$ geany "$d_Qroot""/0_QNial notes.txt" 0_Qnial bug hunt.txt - for debugging setup work! https://github.com/danlm/QNial7 ******** show my QNial symbols : qnial> post EACH first symbols 0 ******** to add loaddef comments as each [expression (IS), operator(IS OP)] is loaded : see 'link d_Qnial_root 'setup.ndf' qnial> add_ISOP_debug (link d_Qndfs 'filename') the operator must be preceded by a comment line of standard format the comment in turn provides lightning fast Unix cmd extraction of comments (eg lists of operators) ******** Find words (symbols) -w = word, -i = case insensitive $ find "$d_webRaw" -maxdepth 3 -name "*.html" | tr \\n \\0 | xargs -0 -IFILE grep -w -i --with-filename --line-number ':&file-insert &:' "FILE" | sed 's#:.*##' | sort -u +--+ $ find "$d_Qndfs" -maxdepth 3 -name "*.ndf" | tr \\n \\0 | xargs -0 -IFILE grep -w -i --with-filename --line-number 'd_webDone' "FILE" | sed 's#:.*##' | sort -u +--+ Picking the right "sed separator", example str_replaceIn_path : chr_sed := first ((`/ `# `! `| `@ EACHLEFT subStr_in_str (link strOld strNew)) sublist (`/ `# `! `| `@) ) ; ******** single path : IF (NOT path_exists "r path) THEN EACH write '?path_backupTo_dir error, file unknown : ' path '' ; ELSE file_backupDated_retain path ; multi-paths : IF (NOT AND (EACH path_exists ("r path) ("w path))) THEN EACH write '?path_backupTo_dir error, file unknown, one of : ' path path '' ; ELSE file_backupDated_retain path ; #+-----+ Interim changes to symbols : operators : "mark" the symNew IS OP, for ease of getting back to it [Ctrl->, Ctl-<] in geany sym??Old IS symNew Note the "??" to make sure that symOld isn't overwritten document replace symOld symNew symOld IS symNew remove the "??" variables : sym???Old := symNew ; document replace symOld symNew CAREFUL - that symOld isn't part of another symbol! symOld IS symNew remove the "??" ******** Linux : launch leafpad from filemanager, using command : leafpad --tab-width=3 %f ******** # IF flag_break THEN BREAK ; ENDIF ; ******** Change : +.....+ +.....+ To : +.....+ +.....+ ******** # 25Apr2018 SPECIAL NOTE!! : [syms_inParens, symExtract_strin] mutually call each other, which is new for me! # It should lock up in a loop or something, but seems to be working. ******** at [start,end] on .ndf files : f_strings := 'strings.ndf' ; loaddefs_start f_strings ; loaddefs_ended f_strings ; ******** Short form of types (want 3 letters, accept 4 for some cases, 5+ rarely?) : arry nested rectangular array dir directory (full directory) fHand file number mostly use shorthand eg 'finn := open path "r' fname file name without directory int integer real real str string path path = fullDirectory baseName extension, used in operator names pinn, pout, ... common shorthand where multiple paths involved pname archaic version used in [IS OP, Tran] arguments word word IMPORTANT! Often, string operations will damage [data, files, directories] because they will result in sub-string substitutions that destroy the original data. Working with words, as defined by either [white, word] character sets, is (hopefully) far more reliable for critical applications. Computer programming is just one great example. ********************* webSite [menuHeadFoot, link, TableOfContents, link] tools.html http://www.billhowell.ca/Software programming & code/Qnial/webSite [menuHeadFoot, link, TableOfContents, link] tools.html ******** 21Oct2020 I need to change ALL : +-----+ WHILE (~= ??eof (line := readfile finn)) DO +-----+ To : +-----+ WHILE (NOT isfault (line := readfile finn)) DO +-----+ ******** 18Oct2020 qnial> add_ISOP_debug (link d_Qroot 'setup.ndf') >> OK, symbols cleaned up a bit in [setup, fileops, strings].ndf Now - back to "Process executables embedded in files (eg html files)" see link d_Qndfs 'website