I fixed one of the VTK bugs tonight. The bug was 0010861 in the VTKPolyDataConnectivityFilter.
I had the source code for VTK straight on my system, so I had to install it through GIT instead. To do this I did the following:
cd VTK
git clone git://vtk.org/VTK.git VTK
git clone git://vtk.org/VTKData.git
./Utilities/SetupForDevelopment.sh
git remote add gerrit asbissad@review.source.kitware.com:VTK
git commit
git add
git push gerrit HEAD:refs/for/master/PrivateToProtected
Sunday, February 27, 2011
Thursday, February 24, 2011
TOS CH7 - 2/24/11
PATCHES
Exercise 7.2.2
The first diff command is VERY different than diff with -u.
The first one gave me:
6c6
< prntf(Hello World.\n");
---
printf("Hello World!\n");
WHEREAS, diff with the -u gave me clearer details.
It showed where the program changed, the lines, everything.
The UNIFIED diff is Unparallelled!
Exercise 7.8:
I received the following input when I did this exercise. Basically null is an empty file so everything in the diff is an addition(+)
--- /dev/null 2010-12-03 09:50:34.426290301 -0500
+++ foo.c 2011-02-24 11:41:59.000000000 -0500
@@ -0,0 +1,8 @@
+/*foo.c*/
+
+#include <stdio.h>
+
+main() {
+printf("This is a new file.\n");
+}
+
Exercise 7.9:
./configure was intense, as was make.
The text is a little confusing. I did not see those as two separate commands at first.
I would suggest some clean up in terms of the typos in the book to make things clearer.
The src/echo command worked.
Exercise 7.2.2
The first diff command is VERY different than diff with -u.
The first one gave me:
6c6
< prntf(Hello World.\n");
---
printf("Hello World!\n");
WHEREAS, diff with the -u gave me clearer details.
It showed where the program changed, the lines, everything.
The UNIFIED diff is Unparallelled!
Exercise 7.8:
I received the following input when I did this exercise. Basically null is an empty file so everything in the diff is an addition(+)
--- /dev/null 2010-12-03 09:50:34.426290301 -0500
+++ foo.c 2011-02-24 11:41:59.000000000 -0500
@@ -0,0 +1,8 @@
+/*foo.c*/
+
+#include <stdio.h>
+
+main() {
+printf("This is a new file.\n");
+}
+
Exercise 7.9:
./configure was intense, as was make.
The text is a little confusing. I did not see those as two separate commands at first.
I would suggest some clean up in terms of the typos in the book to make things clearer.
The src/echo command worked.
Tuesday, February 22, 2011
2/22/11
We had a good meeting this morning. We all met in the computer lab at 9am. I was able to test out the fix for the bug on my system, of which I wrote the how-to instructions on the team wiki. Honestly the code is much simpler than everything else leading up to it (i.e. installing and compiling all the correct files). It is nice to see just a little bit of overlap between this class and CSCI340.
We also just found the VTKSTLWriter Class Reference, which would have been nice to have a week ago. But we have it now so that will be beneficial. We did not find it from our searches, rather through corresponding with someone else on the project who recommended the page to us through comments on the Bug Page. His name is David Doria. He seems to be quite active and quick to respond, which is nice.
One thing we have had an issue with is getting our fix submitted. There was no clear documentation on it, so we placed a post on the VTK mailing list. We were informed that we need specific access to be able to post our bug. So right now that is still up in the air.
We also just found the VTKSTLWriter Class Reference, which would have been nice to have a week ago. But we have it now so that will be beneficial. We did not find it from our searches, rather through corresponding with someone else on the project who recommended the page to us through comments on the Bug Page. His name is David Doria. He seems to be quite active and quick to respond, which is nice.
One thing we have had an issue with is getting our fix submitted. There was no clear documentation on it, so we placed a post on the VTK mailing list. We were informed that we need specific access to be able to post our bug. So right now that is still up in the air.
Thursday, February 17, 2011
Alumni Symposium
I attended the Alumni Symposium on February 15, 2011. It was definitely interesting as there were past students from many different years.
Some of the key points I learned were:
Business Development - How program development should be. Its good to read books on this.
When applying to a job:
If its a Big company- Learn a specialized language.
If its a Small company - Know lots of languages
One of the most important things is:
What is your interest in computing? - Questions they may ask: Whats the last book you read about computing? Where do u get your computer news?
Good characteristics:
Creativity - Play an instrument. Its good to have a combination if the technical and creativity.
Good skill:
Know Databases. Web Applications are becoming very popular.
Some of the key points I learned were:
Business Development - How program development should be. Its good to read books on this.
