Search Results

Search found 18096 results on 724 pages for 'let'.

Page 485/724 | < Previous Page | 481 482 483 484 485 486 487 488 489 490 491 492  | Next Page >

  • Music Notation Editor - Refactoring view creation logic elseware

    - by Cyril Silverman
    Let me preface by saying that knowing some elementary music theory and music notation may be helpful in grasping the problem at hand. I'm currently building a Music Notation and Tablature Editor (in Javascript). But I've come to a point where the core parts of the program are more or less there. All functionality I plan to add at this point will really build off the foundation that I've created. As a result, I want to refactor to really solidify my code. I'm using an API called VexFlow to render notation. Basically I pass the parts of the editor's state to VexFlow to build the graphical representation of the score. Here is a rough and stripped down UML diagram showing you the outline of my program: In essence, a Part has many Measures which has many Notes which has many NoteItems (yes, this is semantically weird, as a chord is represented as a Note with multiple NoteItems, individual pitches or fret positions). All of the relationships are bi-directional. There are a few problems with my design because my Measure class contains the majority of the entire application view logic. The class holds the data about all VexFlow objects (the graphical representation of the score). It contains the graphical Staff object and the graphical notes. (Shouldn't these be placed somewhere else in the program?) While VexFlowFactory deals with actual creation (and some processing) of most of the VexFlow objects, Measure still "directs" the creation of all the objects and what order they are supposed to be created in for both the VexFlowStaff and VexFlowNotes. I'm not looking for a specific answer as you'd need a much deeper understanding of my code. Just a general direction to go in. Here's a thought I had, create an MeasureView/NoteView/PartView classes that contains the basic VexFlow objects for each class in addition to any extraneous logic for it's creation? but where would these views be contained? Do I create a ScoreView that is a parallel graphical representation of everything? So that ScoreView.render() would cascade down PartView and call render for each PartView and casade down into each MeasureView, etc. Again, I just have no idea what direction to go in. The more I think about it, the more ways to go seem to pop into my head. I tried to be as concise and simplistic as possible while still getting my problem across. Please feel free to ask me any questions if anything is unclear. It's quite a struggle trying to dumb down a complicated problem to its core parts.

    Read the article

  • *Un*installing with Ubuntu Software Center (Centre) doesn't work on 64-bit 12.04.1

    - by likethesky
    Not sure if I'm doing something wrong, or if the .deb package I'm installing is broken in some way (I've built it, using NetBeans 7.2), or if indeed this is a bug in Software Center. When I install this particular 32-bit .deb on Ubuntu 10.04 LTS--all updates applied--(where it was built), GDebi shows it and has an 'Uninstall' button next to it. So it works fine to uninstall it there, via the GDebi GUI. However, when I install it on 12.04.1 LTS--all updates applied--it installs fine, but then does not show up in Ubuntu Software Center as available to be uninstalled. No combination of searching finds it. However, I can from the command line, do sudo apt-get purge javafxapplication1 and it finds it and deletes it. The same thing happens when I build a 64-bit .deb and attempt to install it to the same (64-bit AMD) or a different 64-bit Ubuntu 12.04.1 system. So it seems to be isolated to this NetBeans-generated .deb and the 64-bit AMD build (though I haven't tried it on a 32-bit 12.04.1 install yet). These are all on VirtualBox VMs, btw, if that matters. Any way to 'clean up' my Software Center and see if it's something I've done to get it in this state? Could this behavior be due to how this particular .deb has been built? (It doesn't have an 'Installed-Size' control field, so I do get the "Package is of bad quality" warning when I install it--which I do by clicking 'Ignore and install' button.) If you want all the gory details about why this happening--a bug has been reported against NetBeans for this behavior here: http://javafx-jira.kenai.com/browse/RT-25486 (EDIT: Just to be clear, the app installs fine, runs fine, all works as intended--I just can't get that 'bad package' message to go away, and now... I also can't uninstall it via Software Center, but rather, need to use sudo apt-get purge to uninstall it, after it installs. /END EDIT) Thanks for any pointers. I'm happy to report this as a bug against Ubuntu Software Center/Centre too, if that's what it seems to be, just tell me where to do so (a link). I'm a relative Ubuntu, NetBeans, and JavaFX newbie, though a long-time programmer. If I report it as a bug, I'll try it on the 32-bit build of 12.04.1 as well. Also, if I should add any more detail to the bug reported against NetBeans above, let me know--or feel free to add it yourself to the bug report above, if you would like. Thanks again!

    Read the article

  • How to get tens of millions of pages indexed by Google bot?

    - by Chris Adragna
    We are currently developing a site that currently has 8 million unique pages that will grow to about 20 million right away, and eventually to about 50 million or more. Before you criticize... Yes, it provides unique, useful content. We continually process raw data from public records and by doing some data scrubbing, entity rollups, and relationship mapping, we've been able to generate quality content, developing a site that's quite useful and also unique, in part due to the breadth of the data. It's PR is 0 (new domain, no links), and we're getting spidered at a rate of about 500 pages per day, putting us at about 30,000 pages indexed thus far. At this rate, it would take over 400 years to index all of our data. I have two questions: Is the rate of the indexing directly correlated to PR, and by that I mean is it correlated enough that by purchasing an old domain with good PR will get us to a workable indexing rate (in the neighborhood of 100,000 pages per day). Are there any SEO consultants who specialize in aiding the indexing process itself. We're otherwise doing very well with SEO, on-page especially, besides, the competition for our "long-tail" keyword phrases is pretty low, so our success hinges mostly on the number of pages indexed. Our main competitor has achieved approx 20MM pages indexed in just over one year's time, along with an Alexa 2000-ish ranking. Noteworthy qualities we have in place: page download speed is pretty good (250-500 ms) no errors (no 404 or 500 errors when getting spidered) we use Google webmaster tools and login daily friendly URLs in place I'm afraid to submit sitemaps. Some SEO community postings suggest a new site with millions of pages and no PR is suspicious. There is a Google video of Matt Cutts speaking of a staged on-boarding of large sites, too, in order to avoid increased scrutiny (at approx 2:30 in the video). Clickable site links deliver all pages, no more than four pages deep and typically no more than 250(-ish) internal links on a page. Anchor text for internal links is logical and adds relevance hierarchically to the data on the detail pages. We had previously set the crawl rate to the highest on webmaster tools (only about a page every two seconds, max). I recently turned it back to "let Google decide" which is what is advised.

