives another request while a previous request is running, then the new request is accepted and executed in a second thread. Are there any best-practices to ensure that a Worker accepts no further jobs until the current job is complete? Things I've considered: synchronize the job on the server, with a lock on the singleton resource. This would work, but is far from ideal. A call from a Master would block until the current Worker thread completes, even if other Workers become free in the meantime unregister the Worker from the registry while the job is running, then re-register when it completes. Might work OK, but something doesn't smell right with this idea... Of course, I'm quite happy to be told that there are more appropriate technologies than Jini... but I wouldn't want anything too heavyweight, like rolling out EJB containers all over the place.

    Read the article

  • Google Analytics: How long does it take users to trigger an event

    - by Stephen Ostermiller
    I implemented Google Analytics event tracking on my currency conversion website. The typical user flow is: User lands on a page about two currencies. User enters an amount to be converted. The site shows the user the value in the other currency. The JavaScript sends Google Analytics an "converted" event when the currency conversion is done. Because most of the sessions on my site are single page, the event tracking is very important to me to be able to know if users find my page useful. I'm looking for a way to be able to figure out how long it typically takes users to enter a value in the form. I expect that this data would form a bell curve with around a specific amount of time after page load. If I can't get a graph, I could make do with a median value. I would like to be able to use this as a core metric around usability testing. Is there a way to get this information out of Google Analytics?

    Read the article

  • The most efficent ways for drawing lines all day long with OpenGL

    - by nkint
    I'd like to put a computer screen that is running an OpenGL programs in a room. It has to run all day long (not in the night). I'd like to draw lines that are slowly fading in the background. The setting is simple: a uniform color background (say, black) and colored lines (say, white) that are slowly fading out. With slowly I mean.. hours. Say that the first line I draw is with alpha 255 (fully visible), after one hours is 240. After 10 hours is 105. One line could have 250 points and there will be like 300 line in one day. For now I have done a prototype with very rudimentary method like: glBegin( GL_LINE_STRIP ); iterator = point_list.begin(); for (++iterator, end = point_list.end(); iterator != end; ++iterator) { const Vec3D &v = *iterator; glVertex2f(v.x(), v.y()); } glEnd(); More efficient method?

    Read the article

  • Ubuntu 12.04 takes too long to boot

    - by msPeachy
    I've recently encountered the following error message: mount: mounting /dev/disk/by-uuid/3f7f5cd9d-6ea3-4da7-b5ec-**** on /root failed: Invalid argument mount: mounting /sys on /root/sys failed: No such file or directory mount: mounting /dev on /root/dev failed: No such file or directory mount: mounting /sys on /root/sys failed: No such file or directory mount: mounting /proc on /root/proc failed: No such file or directory Target file system doesn't have /sbin/init. No init found. Try passing init= bootarg. Busybox v1.18.5 (Ubuntu 1:1.18.5-1ubuntu4) built-in shell (ash) Enter 'help' for a list of built-in commands. (initramfs) _ I run sudo fsck /dev/sda2 which is the Ubuntu ext4 root partition via LiveCD. It checked and fixed the file system. The next time I boot, Ubuntu started to load with the Ubuntu logo and the dots underneath for several hours (with the mouse pointer active on the screen), I even let the computer on overnight but still it did not successfully boot or got to the login screen in the morning. I booted again with the LiveCD and checked the NTFS partitions with ntfsfix and again the NTFS partitions was checked and fixed successfully. I also edited my fstab and commented out the lines that auto-mounts the NTFS partitions. The next time I boot, it took almost 20 minutes for Ubuntu to get to the login screen, after typing the password it took an additional 10 minutes for Ubuntu to get to the desktop. On the desktop, it take several minutes to open any program, displaying the Dash alone takes 5 minutes! Is there a fix for this without having to reinstall Ubuntu? I don't see or get any errors, Ubuntu is just taking too long to boot and to run programs. Please help!

    Read the article

  • CI - How long is continous?

    - by Andy
    We currently are using CCNet as our continous integration server. Most projects check for changes every 30 seconds (the default) and if needed perform a build (unit tests, stylecop, fxcop, etc). We've gotten quite a few projects now, and the server spends most of its time near 100% cpu utilization. This has alarmed some of the development team, even though the server is responsive and builds are still about the same length of time they've always been. Its been suggested that we lower the check interval to about five minutes. To me that seems too long, and we risk people committing code and then going home for the weekend and now there's a broken build possibly holding up others. In response, the suggestion is that if someone needs to know the results they can force the build. But that seems to defeat the purpose of CI, as I thought it was supposed to be automated. My proposed solution is just to get another build server and split the builds amongst the servers. Am I thinking about this the wrong way, or is there a point where if integration isn't often enough you're not really doing CI anymore?

    Read the article

  • what's wrong with mysql to create very long table fields?