urls notes.ndf' ******** 17Oct2020 https://en.wikipedia.org/wiki/XML I should be looking at XML? maybe, when I need it... ******** 16Oct2020 continue with "word versus string" operators in [fileops, strings].ndf ******** 16Oct2020 Add a tiny bit of formatting to .ndf files Titles : #] #] ********************* #] loaddefs link d_Qndfs 'sortup_cull_listCollection- tests.ndf' #] loaddefs link d_Qtests 'sortup_cull_listCollection- tests.ndf' #] loaddefs link d_bin 'sortup_cull_listCollection- tests.ndf' within a file, to searate sections : #] +-------------------+ #] DEBUGGING operators ******** 15Oct2020 15Oct2020 - oops - I chopped chrs_symbols_matlab - fix later see link d_Qndfs 'QNial symbol changes.ndf' ******** 21Sep2020 emto Michael Jenkins, John Gibbons thanks for making QNial64 available!! Bill Howell's home page : http://www.BillHowell.ca/ QNial root - includes setup files versions of QNial, easy directory listing http://www.BillHowell.ca/Qnial/ QNial programs - includes my own programs and the standard QNial library (sometimes modified) http://www.BillHowell.ca/Qnial/MY_NDFS/ list of QNial [data, op, prgm] files - simple text file listing, mostly complete as of 23Sep2020 /media/bill/PROJECTS/Qnial/MY_NDFS/0_list of Howells QNial programs.txt Linux bash scripting : http://www.BillHowell.ca/bin/ GNUplot scripts (new) - as I've just started using gnuplot (should have done so decades ago!) there isn't much yet http://www.BillHowell.ca/gnuplot/ ************ 27Mar2020 'trig [now, cum, seq, tim, mix] tests.ndf' $ d_tests="/home/bill/PROJECTS/Qnial/MY_NDFS/MindCode/code test/" $ diff "$d_tests""MindCode [time, scalars] tests.ndf" "$d_tests""convert [bits, intg, real] tests.ndf" +-----+ $ diff "$d_tests""MindCode [time, scalars] tests.ndf" "$d_tests""convert [bits, intg, real] tests.ndf" 1c1 < # loaddefs link d_QNial_mine 'MindCode/code test/MindCode [time, scalars] tests.ndf' --- > # loaddefs link d_QNial_mine 'MindCode/code test/convert [bits, intg, real] tests.ndf' 3d2 < # header - [architectures, rules] for MindCode neuron - with [DNA-mRNA-epi, addressable synapses] 5c4 < # /home/bill/PROJECTS/Qnial/MY_NDFS/MindCode/code test/MindCode [time, scalars] tests.ndf --- > # /home/bill/PROJECTS/Qnial/MY_NDFS/MindCode/code test/convert [bits, intg, real] tests.ndf 12a12,15 > f_convert_tests := 'convert [bits, intg, real] tests.ndf' ; > loaddefs_start f_convert_tests ; > > 17c20 < # p_log is [open, close]ed for writing in this file --- > # p_log is [open, close]ed for writing in that file 31c34 < writefile p_log (link 'tests of : ' d_QNial_mine 'MindCode/MindCode neuron - [time, scalars].ndf') ; --- > writefile p_log (link 'tests of : ' d_QNial_mine 'MindCode/code test/convert [bits, intg, real] tests.ndf') ; 175c178 < --- > loaddefs_ended f_convert_tests ; +-----+ >> it's OK, nothing lost, just an update ************ 22Feb2020 int_to_bit routines $ grep -n "byte" "/home/bill/PROJECTS/Qnial/" ************* 06Feb2020 build extended QNial7 on Suse from : /run/media/bill/HOWELL_BASE/Software installations & maintenance/QNial7 to Suse/BuildNial/README.md OK, Installed. Fairly [easy, straightforward]. Several packages wouldn't make due to missing headers or pointer problems. Feature Capability (# denotes couldn't add, missing header file, etc. See process below) DEBUGINCLUDED adds DEBUG capability to basic nial MEMSPACES supports shared memory spaces (Linux/OSX) NCOMPLEX complex arithmetic package NDYNLOAD supports dynamic loading in Linux and OSXNFILES | simple file features NIALDSP digital signal processing support #NIAL_FFTW fast fourier transform support #NSFML_AUDIO Audio support NTABLES user hash tables for associative array features #PROCESS Unix style process control (Linux/OSX) SOCKETS Unix/Windows socket support