When applying to a job:
If its a Big company- Learn a specialized language.
If its a Small company - Know lots of languages
One of the most important things is:
What is your interest in computing? - Questions they may ask: Whats the last book you read about computing? Where do u get your computer news?
Good characteristics:
Creativity - Play an instrument. Its good to have a combination if the technical and creativity.
Good skill:
Know Databases. Web Applications are becoming very popular.
Thursday, February 10, 2011
2/10/11
We looked through a whole list of bugs on Tuesday, and when I was looking at the descriptions I found one that looked like it was doable to fix by the description.
Description: vtkSTLWriter only writes the first 3 points of each polygon. If the input is a triangle strip, for example, only the first 3 points of the strip are written. I'd suggest that it throw an error (or at least a warning) indicating that the input is not what it is expecting. Maybe this could be done by checking the number of points that were written to the file against the number of input points?
Here is a screenshot of the code of the bug we are attempting to fix.
Current Proposed Fix:
Step 1. Find out the number of points passed to the method in the pts list.
Step 2. Use an IF to find out the number of points and if it exceeds the # from step 1
Step 3. If it does, have a message in standard output that tells them only the first three points were used.
As I said we are about to learn C in Operating Systems. But I have not learned C++. We have learned Java in previous classes, so I will be looking at Java to C++ tutorials. Here is one of the good ones that Dr. Leclerc recommended to me: http://www.horstmann.com/ccj2/ccjapp3.html
 | 0010854 | minor | new | 2010-06-19 | vtkSTLWriter should warn when input is not all written |
Description: vtkSTLWriter only writes the first 3 points of each polygon. If the input is a triangle strip, for example, only the first 3 points of the strip are written. I'd suggest that it throw an error (or at least a warning) indicating that the input is not what it is expecting. Maybe this could be done by checking the number of points that were written to the file against the number of input points?
Here is a screenshot of the code of the bug we are attempting to fix.
Current Proposed Fix:
Step 1. Find out the number of points passed to the method in the pts list.
Step 2. Use an IF to find out the number of points and if it exceeds the # from step 1
Step 3. If it does, have a message in standard output that tells them only the first three points were used.
As I said we are about to learn C in Operating Systems. But I have not learned C++. We have learned Java in previous classes, so I will be looking at Java to C++ tutorials. Here is one of the good ones that Dr. Leclerc recommended to me: http://www.horstmann.com/ccj2/ccjapp3.html
Tuesday, February 8, 2011
2/8/11
6.4
0000435: Coverage should use CoverageCommand instead of gcov
We learned about C and gcc yesterday in Operating Systems. This is the second time there has been some type of parallels in these two classes, no matter how little. Unfortunately I found out this bug was for a different sub project.
For some reason the bug tracker is loading extremely slowly. I have been narrowing down categories such as date and project and it takes a minute to load each.
It also seems odd that there could be bugs titled as "new" that are from 2003.
0007760: vtkImplicitPlaneRepresentation Edges Property Not Set
This is the entire description: The following is missing from vtkImplicitPlaneRepresentation's constructor:
this->EdgesActor->SetProperty(this->EdgesProperty);
The chapter states:
Good bugs provide relevant system information
Good bugs only report a single issue
Good bugs are followed by people who care about a solution
Good bugs are reproducible with the fewest possible steps
No one has worked on this because it was not well described. It was definitely a short description. IT certainly did not follow these guidelines.
6.5
I have successfully added an account on Bugtracker under the screen name andrew.bissada
It was very easy.
6.6
I looked at some bugs to fix but honestly I don't understand all of it well enough to fix yet. It doesn't appear that anyone that can code could just go in right away and start working on it.
6.7
To triage for our project is not bad at all. The Bug tracker is a lot better than I expected. You can apply many filters to your search including status and severity. Status includes: new, feedback, acknowledged, confirmed, assigned, resolved, closed. Under severity is includes: feature, block, trivial, text, tweak, minor, crash, and major.
Five bugs that cause VTK to crash are:
0000435: Coverage should use CoverageCommand instead of gcov
We learned about C and gcc yesterday in Operating Systems. This is the second time there has been some type of parallels in these two classes, no matter how little. Unfortunately I found out this bug was for a different sub project.
For some reason the bug tracker is loading extremely slowly. I have been narrowing down categories such as date and project and it takes a minute to load each.
It also seems odd that there could be bugs titled as "new" that are from 2003.