    - by peng
    mysql> create table balance_sheet( -> Cash_and_cash_equivalents VARCHAR(20), -> Trading_financial_assets VARCHAR(20), -> Note_receivable VARCHAR(20), -> Account_receivable VARCHAR(20), -> Advance_money VARCHAR(20), -> Interest_receivable VARCHAR(20), -> Dividend_receivable VARCHAR(20), -> Other_notes_receivable VARCHAR(20), -> Due_from_related_parties VARCHAR(20), -> Inventory VARCHAR(20), -> Consumptive_biological_assets VARCHAR(20), -> Non_current_asset(expire_in_a_year) VARCHAR(20), -> Other_current_assets VARCHAR(20), -> Total_current_assets VARCHAR(20), -> Available_for_sale_financial_assets VARCHAR(20), -> Held_to_maturity_investment VARCHAR(20), -> Long_term_account_receivable VARCHAR(20), -> Long_term_equity_investment VARCHAR(20), -> Investment_real_eastate VARCHAR(20), -> Fixed_assets VARCHAR(20), -> Construction_in_progress VARCHAR(20), -> Project_material VARCHAR(20), -> Liquidation_of_fixed_assets VARCHAR(20), -> Capitalized_biological_assets VARCHAR(20), -> Oil_and_gas_assets VARCHAR(20), -> Intangible_assets VARCHAR(20), -> R&d_expense VARCHAR(20), -> Goodwill VARCHAR(20), -> Deferred_assets VARCHAR(20), -> Deferred_income_tax_assets VARCHAR(20), -> Other_non_current_assets VARCHAR(20), -> Total_non_current_assets VARCHAR(20), -> Total_assets VARCHAR(20), -> Short_term_borrowing VARCHAR(20), -> Transaction_financial_liabilities VARCHAR(20), -> Notes_payable VARCHAR(20), -> Account_payable VARCHAR(20), -> Item_received_in_advance VARCHAR(20), -> Employee_pay_payable VARCHAR(20), -> Tax_payable VARCHAR(20), -> Interest_payable VARCHAR(20), -> Dividend_payable VARCHAR(20), -> Other_account_payable VARCHAR(20), -> Due_to_related_parties VARCHAR(20), -> Non_current_liabilities_due_within_one_year VARCHAR(20), -> Other_current_liabilities VARCHAR(20), -> Total_current_liabilities VARCHAR(20), -> Long_term_loan VARCHAR(20), -> Bonds_payable VARCHAR(20), -> Long_term_payable VARCHAR(20), -> Specific_payable VARCHAR(20), -> Estimated_liabilities VARCHAR(20), -> Deferred_income_tax_liabilities VARCHAR(20), -> Other_non_current_liabilities VARCHAR(20), -> Total_non_current_liabilities VARCHAR(20), -> Total_liabilities VARCHAR(20), -> Paid_in_capital VARCHAR(20), -> Contributed_surplus VARCHAR(20), -> Treasury_stock VARCHAR(20), -> Earned_surplus VARCHAR(20), -> Retained_earnings VARCHAR(20), -> Translation_reserve VARCHAR(20), -> Nonrecurring_items VARCHAR(20), -> Total_equity(non) VARCHAR(20), -> Minority_interests VARCHAR(20), -> Total_equity VARCHAR(20), -> Total_liabilities_&_shareholder's_equity VARCHAR(20)); '> '> when i press enter,there is the output of ' ,no other reaction ,what's wrong?

    Read the article

  • Why is my long polling code for a notification system not updating in real time? PHP MYSQL

    - by tjones
    I am making a notification system similar to the red notification on facebook. It should update the number of messages sent to a user in real time. When the message MYSQL table is updated, it should instantly notify the user, but it does not. There does not seem to be an error inserting into MYSQL because on page refresh the notifications update just fine. I am essentially using code from this video tutorial: http://www.screenr.com/SNH (which updates in realtime if a data.txt file is changed, but it is not written for MYSQL like I am trying to do) Is there something wrong with the below code: **Javascript** <script type="text/javascript"> $(document).ready(function(){ var timestamp = null; function waitForMsg(){ $.ajax({ type: "GET", url: "getData.php", data: "userid=" + userid, async: true, cache: false, success: function(data){ var json = eval('(' + data + ')'); if (json['msg'] != "") { $('.notification').fadeIn().html(json['msg']); } setTimeout('waitForMsg()',30000); }, error: function(XMLHttpRequest, textStatus, errorThrown){ setTimeout('waitForMsg()',30000); } }); } waitForMsg(); </script> <body> <div class="notification"></div> **PHP*** <?php if ($_SERVER['REQUEST_METHOD'] == 'GET' ) { $userid = $_GET['userid']; include("config.php"); $sql="SELECT MAX(time) FROM notification WHERE userid='$userid'"; $result = mysql_query($sql); $row = mysql_fetch_assoc($result); $currentmodif = $row['MAX(time)']; $s="SELECT MAX(lasttimeread) FROM notificationsRead WHERE submittedby='$userid'"; $r = mysql_query($s); $rows = mysql_fetch_assoc($r); $lasttimeread = $rows['MAX(lasttimeread)']; while ($currentmodif <= $lasttimeread) { usleep(10000); clearstatcache(); $currentmodif = $row['MAX(time)']; } $response = array(); $response['msg'] = You have new messages; echo json_encode($response); } ?>

    Read the article

  • Taking too long to get skills for entry level programmer position [closed]

    - by greenonion
    I don't have the skills for an entry level position as a .Net programmer. I am trying to learn what I need but there is too much to learn and too little time. What can I do? About two months ago, I went to a job interview for an entry level C# .Net programming/consultant position in NYC. When I heard back from them, they told me that the knowledge gap between what I knew and what they needed me to know was too big and I might have been a better fit if I had 6 months of experience. This was the first interview that I went on since graduating college. before the interview, I read a book on visual C#. Turns out it wasn't a very good book and I was missing a lot of key areas of knowledge such as ADO.net SQL (I had learned some LINQ) A little bit about how memory is handled Multiple threaded programming, etc. Because the book wasn't very good, the stuff I did know, I didn't know very well. I felt crushed. I've applied for jobs to gain experience but when recruiters hear that I have no experience they lose interest. I figured that I can at least work on my knowledge. Since then, I read "SQL Essentials" to cover the SQL bit and I found a pretty awesome book that is good enough to clear up what's hazy in my mind and covers almost all of the extra topics. The book is "C# 4.0: The Complete Reference" by Herbert Schildt. I'm even learning a lot about the topics I was familiar with. For a month now I've been working my way through this beast of a book. However, gaining the knowledge I need is taking too long. I can't hold off not having a full-time job much longer. I'm not stupid and I'm studying constantly pouring through the book, asking questions on stackoverflow, referencing the C# specification, etc. I have made great progress but there is just too much ground to cover. I'm on chapter 12 which is about a 3rd through the book. To get an idea of what I know vs don't know, the table of contents is on amazon: http://www.amazon.com/C-4-0-The-Complete-Reference/dp/007174116X How on earth can someone know enough to function as a programmer in the real world? Can I try for a job in academia? Will I have time to finish learning the rest of the C# language or am I just un-hireable?