QSORT Quicksort algorithm REGEXP Posix regular expression support (Linux/OSX) SPROCESS support for streams (Linux/OSX) #WINPROCESS process support for Windows (Windows) Modified ~/bin/qnial bash script. $ qnial /home/bill/bin/qnial: line 27: /home/bill/PROJECTS/Qnial/nial7_Suse64 -i -defs "/home/bill/PROJECTS/Qnial/setup.ndf": No such file or directory >> YIKES!!! Stupid QNial doesn't accept quotation marks??? >> Probably stupid Bill Howell for not providing final form? For now, luckily I can remove the quotation marks - maybe not for RaspPi. ********** 25Jan2020 QNial Language Definition Manual Invoking Q'Nial3Invoking Q'Nial There are a number of command line options in invoking the console version of Q'Nial which provide flexibility for its use in applications. A Q'Nial session is invoked by the command: nial It has the following syntax: nial [Wsname] [-size Wssize] [-defs Filename] [-q] [-s] [-b] [-m] [-d] [-h] -size Wssize -defs Filename -q The quiet option suppresses the banner and other default output. -s This option suppresses the trigger mechanism that interrupts computation whenever a fault is created. -b This option changes the box characters used for output display -m This option suppresses the warning messages that advise the user when the workspace expands and other internal tables increase in size. -d This option turns off the default debugging and tracing capabilities. -h display the help information for invoking Q?Nial. So only the workspace is specified, not the stack? That contradicts +-----+ Workspace Status The expression Status provides information about the workspace. Its results are explained as follows: Position Value Value Given Meaning 0 number of free words amount of space available in workspace 1 largest free block size determines largest object that can be built 2 number of free blocks indicates the fragmentation of memory 3 workspace size gives the current size of the workspace 4 size of internal stack used to hold intermediate values in evaluation 5 size of phrase/fault table area used for hash table for phrases and faults 6 internal buffer size used for temporary space by many operations The last three items indicate the size of three internal areas that grow if required, but never shrink in size. After some large computations these components may be larger than needed. They are reset if the workspace saved and reloaded. ********* 15Jan2020 WCCI2020 HELP system loaddefs link d_QNial_mine 'email - split HELPER_WCCI2020.ndf' +-----+ >>> ?local in nonlocal list 16Jan2020 I don't understand the problem? seems well-declared { LOCAL filters filters_rows details ; NONLOCAL d_log d_archive p_log d_fldrSrc f_Src p_fldrSrc p_fldrTwix p_fldrTmp f_stats p_stats email_filter_phrase mime_expand_flag massEmailRemove_flag folders keyStrsEachFldr keyStrsList TosList CCsList filterFolder_list p_folders keyStrsALL ; % ; qnial> loaddefs link d_QNial_mine 'email - split HELPER_WCCI2020.ndf' ?local in nonlocal list: KEYSTRSLIST TOSLIST CCSLIST <***> ; errors found: 1 >> OK, I had repeated a variable in both [LOCAL, NONLOCAL] but couldn't see it (brain dead) ************************ 17May2018 ?undefined identifier: % ; symTranslate_file_test IS { LOCAL d_Lucas test_name pname_diff pname_inn pname_out ; NONLOCAL flag_test_comment_log ; % ; test_name := 'symTranslate_file_test' ; IF flag_test_comment_log THEN open_test_comment_logFile test_name ; ENDIF ; EACH write_testStr ' ' '+---------------+' test_name ; write_testStr 'symTranslate_string translates "simple symbols" and "composite expressions"' ; % ; d_Lucas := '/media/bill/PROJECTS/Lucas - Universal Force/' ; pname_inn := link d_Lucas 'Howell - math of Lucas Universal Force 180418.txt' ; pname_out := link d_QNial_temp 'Howell - math of Lucas Universal