    Read the article

  • Test your internet connection - Emtel Mobile Internet

    After yesterday's report on Emtel Fixed Broadband (I'm still wondering where the 'fixed' part is), I did the same tests on Emtel Mobile Internet. For this I'm using the Huawei E169G HSDPA USB stick, connected to the same machine. Actually, this is my fail-safe internet connection and the system automatically switches between them if a problem, let's say timeout, etc. has been detected on the main line. For better comparison I used exactly the same servers on Speedtest.net. The results Following are the results of Rose Hill (hosted by Emtel) and respectively Frankfurt, Germany (hosted by Vodafone DE): Speedtest.net result of 31.05.2013 between Flic en Flac and Rose Hill, Mauritius (Emtel - Mobile Internet) Speedtest.net result of 31.05.2013 between Flic en Flac and Frankfurt, Germany (Emtel - Mobile Internet) As you might easily see, there is a big difference in speed between national and international connections. More interestingly are the results related to the download and upload ratio. I'm not sure whether connections over Emtel Mobile Internet are asymmetric or symmetric like the Fixed Broadband. Might be interesting to find out. The first test result actually might give us a clue that the connection could be asymmetric with a ratio of 3:1 but again I'm not sure. I'll find out and post an update on this. It depends on network coverage Later today I was on tour with my tablet, a Samsung Galaxy Tab 10.1 (model GT-P7500) running on Android 4.0.4 (Ice Cream Sandwich), and did some more tests using the Speedtest.net app. The results are actually as expected and in areas with better network coverage you will get better results after all. At least, as long as you stay inside the national networks. For anything abroad, it doesn't really matter. But see for yourselves: Speedtest.net result of 31.05.2013 between Cascavelle and servers in Rose Hill, Mauritius (Emtel - Mobile Internet), Port Louis, Mauritius and Kuala Lumpur, Malaysia It's rather shocking and frustrating to see how the speed on international destinations goes down. And the full capability of the tablet's integrated modem (HSDPA: 21 Mbps; HSUPA: 5.76 Mbps) isn't used, too. I guess, this demands more tests in other areas of the island, like Ebene, Pailles or Port Louis. I'll keep you updated... The question remains: Alternatives? After the publication of the test results on Fixed Broadband I had some exchange with others on Facebook. Sadly, it seems that there are really no alternatives to what Emtel is offering at the moment. There are the various internet packages by Mauritius Telecom feat. Orange, like ADSL, MyT and Mobile Internet, and there is Bharat Telecom with their Bees offer which is currently limited to Ebene and parts of Quatre Bornes.

    Read the article

  • Skyrim Creation Kit with Xbox 360

    - by funseiki
    I posted this on stackoverflow but was advised to post here (here is a link to the stackoverflow question). I'm hoping for constructive feedback on its plausibility. Update on progress: It looks like there are ways to stuff files back onto the console (horizon, modio, xplorer360, etc) and they do require some form of signing. As of now, though, I've had no luck. I was hoping I could get away with just placing the ".esp" into the directory containing marketplace downloads for Skyrim, along with the signed ".bsa" file (basically a zipped up file containing any extra content the .esp will need to refer that doesn't exist in the basic game). This doesn't work, at least not in the ways I've tried, so next I'm going to try install the entire game to my flash drive (if possible) and attempt to traverse through the game's directory (this is probably unlikely). If anyone else has suggestions or luck or wants more detail on my failures comment/answer away. Here is the question: I'm thinking about buying the PC version of Skyrim to get the Creation Kit (I already own a copy for the Xbox). I have read the faq and scoured plenty of forums to see if there was some way to mod Skyrim for a console (Xbox 360, in particular), but they are generally coming up negative. I realize the CreationKit is on the PC, but I was wondering if there was a way to set up the '.esp' (hopefully I'm interpreting this correctly) files to be placed on the Xbox 360 file system in a similar manner to how game add-ons are downloaded from the Xbox Live Marketplace. I believe it is possible to transfer saves between the console and the PC (e.g. google: 'skyrim mod xbox360'), but these are referencing items that already exist in the game (e.g. a console command for maximum carry weight does not require reference to new animations or models). It would probably be easier if one could navigate through the xbox's file system to see where the games' files are placed, but with the current setup, the file system is abstracted away. Any help or insight on the matter would be much appreciated. I would love to work on a project that would make it possible to let console gamers experience and enjoy all the great mods available to the PC community.

    Read the article

  • Customer won't decide, how to deal?

    - by Crazy Eddie
    I write software that involves the use of measured quantities, many input by the user, most displayed, that are fed into calculation models to simulate various physical thing-a-majigs. We have created a data type that allows us to associate a numeric value with a unit, we call these "quantities" (big duh). Quantities and units are unique to dimension. You can't attach kilogram to a length for example. Math on quantities does automatic unit conversion to SI and the type is dimension safe (you can't assign a weight to a pressure for example). Custom UI components have been developed that display the value and its unit and/or allow the user to edit them. Dimensionless quantities, having no units, are a single, custom case implemented within the system. There's a set of related quantities such that our target audience apparently uses them interchangeably. The quantities are used in special units that embed the conversion factors for the related quantity dimensions...in other words, when using these units converting from one to another simply involves multiplying the value by 1 to the dimensional difference. However, conversion to/from the calculation system (SI) still involves these factors. One of these related quantities is a dimensionless one that represents a ratio. I simply can't get the "customer" to recognize the necessity of distinguishing these values and their use. They've picked one and want to use it everywhere, customizing the way we deal with it in special places. In this case they've picked one of the dimensions that has a unit...BUT, they don't want there to be a unit (GRR!!!). This of course is causing us to implement these special overrides for our UI elements and such. That of course is often times forgotten and worse...after a couple months everyone forgets why it was necessary and why we're using this dimensional value, calling it the wrong thing, and disabling the unit. I could just ignore the "customer" and implement the type as the dimensionless quantity, which makes most sense. However, that leaves the team responsible for figuring it out when they've given us a formula using one of the other quantities. We have to not only figure out that it's happening, we have to decide what to do. This isn't a trivial deal. The other option is just to say to hell with it, do it the customer's way, and let it waste continued time and effort because it's just downright confusing as hell. However, I can't count the amount of times someone has said, "Why is this being done this way, it makes no sense at all," and the team goes off the deep end trying to figure it out. What would you do? Currently I'm still attempting to convince them that even if they use terms interchangeably, we at the least can't do that within the product discussion. Don't have high hopes though.