    Read the article

  • Slow Ubuntu 10.04 after long time unused

    - by Winston Ewert
    I'm at spring break so I'm back at my parent's house. I've turned my computer on which has been off since January and its unusably slow. This was not the case when I last used the computer in January. It is running 10.04, Memory: 875.5 MB CPU: AMD Athlon 64 X2 Dual Core Processor 4400+ Available Disk Space: 330.8 GB I'm not seeing a large usage of either memory or Disk I/O. If I look at my list of processes there is only a very small amount of CPU usage. However, if I hover over the CPU usage graph that I've on the top bar, I sometimes get really high readings like 100%. It took a long time to boot, to open firefox, to open a link in firefox. As far as I can tell everything that the computer tries to do is just massively slow. Right now, I'm apt-get dist-upgrading to install any updates that I will have missed since last time this computer was on. Any ideas as to what is going on here? UPDATE: I thought to check dmesg and it has a lot of entries like this: [ 1870.142201] ata3.00: exception Emask 0x0 SAct 0x7 SErr 0x0 action 0x0 [ 1870.142206] ata3.00: irq_stat 0x40000008 [ 1870.142210] ata3.00: failed command: READ FPDMA QUEUED [ 1870.142217] ata3.00: cmd 60/08:10:c0:4a:65/00:00:03:00:00/40 tag 2 ncq 4096 in [ 1870.142218] res 41/40:00:c5:4a:65/00:00:03:00:00/40 Emask 0x409 (media error) <F> [ 1870.142221] ata3.00: status: { DRDY ERR } [ 1870.142223] ata3.00: error: { UNC } [ 1870.143981] ata3.00: SB600 AHCI: limiting to 255 sectors per cmd [ 1870.146758] ata3.00: SB600 AHCI: limiting to 255 sectors per cmd [ 1870.146761] ata3.00: configured for UDMA/133 [ 1870.146777] ata3: EH complete [ 1872.092269] ata3.00: exception Emask 0x0 SAct 0x7 SErr 0x0 action 0x0 [ 1872.092274] ata3.00: irq_stat 0x40000008 [ 1872.092278] ata3.00: failed command: READ FPDMA QUEUED [ 1872.092285] ata3.00: cmd 60/08:00:c0:4a:65/00:00:03:00:00/40 tag 0 ncq 4096 in [ 1872.092287] res 41/40:00:c5:4a:65/00:00:03:00:00/40 Emask 0x409 (media error) <F> [ 1872.092289] ata3.00: status: { DRDY ERR } [ 1872.092292] ata3.00: error: { UNC } [ 1872.094050] ata3.00: SB600 AHCI: limiting to 255 sectors per cmd [ 1872.096795] ata3.00: SB600 AHCI: limiting to 255 sectors per cmd [ 1872.096798] ata3.00: configured for UDMA/133 [ 1872.096814] ata3: EH complete [ 1874.042279] ata3.00: exception Emask 0x0 SAct 0x7 SErr 0x0 action 0x0 [ 1874.042285] ata3.00: irq_stat 0x40000008 [ 1874.042289] ata3.00: failed command: READ FPDMA QUEUED [ 1874.042296] ata3.00: cmd 60/08:10:c0:4a:65/00:00:03:00:00/40 tag 2 ncq 4096 in [ 1874.042297] res 41/40:00:c5:4a:65/00:00:03:00:00/40 Emask 0x409 (media error) <F> [ 1874.042300] ata3.00: status: { DRDY ERR } [ 1874.042302] ata3.00: error: { UNC } [ 1874.044048] ata3.00: SB600 AHCI: limiting to 255 sectors per cmd [ 1874.046837] ata3.00: SB600 AHCI: limiting to 255 sectors per cmd [ 1874.046840] ata3.00: configured for UDMA/133 [ 1874.046861] sd 2:0:0:0: [sda] Unhandled sense code [ 1874.046863] sd 2:0:0:0: [sda] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE [ 1874.046867] sd 2:0:0:0: [sda] Sense Key : Medium Error [current] [descriptor] [ 1874.046872] Descriptor sense data with sense descriptors (in hex): [ 1874.046874] 72 03 11 04 00 00 00 0c 00 0a 80 00 00 00 00 00 [ 1874.046883] 03 65 4a c5 [ 1874.046886] sd 2:0:0:0: [sda] Add. Sense: Unrecovered read error - auto reallocate failed [ 1874.046892] sd 2:0:0:0: [sda] CDB: Read(10): 28 00 03 65 4a c0 00 00 08 00 [ 1874.046900] end_request: I/O error, dev sda, sector 56969925 [ 1874.046920] ata3: EH complete I'm not certain, but that looks like my problem may be a failing hard drive. But the drive is less then a year old, it really shouldn't be failing now...