Force ' timestamp_YYMMDD '.txt' ; pname_diff := link d_QNial_temp 'trans_file_diff.txt' ; % ; write_testStr '# symTranslate_file_test example 1: compare test results in : ' ; write_testStr link ' pname_inn= ' pname_inn ; write_testStr link ' pname_out= ' pname_out ; write_testStr link ' pname_dif= ' pname_diff ; write_testStr link ' (ideally "' pname_diff '" will be empty except for "good" substitution lines)' ; % ; symTranslate_file pname_inn pname_out ; write_testStr timestamp ; % ; IF flag_test_comment_log THEN close_test_comment_logFile ; ENDIF ; } >> This gives the error : qnial> loaddefs '/media/bill/PROJECTS/Qnial/code tests/symbols translate- tests.ndf' >>> loading start : symbols translate- tests.ndf >>>>>> loading start : strings.ndf >>>>>>>>> loading start : file_ops.ndf <<<<<<<<< loading ended : file_ops.ndf <<<<<< loading ended : strings.ndf >>>>>> loading start : symbols translate.ndf >>>>>>>>> loading start : symbols stdForm.ndf <<<<<<<<< loading ended : symbols stdForm.ndf >>>>>>>>> loading start : symbols extract & sort.ndf <<<<<<<<< loading ended : symbols extract & sort.ndf >>>>>>>>> loading start : tranList Lucas bads.ndf <<<<<<<<< loading ended : tranList Lucas bads.ndf >>>>>>>>> loading start : tranList Lucas to HLFN.ndf <<<<<<<<< loading ended : tranList Lucas to HLFN.ndf >>>>>>>>> loading start : tranList HLFN bads.ndf <<<<<<<<< loading ended : tranList HLFN bads.ndf <<<<<< loading ended : symbols translate.ndf loading symTranslate_file_test ?undefined identifier: % ; loading symTranslate_file_test - old tests <<< loading ended : symbols translate- tests.ndf errors found: 1 However, I can't see ANY problems with the code Removing the "% ;" lines just shift the error to the line before or after Problem was symTran_file was defined, not symTranslate_file ******************* # 18Apr2018 YIKES! non-standard ASCII codes for QNial? qnial> EACH charrep 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 >> OK qnial> shape EACH charrep 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 26 qnial> shape 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 26 qnial> EACH charrep 'abcdefghijklmnopqrstuvwxyz' 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 >> WRONG!! qnial> shape 'abcdefghijklmnopqrstuvwxyz' 26 qnial> shape EACH charrep 'abcdefghijklmnopqrstuvwxyz' 26 qnial> EACH charrep 'αβγδεζηθικλμνξοπρςστυφχψω' -50 -79 -50 -78 -50 -77 -50 -76 -50 -75 -50 -74 -50 -73 -50 -72 -50 -71 -50 -70 -50 -69 -50 -68 -50 -67 -50 -66 -50 -65 -49 -128 -49 -127 -49 -126 -49 -125 -49 -124 -49 -123 -49 -122 -49 -121 -49 -120 -49 -119 qnial> shape 'αβγδεζηθικλμνξοπρςστυφχψω' 50 qnial> shape EACH charrep 'αβγδεζηθικλμνξοπρςστυφχψω' 50 qnial> EACH char (tell 256) !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklm nopqrstuvwxyz{|}~������������������������������������������������������������������������������������������ ������������������������������������ >> Something is wrong with QNial charrep operator!! - small letters give wrong answer! but char seems correct >> For now, fudge it. ******************* 13Apr2018 QNial startup failure - is this related? $ qnial Q'Nial V7.0 Open Source Edition Intel x86 64bit Linux Jul 6 2017 Copyright (c) NIAL Systems Limited Segmentation fault RaspPi - out of memory? I shut down some LibreCalc files and qnial worked >> not good enough I cold rebooted RaspPi - now qnial launches ************************* 11Apr2018 Problem with : ?undefined identifier: % ; I've often had this problem in the past, but can't remember the [cause, remedy] Removing this "spacer line" doesn't help - the next line then becomes the problem. This time, it occurred for two operators in "translate symbols.ndf", one of which is : translate_file IS OP pname_inn pname_prep pname_out { LOCAL fin