0007760: vtkImplicitPlaneRepresentation Edges Property Not Set
This is the entire description: The following is missing from vtkImplicitPlaneRepresentation's constructor:
this->EdgesActor->SetProperty(this->EdgesProperty);
The chapter states:
Good bugs provide relevant system information
Good bugs only report a single issue
Good bugs are followed by people who care about a solution
Good bugs are reproducible with the fewest possible steps
No one has worked on this because it was not well described. It was definitely a short description. IT certainly did not follow these guidelines.
6.5
I have successfully added an account on Bugtracker under the screen name andrew.bissada
It was very easy.
6.6
I looked at some bugs to fix but honestly I don't understand all of it well enough to fix yet. It doesn't appear that anyone that can code could just go in right away and start working on it.
6.7
To triage for our project is not bad at all. The Bug tracker is a lot better than I expected. You can apply many filters to your search including status and severity. Status includes: new, feedback, acknowledged, confirmed, assigned, resolved, closed. Under severity is includes: feature, block, trivial, text, tweak, minor, crash, and major.
Five bugs that cause VTK to crash are:
| 0011607 | crash | new | 2010-12-15 | Segfault with vtkXMLPImageDataReader | |||||
| 0006461 | crash | new | 2010-12-06 | vtkDataArray::Size wrong after call to IDirect3DDevice9::CreateDevice | |||||
| 0011102 | crash | new | 2010-08-06 | vtkDelaunay2D with some pairs of datasets and transforms crashes | |||||
| 0009783 | crash | new | 2010-04-20 | Crash in vtkline.cxx - access uninitialized data | |||||
| 0010464 | crash | new | 2010-04-20 | vtkSeedWidget CRASH!!! |
Thursday, February 3, 2011
Feb 3, 2011
I was not able to log into Ubuntu on the computers in the lab, so I had to install it (via Wibu)on my system. It was messing up left and right. I tried getting subversion on it but I had to update my Ubuntu version first, which the one downloaded from Windows was 10.10. System -> Administrator -> Update Manager.
I had subversion on windows over a week ago but I was informed I was going to have problems if I wasn't working on Ubuntu.
Next I Under the Terminal from Applications and accessories) type in svn co svn://svn.gna.org/svn/freeciv/trunk freeciv
Wow, freeciv has so many little files to install.
(EDITING)
Then I typed:[gregdek@ip-10-242-118-147 freeciv]$ rpm -q gcc
It told me rpm was not installed so I installed it.
Then I typed the command again and it said: package gcc is not installed
same thing with yum install gcc
sudo apt-get x...gotta love it.
Basically every time there was RPM I had to replace it with sudo apt-get install x
To get glib to work I had to install the package libtoolize first. I did this by going to System then administration then synaptic package manager. Then I had to get a C++ compiler, but when I tried again it said it wasnt there for some reason.
I had subversion on windows over a week ago but I was informed I was going to have problems if I wasn't working on Ubuntu.
Next I Under the Terminal from Applications and accessories) type in svn co svn://svn.gna.org/svn/freeciv/trunk freeciv
Wow, freeciv has so many little files to install.
(EDITING)
Then I typed:[gregdek@ip-10-242-118-147 freeciv]$ rpm -q gcc
It told me rpm was not installed so I installed it.
Then I typed the command again and it said: package gcc is not installed
same thing with yum install gcc
sudo apt-get x...gotta love it.
Basically every time there was RPM I had to replace it with sudo apt-get install x
To get glib to work I had to install the package libtoolize first. I did this by going to System then administration then synaptic package manager. Then I had to get a C++ compiler, but when I tried again it said it wasnt there for some reason.
Tuesday, February 1, 2011
February 1, 2011
Installation:
So last week I downloaded and installed many parts of ITK. That was quite the headache. It felt like a never ending path of files and commands in the prompt. I also had to install GIT and CMake. The commands themselves were not hard, just a simple copy and paste into the prompt, there were just a ton of steps in installing different pieces.
Letdown:
We discovered that the list of bugs for ITK was smaller than we originally thought. Apparently there was only one that was SPECIFICALLY for ITK.
Change?
We are currently considering changing projects. Of course when you come up with your list of original projects you never expect that you will pick one that you will end up changing.
So last week I downloaded and installed many parts of ITK. That was quite the headache. It felt like a never ending path of files and commands in the prompt. I also had to install GIT and CMake. The commands themselves were not hard, just a simple copy and paste into the prompt, there were just a ton of steps in installing different pieces.
Letdown:
We discovered that the list of bugs for ITK was smaller than we originally thought. Apparently there was only one that was SPECIFICALLY for ITK.
Change?
We are currently considering changing projects. Of course when you come up with your list of original projects you never expect that you will pick one that you will end up changing.
Subscribe to:
Posts (Atom)