| Latest revision |
Your text |
| Line 95: |
Line 95: |
| | | | |
| | 14 June: prepared for tomorrow's exam, no work for GSoC | | 14 June: prepared for tomorrow's exam, no work for GSoC |
| − |
| |
| − | '''Week 4'''
| |
| − |
| |
| − | 15 June: been at school for exam almost all day, did some small research on lemon grammar
| |
| − |
| |
| − | 16 June: started reworking on group() parsing, it seems that the latest patch didn't applied cleanly
| |
| − |
| |
| − | 17 June: sent the new patch which applies very well on my clean checkout; did some more work with the group_content parsing
| |
| − |
| |
| − | 18 June: further work on group_content parsing
| |
| − |
| |
| − | 19 June: finished the last exam, more work on group_content parsing, resolved some of the errors, hope to come up with a valid patch soon
| |
| − |
| |
| − | 20 June: finalised group_content and group_end parsing, made a [https://sourceforge.net/p/brlcad/patches/383/ patch], tested it on a clean checkout, posted it on sourceforge and sent a mail to the list with project updates
| |
| − |
| |
| − | 21 June: last patch got committed by Cliff, added tokens for all the functions/methods that can appear in a .csg file, started working on grammar for union() and difference()
| |
| − |
| |
| − | '''Week 5 (pre-Midterm)'''
| |
| − |
| |
| − | 22 June: further work on the functions token, trying to expand the regex for group_content to cover all the functions formats in a .csg file and also trying to expand the grammar
| |
| − |
| |
| − | 23 June: made token for parsing dimension parameters for cube(), sphere(), cylinder() and polyhendron(), linked it with the existing grammar
| |
| − |
| |
| − | 24 June: designed non-terminals for assignation of parameters ( for example for parsing things like "convexity = 1" or "size = [3,4,5]")
| |
| − |
| |
| − | 25 June: heavy work on the csg grammar, lots of tests, working on edge cases, almost done & ready for a patch submission
| |
| − |
| |
| − | == Midterm Evaluation Period ==
| |
| − |
| |
| − | At the beginning of the coding period I haven't had anything related to lemon and re2c integrated into BRL-CAD. I decided to take some time to study how to use this parsing tools before getting to the actual writing of the code. Because of my summer exam session, my progress was little in the first few weeks and, after working out some simple calculators using lemon, I finally started working on the .csg grammar. Initially, I though I could gradually parse and interpret functions and methods but I soon realised that having a fully functional grammar first is a better idea. With lots of help from Cliff and Andrei Constantin Popescu, I managed to finish the grammar just 2 days after the Midterm Evaluation Period opening and have it merged.
| |
| − |
| |
| − | To run my code, just go in build/src/conv/csg , write an input file with some .csg text in it (or save a .csg there) and then just run ./csg input.txt output.txt . It should print on the screen various messages containing the tokens matched in the parsing process.
| |
| − |
| |
| − | 26 June: sent a [https://sourceforge.net/p/brlcad/patches/386/ patch] regarding the .csg grammar which is almost complete , needs some work in the priorities section to be just right
| |
| − |
| |
| − | 27 June: filled midterm form on melange, worked on priorities
| |
| − |
| |
| − | 28 June: sorted out the priorities section, changed some of the internal structure of the grammar to parse imbricated functions better, added a new rule for assignation, made a [https://sourceforge.net/p/brlcad/patches/388/ new patch] with the complete grammar and sent a mail to the list with the latest updates
| |
| − |
| |
| − | 29 June: talked with teepee from OpenSCAD about the conversion from .scad to .csg, found out how to properly test the proposed grammar (running the OpenSCAD test suite) and also got some insight about the functions that generate the .csg from the .scad ( toString for [https://github.com/openscad/openscad/blob/master/src/primitives.cc#L610 primitives] and [https://github.com/openscad/openscad/blob/master/src/transform.cc#L190 special functions] )
| |
| − |
| |
| − | 30 June: did nothing for GSoC
| |
| − |
| |
| − | 1 July: tested my grammar using some .csg files provided by OpenSCAD, modified it to also parse "group();" , not only "group() {...}", added DEBUG macro and tried to print some matched functions using bu_vls_addr(&A->value) but with no luck
| |
| − |
| |
| − | 2 July: further work on printing matched functions, renamed some tokens in the grammar for a clearer code
| |
| − |
| |
| − | == Coding Period (2nd half) ==
| |
| − |
| |
| − | 3-6 July : had a mental breakdown because of the lack of progress and sleep in the last few days, took some time to organise my days and get back on track with everything
| |
| − |
| |
| − | '''Week 6'''
| |
| − |
| |
| − | 7 July : restarted reading dom2dox code to learn about interpreting lemon grammar, understood some basic principles
| |
| − |
| |
| − | 8 July : cleared some confusions about the "END_TEXT;" macro after a discussion with Sean on IRC, decided not to use macros to prevent complexity issues
| |
| − |
| |
| − | 9 July : found the /doc/parsers/writing_perplex_lemon_parsers.doc and also the templates in there and read through them , they seem well explained and detailed
| |
| − |
| |
| − | 10-11-12 July : not much progress, tried a different approach by modifying the grammar to match functions as 1 token to reduce the number of terminals, ended up not parsing the function parameters
| |
| − |
| |
| − | '''Week 7'''
| |
| − |
| |
| − | did nothing for GSoC
| |
| − |
| |
| − | '''Week 8'''
| |
| − |
| |
| − | 20 July : got Sean's mail about communicating more and staying more on IRC, tried to install tmux to be 24/7 on IRC but failed due to some incompatibility issues, decided to stick to XChat
| |
| − |
| |
| − | 21-23 July : documented about perplex, re2c and lemon in order to understand better what is not working and why and asking the right questions
| |
| − |
| |
| − | 24 July : sent an email to Sean and Isaac with questions about why fprintf(appData->outfile, "%s", bu_vls_addr(&A->value)) does not print anything but the token is matched, about the generated tree and also discussed updating deliverables
| |
| − |
| |
| − | 25 July : more work on sending matched strings to main, restrained the problem to a memory allocation issue
| |