fot i line_in n_fLine textnew ; NONLOCAL sym_chars trans_old trans_new trans_msg trans_rows ; % ; sed_prep_file pname_inn pname_prep ; fin := open pname_prep "r ; fot := open pname_out "w ; % ; n_fLine := 0 ; WHILE (~= ??eof (line_in := readfile fin)) DO textnew := line_in ; write link (string n_fLine) 'a ' textnew ; FOR i WITH (tell trans_rows) DO sed_expr := link '"s/\([^' sym_chars ']\)\(' trans_old@i '\)\([^' sym_chars ']\)/\1' trans_new@i '\3/g"' ; IF (< i 2) THEN write link 'sed_expr : ' sed_expr ; ENDIF ; IF (< i 2) THEN write link 'FOR 1= ' textnew ; ENDIF ; textnew := translate_sed sed_expr textnew ; IF (< i 2) THEN write link 'FOR 2= ' textnew ; ENDIF ; ENDFOR ; write link (string n_fLine) 'b ' textnew ; writefile fot textnew ; n_fLine := n_fLine + 1 ; ENDWHILE ; % ; EACH close fin fot ; } This had loadded many times without problems over the last week, so why now the problem? Related to same problem : ?undefined identifier: % ; ?undefined identifier: 'trans_long_out.txt' ; TRANSLATE_FILE_LONG <***> PNAME_DIFF PNAME_INN >> bye QNial, then ree-starting doeswn't help. see "/media/bill/PROJECTS/Qnial/MY_NDFS/translate symbols - coding development.txt" 13Apr2018 # QNial file "/media/bill/PROJECTS/Qnial/MY_NDFS/strings.ndf" # 13Apr2018 Example for project "Lucas Universal Force" # find_chrs_special_file flag_chrs_include chrs_set fname_in find_chrs_special_fTest IS find_chrs_special_file o chrs_writing '/media/bill/PROJECTS/Qnial/MY_NDFS/translate symbols.ndf' >> empty return, which confirms that no "strange or illegal" characters are corrupting the looadefs of "/media/bill/PROJECTS/Qnial/MY_NDFS/translate symbols.ndf" >> however, the Grrek character epsilon (UNICODE) should have failed? 13Apr2018 14:22 Now what??? Maybe the Greek symbols [θ,φ] are causing the trouble? But θ works OK with sed and sed_file? >> Actually, I haven't checked that !!! BUT - earlier loaddefs were OK even with theta in the symbols-to-be-replaced. qnial> loaddefs link d_Qnial_mine 'translate symbols.ndf' >>>>> loading translate symbols.ndf test9 ?undefined identifier: % ; test99 ?undefined identifier: 'trans_file_out.txt' ; TRANSLATE_FILE <***> PNAME_INN PNAME_PREP ?undefined identifier: % ; ?undefined identifier: 'trans_long_out.txt' ; TRANSLATE_FILE_LONG <***> PNAME_DIFF PNAME_INN ?undefined identifier: 'translate symsCheck Lucas.txt' ; TRANSLATE_FILE <***> FNAME_INN <<<<< done loading translate symbols.ndf errors found: 5 >> same problem! Try with original copy of QNial7_RapsPi >> both old & new versions look the same, swap anyways ************************* 14Feb2018 Cull emails from NIPS2017 papers ************************* # 12Feb2018 problems Segmentation fault (see "/media/bill/HOWELL_BASE/Qnial/0_bug hunt QNial.txt") arises within strings_between after several (15-20?) lines have been read note that the failure point varies strangely, so this doesn't seem to be specific to a specific line of fin ??!?!? At one point, the full paper list was written to screen, but a code change and loaddefs generated a "Segmentation fault"???!!? So perhaps : works : write (strings_between URLpaper ender line) ; fails : %fotPapers EACHRIGHT writefile (strings_between URLpaper ender line) ; NYET! just tested without writefile - failed after only 8 lines! # One example froze at : 6642-universal-style-transfer-via-feature-tran ����� Why the stange characters? When I added "write line ; " the line printed and processed, failure occurred later @ paper 6650 ??? Switching to writefile, this time got to paper 6742 Try this on RaspPi ... and Lenovo, and after Dell64 cold boot 13Feb2018 Lenovo - segmentation fault at paper 6687-compression-aware-training-of-deep-networks OK - so something is BAD WRONG with strings_between maybe I have "clean" strings of