| 07:17.20 | *** join/#stepcode hoiji (~hoiji@115.241.233.242) |
| 07:24.06 | hoiji | Hi, adhering to mpictor's comments, I have made a small commit to the forked repository of the stepcode in my git hub handle. https://github.com/hoiji09/stepcode/compare/stepcode:master...master |
| 07:24.10 | hoiji | If possible I would like to merge it with the main code. Should I submit a patch file or should I initiate a pull request? |
| 07:33.35 | hoiji | Oh, I almost forgot, the above commit increases the speed of lazy-loading by 1.5 times on my machine :) |
| 12:07.34 | *** join/#stepcode hoiji (~hoiji@115.242.10.149) |
| 16:31.14 | mpictor | <PROTECTED> |
| 16:31.36 | mpictor | also, I notice that two of the commits are by 'dell' - who is this? |
| 16:33.50 | hoiji | sorry, about 'dell' git was taking my machine name instead of my username. |
| 16:33.58 | hoiji | I have fixed that now |
| 16:43.25 | *** join/#stepcode kesha (~kesha@14.139.122.114) |
| 16:46.49 | mpictor | ok |
| 16:48.54 | mpictor | I just realized that atoll() doesn't catch errors - strtol would be better |
| 16:49.18 | mpictor | that would give you an easy way to check for input numbers that are just barely too big |
| 16:54.46 | hoiji | wow! thanks for the input, I feel embarrassed for overlooking the buffer length folly |
| 16:55.20 | hoiji | I will fix it shortly and commit |
| 16:57.48 | mpictor | live and learn :) |
| 17:02.13 | mpictor | also, I don't want to just print an error and skip the instance. Instead, we should signal an error using _error |
| 17:02.21 | mpictor | and stop trying to read the file |
| 17:02.36 | mpictor | err, stop trying to read the section |
| 17:02.44 | mpictor | maybe other sections of the fill will be ok |
| 17:07.24 | mpictor | *file |
| 17:14.02 | hoiji | that would mean small modification to lazyP21DataSectionReader.cc |
| 17:14.08 | hoiji | Would that be ok? |
| 17:21.24 | mpictor | what do you have in mind? |
| 17:22.14 | mpictor | oh, checking _error? yes, that's fine |
| 17:25.11 | *** join/#stepcode kesha (~kesha@14.139.122.114) |
| 17:25.40 | hoiji | Also can I assume that the largest instance number will be never greater then 2^64 - 1 |
| 17:26.03 | hoiji | The above assumption would be needed to use strtoll |
| 17:26.24 | hoiji | *stroull |
| 17:26.47 | hoiji | *strtoull (sorry) |
| 17:26.54 | mpictor | heh |
| 17:28.12 | mpictor | I'd add an assertion using numeric_limits to check the maximum value |
| 17:28.42 | mpictor | then, it would crash if that assumption was invalid |
| 17:29.50 | mpictor | not as good as handling larger sizes without complaint, but I'd rather a crash than it to continue running and act up |
| 18:14.56 | hoiji | mpictor: I am in doubt, by signaling `_error` you meant `_error` of type `ErrorDescriptor`. Right? |
| 18:17.55 | mpictor | yes. _error->GreaterSeverity( SEVERITY_INPUT_ERROR ) |
| 18:18.09 | mpictor | and you can put your error message in there as well |
| 18:18.40 | mpictor | _error->userMsg( "..." ); |
| 18:21.53 | hoiji | oh! thanks |
| 18:22.45 | mpictor | http://stepcode.org/doxygen/class_error_descriptor.html |
| 18:23.20 | mpictor | that documentation is pretty old now, but still good for a lot of things |
| 18:23.39 | mpictor | needs to regenerate it at the next release... |
| 18:25.09 | mpictor | I'm not sure if I got error handling working completely or not. if it is, any message added via userMsg() ought to print out |
| 18:25.37 | mpictor | it could be that _error is never checked :/ |
| 18:25.55 | *** join/#stepcode shooter (~shooter@188.190.126.62) |
| 18:32.39 | hoiji | oh! then during testing, I will check _error by keeping the value of instanceIDLength very low once. |
| 18:35.17 | hoiji | Also thanks for the doxygen link |
| 19:25.55 | hoiji | sectionReader::_error was never instantiated, and hence kept running into seg. faults. Once I modified the sectionReader constructor everything seems to be working now |
| 19:41.20 | hoiji | mpictor: I have commited the changes, do take a look.. Thanks |
| 19:41.22 | hoiji | https://github.com/hoiji09/stepcode/compare/stepcode:master...master |