Ali Haydar Dev log
Community Bonding Period
- setting up the environment installing ubuntu in dual boot and rebuilding the project in windows and ubuntu
- working on bugs:- the annotation is pointing to the origin fixed and patch submitted the drawing of numerical annotation results in no annotation working on it and notice that after constructing the numerical annotation you can't draw anything in the graphical screen until writing the Z command in MGED or reopening the database.
- using the code to modify the Vlist to draw a box around the label it is like a hack, not a real function just an experiment, it makes me realize that the label drawing should not start from the reference point but a little bit further so the box doesn't touch the label.
- contacting with the mentors and asking about there opinions.
This page is will be edited to include what I've learned so far and served as a log for the coming development.
- May 27 start reading the annotation code and trying to understand the problem in the opengl depiction in Wgl and ogl display managers
- May 28-29 failing to have any result with OpenGL and keep trying the logic exists in dm-X display manager but I am failing to copy it with OpenGL functions. the program is using OpenGL in immediate mode.
- May 30 learning how to manipulate vector lists
- May 31 the annotation vlist is complete when it exists in drawVlist() function in dm-wgl.c starting from the third point we can obtain the coordinates to the label vlist and use them to construct the box around the label
- June 1 learning opengl immediate mode the projection wanted by the society is like the one produced by Glortho() but it is not giving the wanted results
- June 2 working on the short annotation bug when you have a short annotation like one number or one letter it is not drawn in the graphics window one fix suggested by a fellow programmer is commenting the line 79 in vlist.c but this fix isn't approved by the society.
- June 3 write the function ant_label_box() which draw the box around the annotation it was good but we have to optain a pointer to the segments inside the annotations and decide which segments have a box around it and which is not
- June 4 one day off didn't do anything just read some code.
- June 5 reading how the properties of line seg and text segment are entered and read by the annot_in function
- June 6-9 structural dynamics exam didn't do much. and didn't pass the exam :(
- June 10 I am working on the short annotations bug starting in bn_vlist_bbox() we will see what happen today and update.
June 10 the solutions I proposed is not accepted because it is changing the calculations of the bounding box for lines and triangles
- June 11 trying another approach in solving short annotation problem which is another way of calling bn_vlist_bbox() in bound_solid() may be changing the matrix flag in cmd=15 to cmd=12 so the annotation is treated as one point which is the point that is in model coordinates not in display coordinates.
- June 12 submitted a patch to solve the short annotation problem but it didn't solve the long annotation problem because of the long annotation vlist hase multiple chunks.
June 12 trying to add a third segment to the annotation static one defined inside annot_in() with its parameters to see if it will be converted correctly to vlist.
- June 13 the patch was good and I submitted another one to solve the problem of bbox() in long annotations when the annotation vlist is constructed from more than one chunk the display mode is not memorized thought the iterations
- June 14 I managed to draw a circle segment between the annotated point and the reference point replacing the text segment by a carc segment I am working on the entring logic for three segments of the annotation.
June 14 now the annotation contains three segments a leader line, text, and static arc segment I defined it inside the annot_in() function. the annotation is imported correctly and all the segments are converted correctly to vlist.
- June 15 working on the problem of the bounding box, the solution is to change the bbox() function
- June 15 prompt the user if he wants a carc segment in his annotation it seems like the annotation construction will be so hard on the user
- June 16 send the patch of the new bbox() function it needs some modifications
- June 16 the society suggested making an annotation editor using the existing code of sketch editor so I am learning tcl/tk
- June 17 added the annot choice in creat menu and found where to assign the callbacks for classes in tcl/tk script which is done using auto_mkindex command and creats a script called tclindex.tcl
- June 18 after copying the sketch_editor class in skt_ed.tcl and make annot_editor class the menu button still not recognizing the annotation, maybe this is happening because the mged_make() function doesn't now about the annotation so know we will make them meet each other :).
- June 19 made some changes on the new bbox() function, the annotation editor is working now but failing to save the annotation,
- June 20 the error in saving the annotation was in the function which writes the annotation to the database
- June 21 the new bbox() function is accepted solving all the problems, now I am working on modifying the Gui of the annotation editor
- June 22 facing a problem with drawing a line to the same point from the annotation editor and from the in command
- June 23 made the user able to enter the annotated point and store it in V which will be the origin of coordinates to draw the hall annotation
- June 24 the annotation editor stores the chosen vertex twice I don't know why. still working on the Gui to make the user able to enter the label of the annotation and the text placement coordinates work going very slow the placement of the widgets in the window is making me crazy.
- June 25 now its the time to make a method to create the txt segment in tcl script but whenever I add a new method to the script the whole editor goes down.
- June 26 submitted the patch for the bbox() it is correct but contains a lot of mistakes in the indents and white spaces, I am working on making a unit test for the code I wrote.
- June 27 reading the unit tests in bn library it appears there is a file called bn_test.c which calls the test functions and it exists in the configured project I.e the build file you generate using Cmake , I have added the test file that I will write in, to the list of bn_test_srcs in cmakelist in test file of libbn .
- June 28 still working on the unit test program the main function is automatically generated based on the test file name, I have added the tests also to the cmakelist file the program will test the bn_vlist_cmd_cnt() function in vlist.c which returns the number of commands in a vlist so we have to invent a list and send it to the function to see if it works properly.
- June 29 the vlist is a doubly liked list so we have to obtain a head for the list and start adding elements I am thinking to put a for loop which adds points to the vlist depending on the test command in cmakelist file I will use RT_ADD_VLIST macro on the created head to add points to the list. also worked a little on the annotation editor but still not good the new methods I add breaks the hall script I don't know why. at the end of the day the test file works and the main function is called correctly from bn_test program.
- June 30 I have a problem building the test program because of using RT_ADD_VLIST the error is unresolved external RTG which is from the type rt_g defined in global.h, maybe it is a link error to librt, because I noticed that the RT_ADD_VLIST macro works in files which is existed inside librt maybe I am wrong ,
- July 1 the same link error still keeping me from going ahead I tried to define rt_g RTG; object inside the test file and the build is okay but when executing the macro the program get BU_ASSERT error I don't know why I asked the mentors to learn me how to link to librt and i tried a solution by making a function inside annot.c to add the points to the vlist now my new function is the unresolved external object.
- July 2 the same link error we linked librt to the test file the function works but when you use RT_ADD_VLIST inside it you get the BU_ASSERT error and when using RT_ADD_VLIST inside the test program you get the unresolved external error when building the code.
- July 3 found the solution by adding two lines to the test program which is 1- rt_g RTG; 2- BU_LIST_INIT(&RTG.rtg_vlfree); this solution is you need a free list to use the vlist which I learned in a hard way wasting 3 days searching online how to link a library to your code in cmake system.
- July 4 rewrite the test file and adding all the expected cases of the function like Null pointer to no list and very long list. I have sent the file to the society to get opinions about the whole operation I think I finished it. Now looking to the bugs in the annotation we can notice that the annotation label is presented in the top right position no matter what the position flag is I am working on this now and asked for the mentors to revise the goals of the project because the annotation editor is written in tcl/tk which will not be used in the new GUI this will make me do a lot of things that will not be useful. also, I have an exam on the 9th of June.
- July 5 I couldn't do anything due to some problems in my sleeping schedule.
- July 6 I found a bug in the annotation behavior which is the position of the label doesn't change whatever the user chose from the position flags I am working on this and I have sent a suggestion for a partial solution.