unseen characters? added write -> got to 6999-mapping-distinct-timescales-of-functional-interactions-among-brain-networks This is much better, but why? retry -> got to same 6999-mapping-distinct-timescales-of-functional-interactions-among-brain-networks Segmentation fault occurs AFTER strings_between output & BEFORE strings_between write i between_first between_last ; Segmentation fault occurs AFTER strings_between output & BEFORE strings_between 1st FOR Segmentation fault occurs AFTER write 'strings_between - before fronts & BEFORE write link link 'strings_between: fronts=' Problem seems to occur with : fronts rears := strStrt strEndr EACHLEFT find_string strIn ; for rest of testing, see "/media/bill/PROJECTS/Qnial/MY_NDFS/strings.ndf" unicode_to_ascii '/media/bill/PROJECTS/2017 NIPS Long Beach, California/0_NIPS 2017 URLs all raw.txt' '/media/bill/PROJECTS/2017 NIPS Long Beach, California/0_NIPS 2017 URLs all.txt' now run - looking at "/media/bill/ramdisk/strings_between strIn.txt"
  • Minimizing a Submodular Function from Samples Eric Balkanski, Yaron Singer
    • >> NO problem with non-ASCII?? What is going on? For now - just split the html file into 5 parts, and do each separately!! 13Feb2018 note kwrite -> Menu -> Open/Save -> General -> Line Length limit = 4096 why can't I search for escape sequences and regexpr with this file? QNial "bye" doesn't work after running one of these - even if the file is properly processed! lq_confPapers Split the html file into 5 parts (1524 lines / 5 ~ 305 lines each), and name as : ; "/media/bill/PROJECTS/2017 NIPS Long Beach, California/0_NIPS 2017 Proceedings page-1.txt" etc ; extract_paperURLs '0_NIPS 2017 Proceedings page-1' Starts at 6606-wider-and-deeper-cheaper-and-faster-tensorized-lstms-for-sequence-learning This is the FIRST paper in the html file
  • Dykstra's Algorithm, ADMM, and Coordinate Descent: Connections, Insights, and Extensions Ryan J. Tibshirani
    • Remove this record and try again - ends at : 6727-shape-and-material-from-sound This is the last paper in page-1 extract_paperURLs '0_NIPS 2017 Proceedings page-2' Starts at 6728-flexible-statistical-inference-for-mechanistic-models-of-neural-dynamics Crashes at
  • Flexible statistical inference for mechanistic models of neural dynamics Jan-Matthis Lueckmann, Pedro J. Goncalves, Giacomo Bassetto, Kaan Öcal, Marcel Nonnenmacher, Jakob H. Macke
    • Remove this record and try again - does a lot more, crashes at :
  • Data-Efficient Reinforcement Learning in Continuous State-Action Gaussian-POMDPs Rowan McAllister, Carl Edward Rasmussen
    • Remove this record and try again - does a lot more, crashes at :
  • Noise-Tolerant Interactive Learning Using Pairwise Comparisons Yichong Xu, Hongyang Zhang, Aarti Singh, Artur Dubrawski, Kyle Miller
    • Remove this record and try again - does a lot more, crashes at :
  • Analyzing Hidden Representations in End-to-End Automatic Speech Recognition Systems Yonatan Belinkov, James Glass
    • Remove this record and try again - does a lot more, crashes at :
  • Generative Local Metric Learning for Kernel Regression Yung-Kyun Noh, Masashi Sugiyama, Kee-Eung Kim, Frank Park, Daniel D. Lee
    • So the failures are now sequential - WHY??? Split the file here, label end part as '0_NIPS 2017 Proceedings page-2a', add back 6837 & 6838 extract_paperURLs '0_NIPS 2017 Proceedings page-2a' Starts at 6837-noise-tolerant-interactive-learning-using-pairwise-comparisons Crashes at
  • Ensemble Sampling Xiuyuan Lu, Benjamin Van Roy
    • Remove this record and try again - does a lot more, crashes at :
  • Ensemble Sampling Xiuyuan Lu, Benjamin Van Roy