    Read the article

  • Expanding on requestaudit - Tracing who is doing what...and for how long

    - by Kyle Hatlestad
    One of the most helpful tracing sections in WebCenter Content (and one that is on by default) is the requestaudit tracing.  This tracing section summarizes the top service requests happening in the server along with how they are performing.  By default, it has 2 different rotations.  One happens every 2 minutes (listing up to 5 services) and another happens every 60 minutes (listing up to 20 services).  These traces provide the total time for all the requests against that service along with the number of requests and its average request time.  This information can provide a good start in possibly troubleshooting performance issues or tracking a particular issue.   >requestaudit/6 12.10 16:48:00.493 Audit Request Monitor !csMonitorTotalRequests,47,1,0.39009329676628113,0.21034042537212372,1>requestaudit/6 12.10 16:48:00.509 Audit Request Monitor Request Audit Report over the last 120 Seconds for server wcc-base_4444****requestaudit/6 12.10 16:48:00.509 Audit Request Monitor -Num Requests 47 Errors 1 Reqs/sec. 0.39009329676628113 Avg. Latency (secs) 0.21034042537212372 Max Thread Count 1requestaudit/6 12.10 16:48:00.509 Audit Request Monitor 1 Service FLD_BROWSE Total Elapsed Time (secs) 3.5320000648498535 Num requests 10 Num errors 0 Avg. Latency (secs) 0.3531999886035919 requestaudit/6 12.10 16:48:00.509 Audit Request Monitor 2 Service GET_SEARCH_RESULTS Total Elapsed Time (secs) 2.694999933242798 Num requests 6 Num errors 0 Avg. Latency (secs) 0.4491666555404663requestaudit/6 12.10 16:48:00.509 Audit Request Monitor 3 Service GET_DOC_PAGE Total Elapsed Time (secs) 1.8839999437332153 Num requests 5 Num errors 1 Avg. Latency (secs) 0.376800000667572requestaudit/6 12.10 16:48:00.509 Audit Request Monitor 4 Service DOC_INFO Total Elapsed Time (secs) 0.4620000123977661 Num requests 3 Num errors 0 Avg. Latency (secs) 0.15399999916553497requestaudit/6 12.10 16:48:00.509 Audit Request Monitor 5 Service GET_PERSONALIZED_JAVASCRIPT Total Elapsed Time (secs) 0.4099999964237213 Num requests 8 Num errors 0 Avg. Latency (secs) 0.051249999552965164requestaudit/6 12.10 16:48:00.509 Audit Request Monitor ****End Audit Report***** To change the default rotation or size of output, these can be set as configuration variables for the server: RequestAuditIntervalSeconds1 – Used for the shorter of the two summary intervals (default is 120 seconds)RequestAuditIntervalSeconds2 – Used for the longer of the two summary intervals (default is 3600 seconds)RequestAuditListDepth1 – Number of services listed for the first request audit summary interval (default is 5)RequestAuditListDepth2 – Number of services listed for the second request audit summary interval (default is 20) If you want to get more granular, you can enable 'Full Verbose Tracing' from the System Audit Information page and now you will get an audit entry for each and every service request.  >requestaudit/6 12.10 16:58:35.431 IdcServer-68 GET_USER_INFO [dUser=bob][StatusMessage=You are logged in as 'bob'.] 0.08765099942684174(secs) What's nice is it reports who executed the service and how long that particular request took.  In some cases, depending on the service, additional information will be added to the tracing relevant to that  service. >requestaudit/6 12.10 17:00:44.727 IdcServer-81 GET_SEARCH_RESULTS [dUser=bob][QueryText=%28+dDocType+%3cmatches%3e+%60Document%60+%29][StatusCode=0][StatusMessage=Success] 0.4696030020713806(secs) You can even go into more detail and insert any additional data into the tracing.  You simply need to add this configuration variable with a comma separated list of variables from local data to insert. RequestAuditAdditionalVerboseFieldsList=TotalRows,path In this case, for any search results, the number of items the user found is traced: >requestaudit/6 12.10 17:15:28.665 IdcServer-36 GET_SEARCH_RESULTS [TotalRows=224][dUser=bob][QueryText=%28+dDocType+%3cmatches%3e+%60Application%60+%29][Sta... I also recently ran into the case where services were being called from a client through RIDC.  All of the services were being executed as the same user, but they wanted to correlate the requests coming from the client to the ones being executed on the server.  So what we did was add a new field to the request audit list: RequestAuditAdditionalVerboseFieldsList=ClientToken And then in the RIDC client, ClientToken was added to the binder along with a unique value that could be traced for that request.  Now they had a way of tracing on both ends and identifying exactly which client request resulted in which request on the server.

    Read the article

  • The long road to bug-free software

    - by Tony Davis
    The past decade has seen a burgeoning interest in functional programming languages such as Haskell or, in the Microsoft world, F#. Though still on the periphery of mainstream programming, functional programming concepts are gradually seeping into the imperative C# language (for example, Lambda expressions have their root in functional programming). One of the more interesting concepts from functional programming languages is the use of formal methods, the lofty ideal behind which is bug-free software. The idea is that we write a specification that describes exactly how our function (say) should behave. We then prove that our function conforms to it, and in doing so have proved beyond any doubt that it is free from bugs. All programmers already use one form of specification, specifically their programming language's type system. If a value has a specific type then, in a type-safe language, the compiler guarantees that value cannot be an instance of a different type. Many extensions to existing type systems, such as generics in Java and .NET, extend the range of programs that can be type-checked. Unfortunately, type systems can only prevent some bugs. To take a classic problem of retrieving an index value from an array, since the type system doesn't specify the length of the array, the compiler has no way of knowing that a request for the "value of index 4" from an array of only two elements is "unsafe". We restore safety via exception handling, but the ideal type system will prevent us from doing anything that is unsafe in the first place and this is where we start to borrow ideas from a language such as Haskell, with its concept of "dependent types". If the type of an array includes its length, we can ensure that any index accesses into the array are valid. The problem is that we now need to carry around the length of arrays and the values of indices throughout our code so that it can be type-checked. In general, writing the specification to prove a positive property, even for a problem very amenable to specification, such as a simple sorting algorithm, turns out to be very hard and the specification will be different for every program. Extend this to writing a specification for, say, Microsoft Word and we can see that the specification would end up being no simpler, and therefore no less buggy, than the implementation. Fortunately, it is easier to write a specification that proves that a program doesn't have certain, specific and undesirable properties, such as infinite loops or accesses to the wrong bit of memory. If we can write the specifications to prove that a program is immune to such problems, we could reuse them in many places. The problem is the lack of specification "provers" that can do this without a lot of manual intervention (i.e. hints from the programmer). All this might feel a very long way off, but computing power and our understanding of the theory of "provers" advances quickly, and Microsoft is doing some of it already. Via their Terminator research project they have started to prove that their device drivers will always terminate, and in so doing have suddenly eliminated a vast range of possible bugs. This is a huge step forward from saying, "we've tested it lots and it seems fine". What do you think? What might be good targets for specification and verification? SQL could be one: the cost of a bug in SQL Server is quite high given how many important systems rely on it, so there's a good incentive to eliminate bugs, even at high initial cost. [Many thanks to Mike Williamson for guidance and useful conversations during the writing of this piece] Cheers, Tony.