    Read the article

  • SQL SERVER – Backup SQL databases to Box or SkyDrive

    - by Pinal Dave
    To ensure your SQL Server or Azure databases remain safe, you should backup your databases periodically. And it is important to store the backups in a reliable location. Microsoft SkyDrive currently offers 7GB free, Box offers 5GB free – both are reliable and it is simple to send your backups there. SQLBackupAndFTP in it’s latest version 9 added the option to backup to SkyDrive and Box ( in addition to local/network folder, NAS drive, FTP, Dropbox, Google Drive and Amazon S3). Just select the databases that you’d like to backup and select to store the backups in SkyDrive or Box. Below I will show you how to do it in details Select databases to backup First connect to your SQL Server or Azure Sql Database. Then select the databases you’d like to backup. Connect to SkyDrive or Box cloud If you have a free version of SQLBackupAndFTP Box destination is included, but SkyDrive destination will be disabled as it is available in the Standard version or above. Click “Try now” to get 30 days trial on all options On the “SkyDrive Settings” form you’ll need to authorize SQLBackupAndFTP to access your SkyDrive. Click “Authorize…” to open SkyDrive authorization page in your browser, sign in your to SkyDrive account and click at “Allow” . On the next page you will see the field with authorization code. Copy it to the clipboard. Box operation is just the same. After that return to SQLBackupAndFTP, paste the authorization code and click “OK” . After you are authorized, you can enter the path to a backup folder. SQLBackupAndFTP will create the folder if it does not exist. That’s all what has to be done to backup to SkyDrive or Box cloud.  You can now click on “Run Now” button to test this job. Conclusion Whatever is your preference for storing SQL backups, it is easy with SQLBackupAndFTP. Note that at the time of this writing they are running a very rare promotion on volume licenses: 5–9 licenses: 20% off 10–19 licenses: 35% off more than 20 licenses: 50% off Please let me know your favorite options for storing the backups. Reference: Pinal Dave (http://blog.sqlauthority.com)Filed under: PostADay, SQL, SQL Authority, SQL Query, SQL Server, SQL Tips and Tricks, T SQL

    Read the article

  • Class instance clustering in object reference graph for multi-entries serialization

    - by Juh_
    My question is on the best way to cluster a graph of class instances (i.e. objects, the graph nodes) linked by object references (the -directed- edges of the graph) around specifically marked objects. To explain better my question, let me explain my motivation: I currently use a moderately complex system to serialize the data used in my projects: "marked" objects have a specific attributes which stores a "saving entry": the path to an associated file on disc (but it could be done for any storage type providing the suitable interface) Those object can then be serialized automatically (eg: obj.save()) The serialization of a marked object 'a' contains implicitly all objects 'b' for which 'a' has a reference to, directly s.t: a.b = b, or indirectly s.t.: a.c.b = b for some object 'c' This is very simple and basically define specific storage entries to specific objects. I have then "container" type objects that: can be serialized similarly (in fact their are or can-be "marked") they don't serialize in their storage entries the "marked" objects (with direct reference): if a and a.b are both marked, a.save() calls b.save() and stores a.b = storage_entry(b) So, if I serialize 'a', it will serialize automatically all objects that can be reached from 'a' through the object reference graph, possibly in multiples entries. That is what I want, and is usually provides the functionalities I need. However, it is very ad-hoc and there are some structural limitations to this approach: the multi-entry saving can only works through direct connections in "container" objects, and there are situations with undefined behavior such as if two "marked" objects 'a'and 'b' both have a reference to an unmarked object 'c'. In this case my system will stores 'c' in both 'a' and 'b' making an implicit copy which not only double the storage size, but also change the object reference graph after re-loading. I am thinking of generalizing the process. Apart for the practical questions on implementation (I am coding in python, and use Pickle to serialize my objects), there is a general question on the way to attach (cluster) unmarked objects to marked ones. So, my questions are: What are the important issues that should be considered? Basically why not just use any graph parsing algorithm with the "attach to last marked node" behavior. Is there any work done on this problem, practical or theoretical, that I should be aware of? Note: I added the tag graph-database because I think the answer might come from that fields, even if the question is not.

    Read the article

  • Count unique visitors by group of visited places

    - by Mathieu
    I'm facing the problem of counting the unique visitors of groups of places. Here is the situation: I have visitors that can visit places. For example, that can be internet users visiting web pages, or customers going to restaurants. A visitor can visit as much places as he wishes, and a place can be visited by several visitors. A visitor can come to the same place several times. The places belong to groups. A group can obviously contain several places, and places can belong to several groups. Given that, for each visitor, we can have a list of visited places, how can I have the number of unique visitors per group of places? Example: I have visitors A, B, C and D; and I have places x, y and z. I have these visiting lists: [ A -> [x,x,y,x], B -> [], C -> [z,z], D -> [y,x,x,z] ] Having these number of unique visitors per place is quite easy: [ x -> 2, // A and D visited x y -> 2, // A and D visited y z -> 2 // C and D visited z ] But if I have these groups: [ G1 -> [x,y,z], G2 -> [x,z], G3 -> [x,y] ] How can I have this information? [ G1 -> 3, // A, C and D visited x or y or z G2 -> 3, // A, C and D visited x or z G3 -> 2 // A and D visited x or y ] Additional notes : There are so many places that it is not possible to store information about every possible group; It's not a problem if approximation are made. I don't need 100% precision. Having a fast algorithm that tells me that there were 12345 visits in a group instead of 12543 is better than a slow algorithm telling the exact number. Let's say there can be ~5% deviation. Is there an algorithm or class of algorithms that addresses this type of problem?

    Read the article

  • The JDEdwards EnterpriseOne PreSales University

    - by Julien Haye
    Istanbul NOV 5-9 Wednesday, NOV 7 - It is raining outside and I am sitting in my hotel room (#106) in Istanbul and create my first blog entry. Today this blog was enabled and I am excited to have the ability to share my (first) thoughts with the EMEA JDE Partner Community. I am here in Istanbul because we are currently running the JDEdwards PreSales University Event series. This PreSales University is an established event series which we deliver the fifth time now and the first time in the ECEMEA region. Delegates value the openness and competence from the Product Strategy and Product Development Team from Denver and India. Together with the regional Oracle PreSales team we had very valuable discussions around product features and functions and about the business value of the new delivered applications and tools. Additionally the event provides endless opportunities to exchange ideas with other JD Edwards Partner and the Oracle PreSales Team. With its focus on sharing and learning, best practice, user experience and transforming technologies, delegates will leave this event with an abundance of new ideas and best practices to try for your coming projects and existing customer implementations. A day out of the office gives delegates a chance to gain a new perspective on their business processes. Everybody sees better ways of working just by being immersed in an environment where the focus is on using products more effectively. Apps Track: Highly concentrated participants in Istanbul listening to Jeff Erickson presenting the news about OneView Reporting. Jeff: We believe “The things you said”. The event is organized into two tracks, one for Apps and one for Tech. Everybody was able to learn new features and functions and how to position this products. The focus was on the new Apps release 9.1 and Tools Release 9.1.2 and their Value Propositions. For all topics hands-on exercises has been given to the participants. Even very experienced senior consultants did learn a lot from this event. In total we have 55 people registered and we still have some more content to deliver. By the way: Istanbul is a nice place to be. I already booked my next trip to this beautiful city. In two weeks we deliver the JD Edwards EECIS Executive Forum again in Istanbul. Once again a tough Agenda. I will let you know if I had the ability to have a walk outside and see a bit more of this beautiful city. At least I expect to have a different room number. Many greetings Hartmut WieseOracle Alliances & Channels EMEA

    Read the article

  • How do I get long command lines to wrap to the next line?