    • Remove this record and try again - ends at # enddoc Output starts at 6839-generative-local-metric-learning-for-kernel-regression (missing [6837,6838] !!!???) Output ends at 6930-a-universal-analysis-of-large-scale-regularized-least-squares-solutions This is contiguous with '0_NIPS 2017 Proceedings page-3' extract_paperURLs '0_NIPS 2017 Proceedings page-3' Starts at 6931-deep-sets Empty ramdisk file !!??, last success 6957-high-order-attention-models-for-visual-question-answering Retry - Completed on second attempt, "# enddoc" left in file WHY!??!? - maybe it worked on first try? Ends at 7081-dropoutnet-addressing-cold-start-in-recommender-systems extract_paperURLs '0_NIPS 2017 Proceedings page-4' Starts at 7082-a-simple-neural-network-module-for-relational-reasoning Completed at first attempt, "# enddoc" left in file Ends at 7233-towards-generalization-and-simplicity-in-continuous-control extract_paperURLs '0_NIPS 2017 Proceedings page-5' Starts at 7234-random-projection-filter-bank-for-time-series-data Completed at first attempt, "# enddoc" left in file Ends at 7284-multi-view-matrix-factorization-for-linear-dynamical-system-estimation This is the LAST paper listed on the web-age Numbers [3,4,5] are contiguous Finalisation of full list $ cd "/media/bill/PROJECTS/2017 NIPS Long Beach, California/" $ cat "0_NIPS 2017 Proceedings page-1 papers.txt" "0_NIPS 2017 Proceedings page-2 papers.txt" "0_NIPS 2017 Proceedings page-2a papers.txt" "0_NIPS 2017 Proceedings page-3 papers.txt" "0_NIPS 2017 Proceedings page-4 papers.txt" "0_NIPS 2017 Proceedings page-5 papers.txt" >"0_NIPS 2017 URLs all.txt" Then manually add papers that were removed above Total lines in file = 679 compared to 7284 - 6606 + 1 = 679 OK, I've got them all!! WHY was this such a problem? - "bye" is corrupted - "misses" aren't always reproducible, etc etc - illegal characters? $ cd "/media/bill/PROJECTS/2017 NIPS Long Beach, California/" $ grep "\0" "0_NIPS 2017 Proceedings page.html" many (say 1/2) of lines have nulls ************** 16Aug2017 21:03 DesktopSetup LibreCalc spreadsheets [0_contacts,0_References].ods were messed up (appear on screen but couldn't activate or control - I closed them). Also [ToDos,Calendar].ods were the wrong [size ,posn] Possilbly I need to allocate special time delay for the first LibreOffice document open... or just provide a longer, application-specific dealy, as I did with vidProdn ************** 16Aug2017 19:17 DesktopSetup Periodically, I'll have to flush the log file : %comment out for debugging ; IF NOT flag_debug THEN host link 'rm "' (link d_QNial_temp 'USB workspace loads - winList.txt') '"' ; ENDIF ; ************* 15Aug2017 11:59 ?undefined identifier: problem While revamping "/media/bill/HOWELL_BASE/Qnial/MY_NDFS/USB workspace loads.ndf", the following error appeared during loaddefs : ?undefined identifier: IF = ORIENT <***> "landscape THEN S_WIDTH As per 03Sep2016 (see below) this seems to be a QNial quirk. However, it seems to me that ther were specific causes, but I just can't remember where I saw that and what the causes were. % asssign winConfig structure components ; sp_ winConfig "w_computer computer ; sp_ winConfig "w_orientn orientation ; sp_ winConfig "w_utils wConfig_util ; sp_ winConfig "w_text wConfig_text ; sp_ winConfig "w_LibOff wConfig_LO ; sp_ winConfig "w_ToDos wConfigToDos ; sp_ winConfig "w_Calendar wConfigCalendar ; sp_ winConfig "w_MnthCnt wConfigMnthCnt ; >> Probably caused by not having completed the definition of sp_ - so it wasn't in the workspace!???!!! *********** 23Oct2016 redefining IS to IS OP & visa-versa I've seen the solution before Nial Manual p61 erase A Erase the variable or definition specified by the phrase A. Leave the phrase A in the symbol table with its role unchanged but change its value to one of ?no_value, No_expr, no_op or NO_TR depending on the role of A. This operation is used at top level