    Read the article

  • The long road to bug-free software

    - by Tony Davis
    The past decade has seen a burgeoning interest in functional programming languages such as Haskell or, in the Microsoft world, F#. Though still on the periphery of mainstream programming, functional programming concepts are gradually seeping into the imperative C# language (for example, Lambda expressions have their root in functional programming). One of the more interesting concepts from functional programming languages is the use of formal methods, the lofty ideal behind which is bug-free software. The idea is that we write a specification that describes exactly how our function (say) should behave. We then prove that our function conforms to it, and in doing so have proved beyond any doubt that it is free from bugs. All programmers already use one form of specification, specifically their programming language's type system. If a value has a specific type then, in a type-safe language, the compiler guarantees that value cannot be an instance of a different type. Many extensions to existing type systems, such as generics in Java and .NET, extend the range of programs that can be type-checked. Unfortunately, type systems can only prevent some bugs. To take a classic problem of retrieving an index value from an array, since the type system doesn't specify the length of the array, the compiler has no way of knowing that a request for the "value of index 4" from an array of only two elements is "unsafe". We restore safety via exception handling, but the ideal type system will prevent us from doing anything that is unsafe in the first place and this is where we start to borrow ideas from a language such as Haskell, with its concept of "dependent types". If the type of an array includes its length, we can ensure that any index accesses into the array are valid. The problem is that we now need to carry around the length of arrays and the values of indices throughout our code so that it can be type-checked. In general, writing the specification to prove a positive property, even for a problem very amenable to specification, such as a simple sorting algorithm, turns out to be very hard and the specification will be different for every program. Extend this to writing a specification for, say, Microsoft Word and we can see that the specification would end up being no simpler, and therefore no less buggy, than the implementation. Fortunately, it is easier to write a specification that proves that a program doesn't have certain, specific and undesirable properties, such as infinite loops or accesses to the wrong bit of memory. If we can write the specifications to prove that a program is immune to such problems, we could reuse them in many places. The problem is the lack of specification "provers" that can do this without a lot of manual intervention (i.e. hints from the programmer). All this might feel a very long way off, but computing power and our understanding of the theory of "provers" advances quickly, and Microsoft is doing some of it already. Via their Terminator research project they have started to prove that their device drivers will always terminate, and in so doing have suddenly eliminated a vast range of possible bugs. This is a huge step forward from saying, "we've tested it lots and it seems fine". What do you think? What might be good targets for specification and verification? SQL could be one: the cost of a bug in SQL Server is quite high given how many important systems rely on it, so there's a good incentive to eliminate bugs, even at high initial cost. [Many thanks to Mike Williamson for guidance and useful conversations during the writing of this piece] Cheers, Tony.

    Read the article

  • Basic data alignment question

    - by Broken Logic
    I've been playing around to see how my computer works under the hood. What I'm interested in is seeing is what happens on the stack inside a function. To do this I've written the following toy program: #include <stdio.h> void __cdecl Test1(char a, unsigned long long b, char c) { char c1; unsigned long long b1; char a1; c1 = 'b'; b1 = 4; a1 = 'r'; printf("%d %d - %d - %d %d Total: %d\n", (long)&b1 - (long)&a1, (long)&c1 - (long)&b1, (long)&a - (long)&c1, (long)&b - (long)&a, (long)&c - (long)&b, (long)&c - (long)&a1 ); }; struct TestStruct { char a; unsigned long long b; char c; }; void __cdecl Test2(char a, unsigned long long b, char c) { TestStruct locals; locals.a = 'b'; locals.b = 4; locals.c = 'r'; printf("%d %d - %d - %d %d Total: %d\n", (long)&locals.b - (long)&locals.a, (long)&locals.c - (long)&locals.b, (long)&a - (long)&locals.c, (long)&b - (long)&a, (long)&c - (long)&b, (long)&c - (long)&locals.a ); }; int main() { Test1('f', 0, 'o'); Test2('f', 0, 'o'); return 0; } And this spits out the following: 9 19 - 13 - 4 8 Total: 53 8 8 - 24 - 4 8 Total: 52 The function args are well behaved but as the calling convention is specified, I'd expect this. But the local variables are a bit wonky. My question is, why wouldn't these be the same? The second call seems to produce a more compact and better aligned stack. Looking at the ASM is unenlightening (at least to me), as the variable addresses are still aliased there. So I guess this is really a question about the assembler itself allocates the stack to local variables. I realise that any specific answer is likely to be platform specific. I'm more interested in a general explanation unless this quirk really is platform specific. For the record though, I'm compiling with VS2010 on a 64bit Intel machine.