    - by BrianH
    Edit It was my .bashrc file. I've copied the same profile from machine to machine, and I used special characters in my $PS1 that are somehow throwing it off. I'm now sticking with the standard bash variables for my $PS1. Thanks to @ændrük for the tip on the .bashrc! ...End Edit... Something I have noticed in Ubuntu for a long time that has been frustrating to me is when I am typing a command at the command line that gets longer (wider) than the terminal width, instead of wrapping to a new line, it goes back to column 1 on the same line and starts over-writing the beginning of my command line. (It doesn't actually overwrite the actual command, but visually, it is overwriting the text that was displayed). It's hard to explain without seeing it, but let's say my terminal was 20 characters wide (Mine is more like 120 characters - but for the sake of an example), and I want to echo the English alphabet. What I type is this: echo abcdefghijklmnopqrstuvwxyz But what my terminal looks like before I hit the key is: pqrstuvwxyzghijklmno When I hit enter, it echos abcdefghijklmnopqrstuvwxyz so I know the command was received properly. It just wrapped my typing after the "o" and started over on the same line. What I would expect to happen, if I typed this command in on a terminal that was only 20 characters wide would be this: echo abcdefghijklmno pqrstuvwxyz Background: I am using bash as my shell, and I have this line in my ~/.bashrc: set -o vi to be able to navigate the command line with VI commands. I am currently using Ubuntu 10.10 server, and connecting to the server with Putty. In any other environment I have worked in, if I type a long command line, it will add a new line underneath the line I am working on when my command gets longer than the terminal width and when I keep typing I can see my command on 2 different lines. But for as long as I can remember using Ubuntu, my long commands only occupy 1 line. This also happens when I am going back to previous commands in the history (I hit Esc, then 'K' to go back to previous commands) - when I get to a previous command that was longer than the terminal width, the command line gets mangled and I cannot tell where I am at in the command. The only work-around I have found to see the entire long command is to hit "Esc-V", which opens up the current command in a VI editor. I don't think I have anything odd in my .bashrc file. I commented out the "set -o vi" line, and I still had the problem. I downloaded a fresh copy of Putty and didn't make any changes to the configuration - I just typed in my host name to connect, and I still have the problem, so I don't think it's anything with Putty (unless I need to make some config changes) Has anyone else had this problem, and can anyone think of how to fix it? Thanks in advance! Brian

    Read the article

  • Synchronizing ODSEE and OUD

    - by Etienne Remillon
    When it comes to synchronizing between ODSEE and OUD, what should be the best options ? Couple  options are available - Use one of OUD internal capability called Replication Gateway - Use our synchronization tool called Directory Integration Platform part of Oracle Directory Services Plus - Manuel export and import Let's check pro and cons on each method. Replication Gateway is the natural, out of the box solution to perform the task. We created this as a feature of OUD because it works at our replication protocol level. The gateway perform the required adaptation between the ODSEE's replication protocol and OUD's one. The benefits of doing this is that it provide strong consistency between the to type of directories. This fully leverage conflict management implemented in the replication protocols to ensure that changes are applied in a coherent and ordered manner. It does not require specific modification on existing ODSEE production instances such as turning on "retro changelog". Changes are propagated at near speed of replication in both directions. Replication Gateway can also synchronize information that are stored internally in the directory server such as "xxxxx" account locking managed at ODSEE server level and not via the nsyyyy attribute. OUD replication gateway does no require any specific tools or installation specific procedure. It is manged like other OUD component with monitoring and configuration via the standard console. OUD Replication Gateway does not perform adaptation between ODSEE and OUD. Using Directory Integration Protocol as external component to OUD, brings flexibility in remapping and transformations between ODSEE and OUD. There is a price to pay in using DIP to perform the synchronization task. You will have to turn on the retro change log to get access to changes on the ODSEE side (this will impact disk and CPU usage and performances which could be a serious challenge for your existing ODSEE environment (if you have not provisioned additional hardware and instances). You will not benefits of conflict resolution management and this might have to be addressed at application level, which is not always possible to implement. Using export and import seams very simple, but this methodology cannot ensure an highly available deployment with up to date entries on booth sides. This solution can be used if full HA with up-to-date data is not needed (during synchronization time). It often used  if data-cleaning need to take place to avoid polluting a new environment with old un-necessary data.

    Read the article

  • TFS Hosting: discountasp.net TFS

    - by Enrique Lima
    In the last month or so I have been able to test and experience first hand the offering from discountasp.net for hosted TFS 2010. This first part is a description of the setup process for the account itself and getting some additional information on what you will find through the portal on their site. Not long ago, I posted a little tidbit on hosting TFS.  Through it I also did a shameless plug to my employer, our services and the type of hosting we recommend.  So, wouldn’t me running on discountasp.net be an issue?  Actually? NO. Ok, enough rambling.  Let’s get some details here. It is a Software as a Service model.  Through it we get Source Control, Version Control, Work Item Tracking and such.  What about Build?  If your need includes Build Management and such, you may need to look at some other options.  But, still this is a great offering for those that are moving from SourceSafe.  Or organizations who have 3 to 5 developers on staff, and do not foresee getting larger anytime soon.  Can it support more than 5 developers?  Yes, but then we need to get into how are you using TFS.  Do you need more than just Basic?  For example, SharePoint and Reporting Services integration. The signup process was seamless! Very easy to follow, complete and transition to Visual Studio to start working. An email followed the signup process, it contained details on how to get to the Team Foundation Server Control Panel login.  Once there, here is what I saw after the initial setup process of naming my Team Project Collection: So, moving on … once I clicked the area to get my server info, I got the following: Then it was a matter of getting the first user in there: Then on to connecting Visual Studio to my hosted TFS. Getting the server information, and the user account created I will configure those options in Visual Studio. Using Team Explorer, I am adding a new server configuration. Once this is provided, click OK, I will be challenged for a username and password, provide them and you will land on the following screen. Then Click Close. You will now be connected to your server and Team Project Collection. Since this will likely be the first time connecting, you will have no Projects (I already have 2 going). Click Connect, and you will be back in Team Explorer. My next post in the topic will be on Creating your First Team Project and uploading a Project Template to the server.

