| 59 | | immutable. There will be two special elements holding NONE-REF and the strings "AUTO-CHAIN" and "NO CHAIN". Other elements should not be created withh NONE |
| 60 | | |
| 61 | | resource ref. So the constructor should be private and there will be a create method to validate the given refs. [[BR]] [[BR]] |
| 62 | | |
| 63 | | In TextChainingLogic::ON_CHAIN_TEXT_FRAME check what elements was choosen and if it is with none ref change HeadTextFrameR4.KEY_CHAIN_MODE corespondingly, |
| 64 | | |
| 65 | | otherwise chain the frame pointed by the resource ref held in the element to the head of the current frame. This hsould always change the chain mode to |
| 66 | | |
| 67 | | manual. [[BR]][[BR]] |
| 68 | | |
| 69 | | In org.sophie2.main.app.commons.util package create a util class named TemplateUtil. It will contain helper methods for the page and frame templating and for |
| 70 | | |
| 71 | | the creating of new templates. There is a difference in code of templating those elements since the templating of sub elements is not yet final.[[BR]] |
| | 59 | In org.sophie2.main.app.commons.util package create a util class named TemplateUtil. It will contain helper methods for the page and frame templating and for the creating of new templates. There is a difference in code of templating those elements since the templating of sub elements is not yet final.[[BR]] |
| 84 | | logic.[[BR]][[BR]] |
| 85 | | |
| 86 | | The TextChainUtils should be quite extended to hold the helper methods used by the chaing logic. Some of them are moved from the TextChainingLogic. The |
| 87 | | |
| 88 | | method used for auto-chaining will be called reflowText. It would get a head frame view as argument and basically handle all auto-fitting of its text flow |
| 89 | | |
| 90 | | (creatin, deletion of pages). It should get all frame views in this head frame`s chain. Then it gets the number of frames on the last page and runs the |
| 91 | | |
| 92 | | HotLayout.dryrun method with them. If the result is 0 - there is nothing more to do. If the result is -1 this means that the last page is empty. There should |
| 93 | | |
| 94 | | be a ceparate method that hamdles the deletion of extra pages. Its name should be reduceChain. If the result is an positive int this is the number of pages |
| 95 | | |
| 96 | | exactly as the first one to be added. This should be done by the method expandChain.[[BR]] |
| 97 | | |
| 98 | | * expandChain - The chain should be expanded with pages that use templates made from the last page in the current chain and children that use templates made |
| 99 | | |
| 100 | | from the children of the last page`s children. The last page should also be made to use this template so that wnem it is changd it chanes just like the |
| 101 | | |
| 102 | | others in the chain. Those templates are added to the library just like regular ones. There are some x-cases:[[BR]] |
| 103 | | - When there is another auto-chain in the page - the new frames are made the next in their chain. [[BR]] |
| 104 | | - If there is an head frame (that is not autochained) its copy uses the same text resource. [[BR]] |
| 105 | | - when there is a tail frame (or head frame that is auto-chained), its copy is the next in the chain of the source. Its chain order is also placing it |
| 106 | | |
| 107 | | before any other frame of this chain.[[BR]] |
| 108 | | |
| 109 | | * reduceChain - Checks all pages starting from the last one for pages that are all templated. This means that their templated keys should be templated and |
| 110 | | |
| 111 | | all their children should also be templated. Also all of its tail text frames should be empty. If such page is found it is deleted. If there is no more an |
| 112 | | |
| 113 | | empty page in the end of the chain the job is done, otherwise we continue looking ofr pages to remove from the previous page. [[BR]][[BR]] |
| | 76 | * reduceChain - Checks all pages starting from the last one for pages that are all templated. This means that their templated keys should be templated and all their children should also be templated. Also all of its tail text frames should be empty. If such page is found it is deleted. If there is no more an empty page in the end of the chain the job is done, otherwise we continue looking ofr pages to remove from the previous page. [[BR]][[BR]] |
| 124 | | In HeadTextFrameView in the class textFlow used for a the resource property with the same name crate a class named ReflowThread. Its run method will be quite |
| 125 | | |
| 126 | | simple. It should sleep for a second and then invoke later in the event dispatching thread the method TextChainUtils.reflowText giving it this |
| 127 | | |
| 128 | | HeadTextFrameView object. he Headin getTextLayout method that extends Thread and delays the invoking of the TextChainUtils.reflowText method. Now the only |
| 129 | | |
| 130 | | thing that is left is to start the thread and this will happen every time the method getTextLayout() is called on an autochained head view. It would check if |
| 131 | | |
| 132 | | there is a thread runing and if not - strat one. Otherwise its job is done. For optimization there may be a check whether the text flow is going to be the |