    Read the article

  • Real-world SignalR example, ditching ghetto long polling

    - by Jeff
    One of the highlights of BUILD last week was the announcement that SignalR, a framework for real-time client to server (or cloud, if you will) communication, would be a real supported thing now with the weight of Microsoft behind it. Love the open source flava! If you aren’t familiar with SignalR, watch this BUILD session with PM Damian Edwards and dev David Fowler. Go ahead, I’ll wait. You’ll be in a happy place within the first ten minutes. If you skip to the end, you’ll see that they plan to ship this as a real first version by the end of the year. Insert slow clap here. Writing a few lines of code to move around a box from one browser to the next is a way cool demo, but how about something real-world? When learning new things, I find it difficult to be abstract, and I like real stuff. So I thought about what was in my tool box and the decided to port my crappy long-polling “there are new posts” feature of POP Forums to use SignalR. A few versions back, I added a feature where a button would light up while you were pecking out a reply if someone else made a post in the interim. It kind of saves you from that awkward moment where someone else posts some snark before you. While I was proud of the feature, I hated the implementation. When you clicked the reply button, it started polling an MVC URL asking if the last post you had matched the last one the server, and it did it every second and a half until you either replied or the server told you there was a new post, at which point it would display that button. The code was not glam: // in the reply setup PopForums.replyInterval = setInterval("PopForums.pollForNewPosts(" + topicID + ")", 1500); // called from the reply setup and the handler that fetches more posts PopForums.pollForNewPosts = function (topicID) { $.ajax({ url: PopForums.areaPath + "/Forum/IsLastPostInTopic/" + topicID, type: "GET", dataType: "text", data: "lastPostID=" + PopForums.currentTopicState.lastVisiblePost, success: function (result) { var lastPostLoaded = result.toLowerCase() == "true"; if (lastPostLoaded) { $("#MorePostsBeforeReplyButton").css("visibility", "hidden"); } else { $("#MorePostsBeforeReplyButton").css("visibility", "visible"); clearInterval(PopForums.replyInterval); } }, error: function () { } }); }; What’s going on here is the creation of an interval timer to keep calling the server and bugging it about new posts, and setting the visibility of a button appropriately. It looks like this if you’re monitoring requests in FireBug: Gross. The SignalR approach was to call a message broker when a reply was made, and have that broker call back to the listening clients, via a SingalR hub, to let them know about the new post. It seemed weird at first, but the server-side hub’s only method is to add the caller to a group, so new post notifications only go to callers viewing the topic where a new post was made. Beyond that, it’s important to remember that the hub is also the means to calling methods at the client end. Starting at the server side, here’s the hub: using Microsoft.AspNet.SignalR.Hubs; namespace PopForums.Messaging { public class Topics : Hub { public void ListenTo(int topicID) { Groups.Add(Context.ConnectionId, topicID.ToString()); } } } Have I mentioned how awesomely not complicated this is? The hub acts as the channel between the server and the client, and you’ll see how JavaScript calls the above method in a moment. Next, the broker class and its associated interface: using Microsoft.AspNet.SignalR; using Topic = PopForums.Models.Topic; namespace PopForums.Messaging { public interface IBroker { void NotifyNewPosts(Topic topic, int lasPostID); } public class Broker : IBroker { public void NotifyNewPosts(Topic topic, int lasPostID) { var context = GlobalHost.ConnectionManager.GetHubContext<Topics>(); context.Clients.Group(topic.TopicID.ToString()).notifyNewPosts(lasPostID); } } } The NotifyNewPosts method uses the static GlobalHost.ConnectionManager.GetHubContext<Topics>() method to get a reference to the hub, and then makes a call to clients in the group matched by the topic ID. It’s calling the notifyNewPosts method on the client. The TopicService class, which handles the reply data from the MVC controller, has an instance of the broker new’d up by dependency injection, so it took literally one line of code in the reply action method to get things moving. _broker.NotifyNewPosts(topic, post.PostID); The JavaScript side of things wasn’t much harder. When you click the reply button (or quote button), the reply window opens up and fires up a connection to the hub: var hub = $.connection.topics; hub.client.notifyNewPosts = function (lastPostID) { PopForums.setReplyMorePosts(lastPostID); }; $.connection.hub.start().done(function () { hub.server.listenTo(topicID); }); The important part to look at here is the creation of the notifyNewPosts function. That’s the method that is called from the server in the Broker class above. Conversely, once the connection is done, the script calls the listenTo method on the server, letting it know that this particular connection is listening for new posts on this specific topic ID. This whole experiment enables a lot of ideas that would make the forum more Facebook-like, letting you know when stuff is going on around you.