    Read the article

  • More Efficient Data Structure for Large Layered Tile Map

    - by Stupac
    It seems like the popular method is to break the map up into regions and load them as needed, my problem is that in my game there are many AI entities other than the player out performing actions in virtually all the regions of the map. Let's just say I have a 5000x5000 map, when I use a 2D array of byte's to render it my game uses around 17 MB of memory, as soon as I change that data structure to a my own defined MapCell class (which only contains a single field: byte terrain) my game's memory consumption rockets up to 400+ MB. I plan on adding layering, so an array of byte's won't cut it and I figure I'd need to add a List of some sort to the MapCell class to provide objects in the layers. I'm only rendering tiles that are on screen, but I need the rest of the map to be represented in memory since it is constantly used in Update. So my question is, how can I reduce the memory consumption of my map while still maintaining the above requirements? Thank you for your time! Here's a few snippets my C# code in XNA4: public static void LoadMapData() { // Test map generations int xSize = 5000; int ySize = 5000; MapCell[,] map = new MapCell[xSize,ySize]; //byte[,] map = new byte[xSize, ySize]; Terrain[] terrains = new Terrain[4]; terrains[0] = grass; terrains[1] = dirt; terrains[2] = rock; terrains[3] = water; Random random = new Random(); for(int x = 0; x < xSize; x++) { for(int y = 0; y < ySize; y++) { //map[x,y] = new MapCell(terrains[random.Next(4)]); map[x,y] = new MapCell((byte)random.Next(4)); //map[x, y] = (byte)random.Next(4); } } testMap = new TileMap(map, xSize, ySize); // End test map setup currentMap = testMap; } public class MapCell { //public TerrainType terrain; public byte terrain; public MapCell(byte itsTerrain) { terrain = itsTerrain; } // the type of terrain this cell is treated as /*public Terrain terrain { get; set; } public MapCell(Terrain itsTerrain) { terrain = itsTerrain; }*/ }

    Read the article

  • Upgarde from Asp.Net MVC 1 to MVC 2 - how to and issues with JsonRequestBehavior

    - by Renso
    Goal Upgrade your MVC 1 app to MVC 2 Issues You may get errors about your Json data being returned via a GET request violating security principles - we also address this here. This post is not intended to delve into why the Json GET request is or may be an issue, just how to resolve it as part of upgrading from MVC1 to 2. Solution First remove all references from your projects to the MVC 1 dll and replace it with the MVC 2 dll. Now update your web.config file in your web app root folder by simply changing references to assembly="System.Web.Mvc, Version 1.0.0.0 to Version 2.0.0.0, there are a couple of references in your config file, here are probably most of them you may have:         <compilation debug="true" defaultLanguage="c#">             <assemblies>                        <add assembly="System.Web.Mvc, Version=2.0.0.0, Culture=neutral, PublicKeyToken=31BF3856AD364E35" />             </assemblies>         </compilation>           <pages masterPageFile="~/Views/Masters/CRMTemplate.master" pageParserFilterType="System.Web.Mvc.ViewTypeParserFilter, System.Web.Mvc, Version=2.0.0.0, Culture=neutral, PublicKeyToken=31BF3856AD364E35" pageBaseType="System.Web.Mvc.ViewPage, System.Web.Mvc, Version=2.0.0.0, Culture=neutral, PublicKeyToken=31BF3856AD364E35" userControlBaseType="System.Web.Mvc.ViewUserControl, System.Web.Mvc, Version=2.0.0.0, Culture=neutral, PublicKeyToken=31BF3856AD364E35" validateRequest="False">             <controls>                 <add assembly="System.Web.Mvc, Version=2.0.0.0, Culture=neutral, PublicKeyToken=31BF3856AD364E35" namespace="System.Web.Mvc" tagPrefix="mvc" />   Secondly, if you return Json objects from an ajax call via the GET method you ahve several options to fix this depending on your situation: 1. The simplest, as in my case I did this for an internal web app, you may simply do:             return Json(myObject, JsonRequestBehavior.AllowGet);   2. In Mvc if you have a controller base you could wrap the Json method with:         public new JsonResult Json(object data)         {             return Json(data, "application/json", JsonRequestBehavior.AllowGet);                    }   3. The most work would be to decorate your Actions with:         [AcceptVerbs(HttpVerbs.Get)]   4. Another tnat is also a lot of work that needs to be done to every ajax call returning Json is:                             msg = $.ajax({ url: $('#ajaxGetSampleUrl').val(), dataType: 'json', type: 'POST', async: false, data: { name: theClass }, success: function(data, result) { if (!result) alert('Failure to retrieve the Sample Data.'); } }).responseText;   This should cover all the issues you may run into when upgrading. Let me kow if you run into any other ones.

    Read the article

  • How to Make Objects Fall Faster in a Physics Simulation

    - by David Dimalanta
    I used the collision physics (i.e. Box2d, Physics Body Editor) and implemented onto the java code. I'm trying to make the fall speed higher according to the examples: It falls slower if light object (i.e. feather). It falls faster depending on the object (i.e. pebble, rock, car). I decided to double its falling speed for more excitement. I tried adding the mass but the speed of falling is constant instead of gaining more speed. check my code that something I put under input processor's touchUp() return method under same roof of the class that implements InputProcessor and Screen: @Override public boolean touchUp(int screenX, int screenY, int pointer, int button) { // TODO Touch Up Event if(is_Next_Fruit_Touched) { BodyEditorLoader Fruit_Loader = new BodyEditorLoader(Gdx.files.internal("Shape_Physics/Fruity Physics.json")); Fruit_BD.type = BodyType.DynamicBody; Fruit_BD.position.set(x, y); FixtureDef Fruit_FD = new FixtureDef(); // --> Allows you to make the object's physics. Fruit_FD.density = 1.0f; Fruit_FD.friction = 0.7f; Fruit_FD.restitution = 0.2f; MassData mass = new MassData(); mass.mass = 5f; Fruit_Body[n] = world.createBody(Fruit_BD); Fruit_Body[n].setActive(true); // --> Let your dragon fall. Fruit_Body[n].setMassData(mass); Fruit_Body[n].setGravityScale(1.0f); System.out.println("Eggs... " + n); Fruit_Loader.attachFixture(Fruit_Body[n], Body, Fruit_FD, Fruit_IMG.getWidth()); Fruit_Origin = Fruit_Loader.getOrigin(Body, Fruit_IMG.getWidth()).cpy(); is_Next_Fruit_Touched = false; up = y; Gdx.app.log("Initial Y-coordinate", "Y at " + up); //Once it's touched, the next fruit will set to drag. if(n < 50) { n++; }else{ System.exit(0); } } return true; } And take note, at show() method , the view size from the camera is at 720x1280: camera_1 = new OrthographicCamera(); camera_1.viewportHeight = 1280; camera_1.viewportWidth = 720; camera_1.position.set(camera_1.viewportWidth * 0.5f, camera_1.viewportHeight * 0.5f, 0f); camera_1.update(); I know it's a good idea to add weight to make the falling object falls faster once I released the finger from the touchUp() after I picked the object from the upper right of the screen but the speed remains either constant or slow. How can I solve this? Can you help?