| 133 | | |
| 134 | | same as the last time the thread was called. [[BR]][[BR]] |
| | 85 | In HeadTextFrameView in the class textFlow used for a the resource property with the same name crate a class named ReflowThread. Its run method will be quite simple. It should sleep for a second and then invoke later in the event dispatching thread the method TextChainUtils.reflowText giving it this HeadTextFrameView object. he Headin getTextLayout method that extends Thread and delays the invoking of the TextChainUtils.reflowText method. Now the only thing that is left is to start the thread and this will happen every time the method getTextLayout() is called on an autochained head view. It would check if there is a thread runing and if not - strat one. Otherwise its job is done. For optimization there may be a check whether the text flow is going to be the same as the last time the thread was called. [[BR]][[BR]] |
| 137 | | |
| 138 | | Now the logic assumes that in a chain there are never same chain orders. But this is not always this way. when there are same chain orders the order of the |
| 139 | | |
| 140 | | pages and the z order are taken in consideration. Also this is the reason why the getChainedFrames() helper method can not return a map - since there is not |
| 141 | | |
| 142 | | always a single TailFrameH corresponding to a chain order. [[BR]] |
| 143 | | The logic of the two buttons for unchain (previous and next) should be unified. There will be a common method splitChain that they will use. It would make |
| 144 | | |
| 145 | | the given frame - head of a new chain and all after it would be chained to it. It could use the same text resource as the previous one or get a new empty |
| 146 | | |
| 147 | | one. This should be specified by an flag argument. [[BR]] |
| 148 | | This method will also be used when deleting pages. Its logic was broken if deleting from the button but it was ok if deleting by 'del' or backspace. The |
| 149 | | |
| 150 | | delete page button should have the same event as the others (its source should be the PagePreviewPalette). Now before the page is deleted the chains that |
| 151 | | |
| 152 | | start in it are split to the first frame that is not in this page. [[BR]][[BR]] |
| | 88 | Now the logic assumes that in a chain there are never same chain orders. But this is not always this way. when there are same chain orders the order of the pages and the z order are taken in consideration. Also this is the reason why the getChainedFrames() helper method can not return a map - since there is not always a single TailFrameH corresponding to a chain order. [[BR]] |
| | 89 | The logic of the two buttons for unchain (previous and next) should be unified. There will be a common method splitChain that they will use. It would make the given frame - head of a new chain and all after it would be chained to it. It could use the same text resource as the previous one or get a new empty one. This should be specified by an flag argument. [[BR]] |
| | 90 | This method will also be used when deleting pages. Its logic was broken if deleting from the button but it was ok if deleting by 'del' or backspace. The delete page button should have the same event as the others (its source should be the PagePreviewPalette). Now before the page is deleted the chains that start in it are split to the first frame that is not in this page. [[BR]][[BR]] |
| 157 | | model. This could only be found while developing and I don`t think there is an ticket for it.[[BR]] |
| 158 | | |
| 159 | | The other one is in the modify method of the ResourceModel. The problem occures when creating template from templated frame. There is an exception thrown but |
| 160 | | |
| 161 | | there is no ticket also. It this case the values of some of the keys are already null and there is no need for them to be set to null but their 'locked-key's |
| 162 | | |
| 163 | | are still set to false. The moidfy method relies on the fact that when modifing a node corresponding to a key there are chanes meant for it if the sub map of |
| 164 | | |
| 165 | | chanes for entries whose keys start with this key is not empty. This is not true for the case of templated keys though since their locked keys idd start |
| 166 | | |
| 167 | | exactly the same as their ids. This method hsould not rely on such assumptions. [[BR]] |
| 168 | | |
| 169 | | |
| | 95 | The other one is in the modify method of the ResourceModel. The problem occures when creating template from templated frame. There is an exception thrown but there is no ticket also. It this case the values of some of the keys are already null and there is no need for them to be set to null but their 'locked-key's are still set to false. The moidfy method relies on the fact that when modifing a node corresponding to a key there are chanes meant for it if the sub map of changes for entries whose keys start with this key is not empty. This is not true for the case of templated keys though since their locked keys idd start exactly the same as their ids. This method hsould not rely on such assumptions. [[BR]] |