    Read the article

  • Short Season, Long Models - Dealing with Seasonality

    - by Michel Adar
    Accounting for seasonality presents a challenge for the accurate prediction of events. Examples of seasonality include: ·         Boxed cosmetics sets are more popular during Christmas. They sell at other times of the year, but they rise higher than other products during the holiday season. ·         Interest in a promotion rises around the time advertising on TV airs ·         Interest in the Sports section of a newspaper rises when there is a big football match There are several ways of dealing with seasonality in predictions. Time Windows If the length of the model time windows is short enough relative to the seasonality effect, then the models will see only seasonal data, and therefore will be accurate in their predictions. For example, a model with a weekly time window may be quick enough to adapt during the holiday season. In order for time windows to be useful in dealing with seasonality it is necessary that: The time window is significantly shorter than the season changes There is enough volume of data in the short time windows to produce an accurate model An additional issue to consider is that sometimes the season may have an abrupt end, for example the day after Christmas. Input Data If available, it is possible to include the seasonality effect in the input data for the model. For example the customer record may include a list of all the promotions advertised in the area of residence. A model with these inputs will have to learn the effect of the input. It is possible to learn it specific to the promotion – and by the way learn about inter-promotion cross feeding – by leaving the list of ads as it is; or it is possible to learn the general effect by having a flag that indicates if the promotion is being advertised. For inputs to properly represent the effect in the model it is necessary that: The model sees enough events with the input present. For example, by virtue of the model lifetime (or time window) being long enough to see several “seasons” or by having enough volume for the model to learn seasonality quickly. Proportional Frequency If we create a model that ignores seasonality it is possible to use that model to predict how the specific person likelihood differs from average. If we have a divergence from average then we can transfer that divergence proportionally to the observed frequency at the time of the prediction. Definitions: Ft = trailing average frequency of the event at time “t”. The average is done over a suitable period of to achieve a statistical significant estimate. F = average frequency as seen by the model. L = likelihood predicted by the model for a specific person Lt = predicted likelihood proportionally scaled for time “t”. If the model is good at predicting deviation from average, and this holds over the interesting range of seasons, then we can estimate Lt as: Lt = L * (Ft / F) Considering that: L = (L – F) + F Substituting we get: Lt = [(L – F) + F] * (Ft / F) Which simplifies to: (i)                  Lt = (L – F) * (Ft / F)  +  Ft This latest expression can be understood as “The adjusted likelihood at time t is the average likelihood at time t plus the effect from the model, which is calculated as the difference from average time the proportion of frequencies”. The formula above assumes a linear translation of the proportion. It is possible to generalize the formula using a factor which we will call “a” as follows: (ii)                Lt = (L – F) * (Ft / F) * a  +  Ft It is also possible to use a formula that does not scale the difference, like: (iii)               Lt = (L – F) * a  +  Ft While these formulas seem reasonable, they should be taken as hypothesis to be proven with empirical data. A theoretical analysis provides the following insights: The Cumulative Gains Chart (lift) should stay the same, as at any given time the order of the likelihood for different customers is preserved If F is equal to Ft then the formula reverts to “L” If (Ft = 0) then Lt in (i) and (ii) is 0 It is possible for Lt to be above 1. If it is desired to avoid going over 1, for relatively high base frequencies it is possible to use a relative interpretation of the multiplicative factor. For example, if we say that Y is twice as likely as X, then we can interpret this sentence as: If X is 3%, then Y is 6% If X is 11%, then Y is 22% If X is 70%, then Y is 85% - in this case we interpret “twice as likely” as “half as likely to not happen” Applying this reasoning to (i) for example we would get: If (L < F) or (Ft < (1 / ((L/F) + 1)) Then  Lt = L * (Ft / F) Else Lt = 1 – (F / L) + (Ft * F / L)  

    Read the article

  • storing and retrieving socket

    - by Trevor Newhook
    From what I can understand, once I create a socket, I can then create an array to store it with userArray[socket.nickname]=socket; I can then send a message to it with: io.sockets.socket(userArray[data.to]).emit('private message', tstamp(), socket.nickname, message); The basic logic is to store a copy of each socket in an object, identified by nickname. When I want to send a message to that socket, I use the copy of the socket, and send the message via io.sockets.socket(id).emit(). The entire server code is below: io.sockets.on('connection', function (socket) { socket.on('user message', function (msg) { socket.broadcast.emit('user message', tstamp(), socket.nickname, msg); updateLog('user message', socket.nickname, msg); }); socket.on('private message', function(data) { socket.get(data.nickname, function (err, name) { console.log('Chat message by ', name); }); updateLog('private message', socket.nickname, data.message); message=data.message; io.sockets.socket(userArray[data.to]).emit('private message', tstamp(), socket.nickname, message); }); socket.on('get log', function () { updateLog(); // Ensure old entries are cleared out before sending it. io.sockets.emit('chat log', log); }); socket.on('nickname', function (nick, fn) { var i = 1; var orignick = nick; while (nicknames[nick]) { nick = orignick+i; i++; } fn(nick); nicknames[nick] = socket.nickname = nick; userArray[socket.nickname]=socket; socket.set('nickname', nick, function () { socket.emit('ready'); }); socket.broadcast.emit('announcement', nick + ' connected'); // io.sockets.socket(userArray[nick]).emit('newID', 'Your name is: ' + nick, '. Your ID is: '+ userArray[nick]); io.sockets.emit('nicknames', nicknames); });

    Read the article

  • Iteration speed of int vs long

    - by jqno
    I have the following two programs: long startTime = System.currentTimeMillis(); for (int i = 0; i < N; i++); long endTime = System.currentTimeMillis(); System.out.println("Elapsed time: " + (endTime - startTime) + " msecs"); and long startTime = System.currentTimeMillis(); for (long i = 0; i < N; i++); long endTime = System.currentTimeMillis(); System.out.println("Elapsed time: " + (endTime - startTime) + " msecs"); Note: the only difference is the type of the loop variable (int and long). When I run this, the first program consistently prints between 0 and 16 msecs, regardless of the value of N. The second takes a lot longer. For N == Integer.MAX_VALUE, it runs in about 1800 msecs on my machine. The run time appears to be more or less linear in N. So why is this? I suppose the JIT-compiler optimizes the int loop to death. And for good reason, because obviously it doesn't do anything. But why doesn't it do so for the long loop as well? A colleague thought we might be measuring the JIT compiler doing its work in the long loop, but since the run time seems to be linear in N, this probably isn't the case.