    Read the article

  • Can't update Nvidia driver and having error near the end of the installation

    - by user94843
    I had just got Ubuntu (first timer to Ubuntu so be very descriptive). I think there a problem with my Nvida update it won't let me update it. This is the name of the update in update manager NVIDIA binary xorg driver, kernel module and VDPAU library. When i attempt to install it, it starts out fine but near the end i get a window titaled package operation failed with these under the details installArchives() failed: Setting up nvidia-current (295.40-0ubuntu1) ... update-initramfs: deferring update (trigger activated) INFO:Enable nvidia-current DEBUG:Parsing /usr/share/nvidia-common/quirks/put_your_quirks_here DEBUG:Parsing /usr/share/nvidia-common/quirks/dell_latitude DEBUG:Parsing /usr/share/nvidia-common/quirks/lenovo_thinkpad DEBUG:Processing quirk Latitude E6530 DEBUG:Failure to match Gigabyte Technology Co., Ltd. with Dell Inc. DEBUG:Quirk doesn't match DEBUG:Processing quirk ThinkPad T420s DEBUG:Failure to match Gigabyte Technology Co., Ltd. with LENOVO DEBUG:Quirk doesn't match Removing old nvidia-current-295.40 DKMS files... Loading new nvidia-current-295.40 DKMS files... Error! DKMS tree already contains: nvidia-current-295.40 You cannot add the same module/version combo more than once. dpkg: error processing nvidia-current (--configure): subprocess installed post-installation script returned error exit status 3 Processing triggers for bamfdaemon ... Rebuilding /usr/share/applications/bamf.index... Processing triggers for initramfs-tools ... update-initramfs: Generating /boot/initrd.img-3.2.0-31-generic Warning: No support for locale: en_US.utf8 Errors were encountered while processing: nvidia-current Error in function: Setting up nvidia-current (295.40-0ubuntu1) ... update-initramfs: deferring update (trigger activated) INFO:Enable nvidia-current DEBUG:Parsing /usr/share/nvidia-common/quirks/put_your_quirks_here DEBUG:Parsing /usr/share/nvidia-common/quirks/dell_latitude DEBUG:Parsing /usr/share/nvidia-common/quirks/lenovo_thinkpad DEBUG:Processing quirk Latitude E6530 DEBUG:Failure to match Gigabyte Technology Co., Ltd. with Dell Inc. DEBUG:Quirk doesn't match DEBUG:Processing quirk ThinkPad T420s DEBUG:Failure to match Gigabyte Technology Co., Ltd. with LENOVO DEBUG:Quirk doesn't match Removing old nvidia-current-295.40 DKMS files... Loading new nvidia-current-295.40 DKMS files... Error! DKMS tree already contains: nvidia-current-295.40 You cannot add the same module/version combo more than once. dpkg: error processing nvidia-current (--configure): subprocess installed post-installation script returned error exit status 3 Processing triggers for bamfdaemon ... Rebuilding /usr/share/applications/bamf.index... Processing triggers for initramfs-tools ... update-initramfs: Generating /boot/initrd.img-3.2.0-31-generic Warning: No support for locale: en_US.utf8

    Read the article

  • Where can I hire local programmers with very specific skillsets?

    - by Lostsoul
    I have been browsing the site and haven't found a exact fit to this question so I'll post it but if its already answered(since I'm sure its a common problem, then let me know). I have a business and want to create a totally different product in a different industry than I'm currently in, so I learned how to program and created a working prototype. I have a bit of savings and am getting some cash flow from my current business so I can go out and hire a developer(in the future hopefully it can be permenant but right now I just need a person willing to work on contract and code on weekends or their spare time and I just want to pay in cash instead of equity or future promises). At first I wasn't sure what kind of developer to hire but this question helped me understand I should target specific skills I need as opposed to general programmers. This poses a problem for me since general programmers are everywhere but if I want specific skills I'm unsure how to get them. I thought about a list of approaches but it doesn't feel complete or effective since it seems to be assuming good developers are actively looking. If it helps I want someone local(since this is my first developer hire) and looking for skills like cuda, hadoop, hbase, java and c. Any suggestions? As a FYI, I have been thinking of approaching it as: Go to meet ups for one or more skills I need. Use LinkedIn to find people with the skills I need Search for job postings that contain skills I need and then use linkedIn to reach out to that firms employees since many profiles on linkedin are not very updated or detailed but job postings generally are. Send postings to universities and maybe find a student who loves technology so much they learned these tools on their own. Post on job board. Not sure how successful it will be to post to monster. Use Craigslist, not sure if a highly skilled developer would go here for work. What am I missing? I could be wrong but it seems like good/smart/able developers aren't hunting for work non-stop(especially in this tech job market). Plus most successful people I know have work/life balance so I'm not sure if the best ones really care about code after work. Lastly, most of the skills I need aren't used in big corporations so not sure how aggressively smart developers at small shops look for work. I don’t really know any developers personally, so but should I be using the above plan or if they live balanced lives should I be looking outside of the regular resources(and instead focus on asking around my gym or my accountant or something)? Sorry, I'm making huge assumptions here, I guess because developers are a total mystery to me. I kind of wish Jane Goodall wrote a book on understanding developers social behaviour better :-p

    Read the article

  • Programming Language, Turing Completeness and Turing Machine

    - by Amumu
    A programming language is said to be Turing Completeness if it can successfully simulate a universal TM. Let's take functional programming language for example. In functional programming, function has highest priority over anything. You can pass functions around like any primitives or objects. This is called first class function. In functional programming, your function does not produce side effect i.e. output strings onto screen, change the state of variables outside of its scope. Each function has a copy of its own objects if the objects are passed from the outside, and the copied objects are returned once the function finishes its job. Each function written purely in functional style is completely independent to anything outside of it. Thus, the complexity of the overall system is reduced. This is referred as referential transparency. In functional programming, each function can have its local variables kept its values even after the function exits. This is done by the garbage collector. The value can be reused the next time the function is called again. This is called memoization. A function usually should solve only one thing. It should model only one algorithm to answer a problem. Do you think that a function in a functional language with above properties simulate a Turing Machines? Functions (= algorithms = Turing Machines) are able to be passed around as input and returned as output. TM also accepts and simulate other TMs Memoization models the set of states of a Turing Machine. The memorized variables can be used to determine states of a TM (i.e. which lines to execute, what behavior should it take in a give state ...). Also, you can use memoization to simulate your internal tape storage. In language like C/C++, when a function exits, you lose all of its internal data (unless you store it elsewhere outside of its scope). The set of symbols are the set of all strings in a programming language, which is the higher level and human-readable version of machine code (opcode) Start state is the beginning of the function. However, with memoization, start state can be determined by memoization or if you want, switch/if-else statement in imperative programming language. But then, you can't Final accepting state when the function returns a value, or rejects if an exception happens. Thus, the function (= algorithm = TM) is decidable. Otherwise, it's undecidable. I'm not sure about this. What do you think? Is my thinking true on all of this? The reason I bring function in functional programming because I think it's closer to the idea of TM. What experience with other programming languages do you have which make you feel the idea of TM and the ideas of Computer Science in general? Can you specify how you think?