    Read the article

  • How to make Symbolicate iPhone App Crash Reports

    - by bluej3
    Hello~ I retrieved the crash reports from iTunes Connect. I referenced this site. http://webcache.googleusercontent.com/search?q=cache:MmxwdXObZLMJ:www.anoshkin.net/blog/2008/09/09/iphone-crash-logs/+iphone+crash+debig&cd=2&hl=en&ct=clnk I tried.... $ symbolicatecrash report.crash MobileLines.app.dSYM report-with-symbols.crash Error in symbol file for /Developer/Platforms/iPhoneOS.platform/DeviceSupport/3.1.2 (7D11)/Symbols/System/Library/Frameworks/IOKit.framework/Versions/A/IOKit Error in symbol file for /Developer/Platforms/iPhoneOS.platform/DeviceSupport/3.1.2 (7D11)/Symbols/System/Library/PrivateFrameworks/WebCore.framework/WebCore Error in symbol file for /Developer/Platforms/iPhoneOS.platform/DeviceSupport/3.1.2 (7D11)/Symbols/System/Library/Frameworks/Foundation.framework/Foundation Error in symbol file for /Developer/Platforms/iPhoneOS.platform/DeviceSupport/3.1.2 (7D11)/Symbols/usr/lib/libSystem.B.dylib Error in symbol file for /Developer/Platforms/iPhoneOS.platform/DeviceSupport/3.1.2 (7D11)/Symbols/System/Library/PrivateFrameworks/GraphicsServices.framework/GraphicsServices Error in symbol file for /Developer/Platforms/iPhoneOS.platform/DeviceSupport/3.1.2 (7D11)/Symbols/System/Library/Frameworks/UIKit.framework/UIKit Error in symbol file for /Developer/Platforms/iPhoneOS.platform/DeviceSupport/3.1.2 (7D11)/Symbols/System/Library/Frameworks/OpenGLES.framework/MBXGLEngine.bundle/MBXGLEngine Error in symbol file for /Developer/Platforms/iPhoneOS.platform/DeviceSupport/3.1.2 (7D11)/Symbols/System/Library/Frameworks/AudioToolbox.framework/AudioToolbox Error in symbol file for /Developer/Platforms/iPhoneOS.platform/DeviceSupport/3.1.2 (7D11)/Symbols/System/Library/Frameworks/CoreFoundation.framework/CoreFoundation BUT... I didn't result. (find error message) - This directory is located "bulid/Distribution-iphones" - "MYGAME.app" file and "MYGAME.app.dSYM" file is located in same directory. How can i do solve this problem. ? Please help me :) * Crash log (carsh at thread 2 ) Incident Identifier: 95230C2E-CD83-46BF-8DAE-F38BCD46B910 Process: MYGAMELite [303] Path: /var/mobile/Applications/4FB79BEC-2BF0-438B-82A8-C302CD52A85C/MYGAMELite.app/MYGAMELite Identifier: MYGAMELite Version: ??? (???) Code Type: ARM (Native) Parent Process: launchd [1] Date/Time: 2010-06-03 11:43:52.875 +0800 OS Version: iPhone OS 3.1.2 (7D11) Report Version: 104 Exception Type: EXC_BAD_ACCESS (SIGSEGV) Exception Codes: KERN_INVALID_ADDRESS at 0x03e3a002 Crashed Thread: 2 Thread 2 Crashed: 0 AudioToolbox 0x330d708c AU3DMixerEmbedded::SumInput16(unsigned long, AudioBufferList const&, AudioBufferList const&, unsigned long, float, unsigned long) 1 AudioToolbox 0x330d89a0 AU3DMixerEmbedded::Render(unsigned long&, AudioTimeStamp const&, unsigned long) 2 AudioToolbox 0x32fe6bb8 AUBase::DoRender(unsigned long&, AudioTimeStamp const&, unsigned long, unsigned long, AudioBufferList&) 3 AudioToolbox 0x32fe6504 Render 4 AudioToolbox 0x330160b8 AUInputElement::PullInput(unsigned long&, AudioTimeStamp const&, unsigned long, unsigned long) 5 AudioToolbox 0x33023fa8 AUInputFormatConverter2::InputProc(OpaqueAudioConverter*, unsigned long*, AudioBufferList*, AudioStreamPacketDescription*, void) 6 AudioToolbox 0x32fe4b60 AudioConverterChain::CallInputProc(unsigned long) 7 AudioToolbox 0x32fe4a5c AudioConverterChain::FillBufferFromInputProc(unsigned long*, CABufferList*) 8 AudioToolbox 0x32fe4790 BufferedAudioConverter::GetInputBytes(unsigned long, unsigned long&, CABufferList const*&) 9 AudioToolbox 0x33023e30 CBRConverter::RenderOutput(CABufferList*, unsigned long, unsigned long&, AudioStreamPacketDescription*) 10 AudioToolbox 0x32fe4284 BufferedAudioConverter::FillBuffer(unsigned long&, AudioBufferList&, AudioStreamPacketDescription*) 11 AudioToolbox 0x32fe44a4 AudioConverterChain::RenderOutput(CABufferList*, unsigned long, unsigned long&, AudioStreamPacketDescription*) 12 AudioToolbox 0x32fe4284 BufferedAudioConverter::FillBuffer(unsigned long&, AudioBufferList&, AudioStreamPacketDescription*) 13 AudioToolbox 0x32fe3f10 AudioConverterFillComplexBuffer 14 AudioToolbox 0x33023844 AUConverterBase::RenderBus(unsigned long&, AudioTimeStamp const&, unsigned long, unsigned long) 15 AudioToolbox 0x330ce928 AURemoteIO::RenderBus(unsigned long&, AudioTimeStamp const&, unsigned long, unsigned long) 16 AudioToolbox 0x32fe6bb8 AUBase::DoRender(unsigned long&, AudioTimeStamp const&, unsigned long, unsigned long, AudioBufferList&) 17 AudioToolbox 0x330cf308 AURemoteIO::PerformIO(int, unsigned int, unsigned int, AQTimeStamp const&, AQTimeStamp const&) 18 AudioToolbox 0x330cf4cc AURIOCallbackReceiver_PerformIOSync 19 AudioToolbox 0x330c76fc _XPerformIOSync 20 AudioToolbox 0x330181d8 mshMIGPerform 21 AudioToolbox 0x3309cec8 MSHMIGDispatchMessage 22 AudioToolbox 0x330d48d4 AURemoteIO::IOThread::Entry(void*) 23 AudioToolbox 0x32fc9f20 CAPThread::Entry(CAPThread*) 24 libSystem.B.dylib 0x30b5b7b0 _pthread_body

    Read the article

< Previous Page | 3 4 5 6 7 8 9 10 11 12 13 14  | Next Page >