    Read the article

  • Configurable tables in sql database

    - by dot
    I have the following tables in my database: Config Table: ====================================== Start_Range | End Range | Config_id 10 | 15 | 1 ====================================== Available_UserIDs ========================== ID | UserID | Used_YN | 1 | 10 | t | 1 | 11 | f | 1 | 12 | f | 1 | 13 | f | 1 | 14 | f | 1 | 15 | f | ========================== Users ========================== UserId | FName | LName | 10 |John | Doe | ========================== This is used in a reservation system of sorts... which lets an administrator specify a range of numbers that will be assigned to users in the config table. Once the range has been defined, the system then populates the Available_userIDs table with all the numbers in between the range, and sets the Used_YN flag to false As users sign up, they grab the next user_id number that's not in use... and reserve it. Then the system adds a record to the Users table. Once the admin has specified a range, it is possible that they can change it. For example, they can start with 10-15... and then when the range is used up, they should be able to specify another range like 16 - 99. I've put a unique constraint on the Available_UserIDs table, as well as on the Users table - to ensure that UserIds can't be duplicated. My questions are as follows: What's the best way to prevent the admins from using a range that's already in use? I thought of the following options: -- check either the Users table to see if the start range or ending range numbers are being used. If they are, assume that all the numbers in between are in use too, and reject the range. -- let them specify whatever they want, try to populate the Available_UserIDs table. If there are duplicates, just ignore that specific error message from the database and continue on. How do I find gaps in the number ranges? For example, if they specify 10-15, and then 20-25, it'd be nice to be able to somehow suggest on my web page that 16-19 is currently available. I found this article: http://stackoverflow.com/questions/1312101/how-to-find-a-gap-in-running-counter-with-sql But it only seems to return the first available number... so in my example above, it would only return the number 16. I'm sure there's a simpler way to do things that I'm overlooking!

    Read the article

  • PowerShell – Show a Notification Balloon

    - by BuckWoody
    In my presentations for PowerShell I sometimes want to start a process (like a backup) that will take some time. I normally pop up a notification “balloon” at the start, and then do the bulk of the work, and then pop up a balloon at the end to let me know it’s done. You can actually try out this little sample (on a test system, of course) without any other code to see what it does. Then just put the other PowerShell commands in the #Do Some Work part. Oh – throw an icon (.ico file) in a c:\temp directory or point that somewhere else. (No, this probably isn’t original. Can’t remember where I saw the original code, but I’ve modified it a bit anyway, so if you’re the original author and this looks slightly familiar, post a comment.) [void] [System.Reflection.Assembly]::LoadWithPartialName("System.Windows.Forms") $objBalloon = New-Object System.Windows.Forms.NotifyIcon $objBalloon.Icon = "C:\temp\Folder.ico" # You can use the value Info, Warning, Error $objBalloon.BalloonTipIcon = "Info" # Put what you want to say here for the Start of the process $objBalloon.BalloonTipTitle = "Begin Title" $objBalloon.BalloonTipText = "Begin Message" $objBalloon.Visible = $True $objBalloon.ShowBalloonTip(10000) # Do some work # Put what you want to say here for the completion of the process $objBalloon.BalloonTipTitle = "End Title" $objBalloon.BalloonTipText = "End Message" $objBalloon.Visible = $True $objBalloon.ShowBalloonTip(10000) Script Disclaimer, for people who need to be told this sort of thing: Never trust any script, including those that you find here, until you understand exactly what it does and how it will act on your systems. Always check the script on a test system or Virtual Machine, not a production system. Yes, there are always multiple ways to do things, and this script may not work in every situation, for everything. It’s just a script, people. All scripts on this site are performed by a professional stunt driver on a closed course. Your mileage may vary. Void where prohibited. Offer good for a limited time only. Keep out of reach of small children. Do not operate heavy machinery while using this script. If you experience blurry vision, indigestion or diarrhea during the operation of this script, see a physician immediately. Share this post: email it! | bookmark it! | digg it! | reddit! | kick it! | live it!

    Read the article

  • I can't install Ubuntu 12.04.1 on iMac G5

    - by user89004
    So, I have this iMac G5 that doesn't have iSight, only a small light sensor I think undernieth, machine model 8.2. I tried burning a Ubuntu 12.04.1 PowerPC 64bit .iso to a cd but the computer just won't boot it, I don't know why. Next I tried with a USB but it wouldn't let me boot that either, I created the usb on my dad's win7 laptop as the process was way easier than on freakin Mac or Ubuntu (no command typing AT ALL on windows) I'm able to get into openfirmware and type boot usb and it does show some weird writing that scrolls so fast I can't see anything and then it just gives me this huge no sign like a stop sign and freezez. The sign is grey and the line in the middle is tilted towards the left. An other issue I'm having with hdiutil is that I can't convert the stupid .iso I just downloaded into a .img because the file keeps on dissapearing right when it's done converting it. I used the syntax from Ubuntu support how to create a bootable usb drive under Mac OS X. I even didn't include the 2 stupid ~ that are shown in the syntax that are completly worthless, God only know why they put them there, and I even tried running the whole thing as root with sudo su before the command. The funny thing is that if I convert something smaller it works. The command I was using is hdiutil convert -format UDRW -o /path/to/target.img /path/to/ubuntu.iso I even tried hdiutil convert /path/to/ubuntu.iso -format UDRW -o /path/to/target.img but the same thing happens, the dummy .img.dmg file dissapears when the conversion is done no matter where I set the output file to go. I have tried several different folders, the same thing happens with all of them. I also tried burning a Ubuntu mini iso on a cd, can't remember if it was 11.10 or 12.10 but even thoguh holding c when the iMac boots up does show me the cd and I can boot from it, I get this weird error upon hitting install, it says something like invalid memory access, release keys and error strings I can't read. I don't have any original DVDs from this iMac and can't run hardware diagnostics. WHatever option I try at the command prompt from the mini ubuntu cd I get the same result, error code and openfirmware backdrop that's frozen. I noticed that the pen drive I created on my dads Win7 laptop is formated with MS-DOS but I can still mount it no problem, so it shouldn't have a problem booting it, right? I used the advice on ubuntu.com to make it, from here. Also, my partition is HFS+ so I can't use it as a hard drive and boot from it. I don' have 2 partitions either, just one HDD, one partition. Please help!!!

    Read the article

  • Meaning of Crawl errors

    - by com
    My question is about definition of Crawl errors in Google Webmaster Tools. Crawl errors is devided into few sections. Let's first consider HTTP section. I assume that all broken links in this section was somehow found by crawler, this is not the links from sitemap. If all this links was found by scanning pages from sitemap for links, why it doesn't mention what was the source page, like in sitemap section with column Linked From. Please correct me if I am wrong. Sitemap section. Looks like all those links came from my sitemap. But there is Linked From column, I already know, that all those broken links is from sitemap, so in order to fix the error, I should revise my sitemap. Am I wrong? Not followed section. I don't know what does it mean. Looks like it accumulates all links that caused redirect, but for some reason Google considers all those redirect as wrong redirect. Do you know if there are any set of rules how to determine wrong redirect. Actually I found were was my mistake, I tried to normalize URL and redirect it to the right URL, but I did normalization in a wrong way. Not found section. This section like HTTP section but with 404 errors. This section has Linked From column. But very often Linked From has unavailable. What does it mean, Google can not say me how it found this non existing page. How this section related to sitemap section. Does this section contains all 404 links from sitemap too. But there is too many 404 links, much more than in sitemap. I tried to take a look what we have in Linked From, and I saw that this link came from sitemap two month ago. But why Google keeps it indexed, the link is already dead, new sitemap doesn't have it. If there is any expire date for old links? Unreachable section. Looks like this section for 500 errors. This section doesn't contain Linked From column. There are too many completely meaningless links, I really don't know where this stuff came from, and without Linked From I am not able to figure out how to deal with it. Sorry for such a big topic, but I just want to make it clear, what every section stands for, because it's extremely crucial in order to deal with all those problems. Hopefully it will be useful not just for me. Thanks!

    Read the article

  • How do I get long command lines to wrap to the next line?

    - by BrianH
    Edit It was my .bashrc file. I've copied the same profile from machine to machine, and I used special characters in my $PS1 that are somehow throwing it off. I'm now sticking with the standard bash variables for my $PS1. Thanks to @ændrük for the tip on the .bashrc! ...End Edit... Something I have noticed in Ubuntu for a long time that has been frustrating to me is when I am typing a command at the command line that gets longer (wider) than the terminal width, instead of wrapping to a new line, it goes back to column 1 on the same line and starts over-writing the beginning of my command line. (It doesn't actually overwrite the actual command, but visually, it is overwriting the text that was displayed). It's hard to explain without seeing it, but let's say my terminal was 20 characters wide (Mine is more like 120 characters - but for the sake of an example), and I want to echo the English alphabet. What I type is this: echo abcdefghijklmnopqrstuvwxyz But what my terminal looks like before I hit the key is: pqrstuvwxyzghijklmno When I hit enter, it echos abcdefghijklmnopqrstuvwxyz so I know the command was received properly. It just wrapped my typing after the "o" and started over on the same line. What I would expect to happen, if I typed this command in on a terminal that was only 20 characters wide would be this: echo abcdefghijklmno pqrstuvwxyz Background: I am using bash as my shell, and I have this line in my ~/.bashrc: set -o vi to be able to navigate the command line with VI commands. I am currently using Ubuntu 10.10 server, and connecting to the server with Putty. In any other environment I have worked in, if I type a long command line, it will add a new line underneath the line I am working on when my command gets longer than the terminal width and when I keep typing I can see my command on 2 different lines. But for as long as I can remember using Ubuntu, my long commands only occupy 1 line. This also happens when I am going back to previous commands in the history (I hit Esc, then 'K' to go back to previous commands) - when I get to a previous command that was longer than the terminal width, the command line gets mangled and I cannot tell where I am at in the command. The only work-around I have found to see the entire long command is to hit "Esc-V", which opens up the current command in a VI editor. I don't think I have anything odd in my .bashrc file. I commented out the "set -o vi" line, and I still had the problem. I downloaded a fresh copy of Putty and didn't make any changes to the configuration - I just typed in my host name to connect, and I still have the problem, so I don't think it's anything with Putty (unless I need to make some config changes) Has anyone else had this problem, and can anyone think of how to fix it? Thanks in advance! Brian

    Read the article

  • Shelving &ndash; What is it &ndash; and more importantly, can it help me?

    - by Chris Skardon
    Since we shifted to TFS we’ve had the ability to perform what is known as ‘shelving’. Shelving (whilst not a wholly new topic in the world of SCC) is new to us, and didn’t exist in our previous SCC solution – SVN. Soo… what is it? What? Shelving is a way to check-in but not check-in your code. By shelving you submit a copy of your ‘pending changes’ to the SCC server, (which maintains a list of the shelvesets) and once that is done you can either continue working, or undo your changes, safe in the knowledge that a backup copy exists on the server. You can unshelve your code at any time and get back to the state you were when you shelved. Yer, that is great but why not just check it in?? Shelvesets don’t have to build. The shelveset you put in there could be entirely broken, or it might solve every bug in the system – shelves aren’t continuously integrated so you can shelve anything. Hmmmm… What else? Shelving allows us to do some pretty cool stuff that beforehand was quite frankly a pain. For instance – Gated Check-ins are implemented via the shelving mechanism, when code is checked-in, what you’re actually doing is shelving it, the Build Controller will build the shelveset with the original code and if it succeeds, the code will be committed, if it fails – well – it’s only you that has to fix the code :) Other nice features are things like the ability to share code you are working on… For example, if I was having trouble with a particular piece of code, I could shelve it, and then you (yes you) could then get that shelveset and check out the problem for yourself, and if you fix it?? Well – you could check-it in! Nice, but day-to-day shizzle? Let’s say you’ve been working on your project and your project manager comes over to you and says: “Hey, errr, bad times, there is an urgent bug we need you to fix, it needs to go out now!” (also for this to play out – we’ll need to assume you’re currently working in the 'release’ branch for another bug fix (maybe))… You could undo all your current changes (obviously you’ll probably backup your code using zip or something I imagine) fix the bug, then re-copy your backup over the top, or you could shelve and unshelve. Perhaps some other uses will awaken the shelver in you… :) Before each checkin – if you shelve, you no longer need to worry (if indeed you do) about resolving conflicts and mysteriously losing your code… Going home at night? Not checking in straight away? Why not shelve, this way – should the worst come to the worst and your local pc gives up, you can just get the shelveset onto another machine and be up and running in literally seconds minutes…

    Read the article

< Previous Page | 481 482 483 484 485 486 487 488 489 490 491 492  | Next Page >