to remove a user defined object from the current workspace. It permits the use of the identifier for a different purpose but not a differ­ent role. It cannot be used to erase a name in a local scope. >> doesn't work to convert IS to IS OP etc *********** 12Sep2016 removes_extract : crashed QNial yet still produced output files (empty for files labelled 3 through 6) was this due to empty files, or an infinite loop? process_file -> did NOT extract all email addresses! Might be due to new code i wrote since mass email to captue >1 email/line? HOWEVER : Bad adresses works fine!??? why? PROBLEM : too short a wait between closing & opening files (same file?) host 'sleep 2s' seemed to solve it ***************** 03Sep2016 is_blankstring isn't working? OOPS! in strings.ndf, there is a mix of charrep and chars!! : changed to : char_tab := string char 9 ; problem was probably that strings.ndf wasn't loading? changed to is_blankstring IS OP Str { AND OR (Str EACHLEFT in blank_chars) } #***************************** 03Sep2016 "?undefined identifier: % [mark, cull, sort] emails ; " I've had this headached several times, and forget the [problem, solution] !! If I remember correctly, it's a Qnial quirk? clearws, re-load >> didn't work I just re-programmed a bit and the problem disappeared (must have been stray character?) #***************************** 05May2016 Moved QNial onto my USB memory stick to make it easy to change between computers (desktop, laptop, other [individual, organisation] nial won't run - insufficient priviledges. Try chown to clearly make it me who owns man chown : -v verbose --recursive sudo -i chown --recursive bill:bill "/media/bill/USB DISK/Qnial/z_Old" >> OK - that worked. +-----+ make "/media/bill/USB DISK/Qnial/nial" executable chmod a+x "/media/bill/USB DISK/Qnial/nial" >> No can do!!!??? Maybe USB memory can't be the "host" of executables? Go back to simply using nial within /home/bill/Qnial #**************************************** 03Nov2015 ?L - logic error Arguments like a list of Booleans cause this? IF lloolo #**************************************** 02May10 Fault trapping within Operators I am using linux's "ls" command to see is a filename being assigned is unique. However, when I try to trap this condition with "isfault", Nial stops processing with the fault! How do I turn off fault-based interrupt, and just keep going? /home/bill/nial/v6testing/0a_notebook.txt Mentions "settrigger": settrigger Check the the fault triggering mechanism can be toggled with this operation. --->> [Howell 30Mar09 - what is this?] /home/bill/nial/v6testing/0a_notebook.txt 7. Command line options OK -s supresses fault triggering THIS WORKED!!: command line nial -s !!!! #**************************************** 28May07 ?missing_op triggered in: 28May07 building Paillard project The right number of parameters with the right values were input. I f forget what the problem was (arguments to a function out of whack). #**************************************** ?assignment to referenced name: 16Feb07 'Laskar - prepsub.ndf' can't do EXTERNALs for variables that are assigned!!! see Telor It seems to me that this is a Q'Nial interpreter error, as I should be able to declare a variable as an EXTERNAL, and yet still assign values or use it at will. Work-around: I had to assign the variables within the Tlor operation that used them. This isn't going to help with more general situations! #**************************************** ?eof encountered 16Feb07 building Laskar project It seems that the host command doesn't work at all. Very invonvenient (eg setting default directory). Work-around: None so far. #**************************************** ?assignment to referenced name: 16Feb07 'Laskar - prepsub.ndf' can't do EXTERNALs for variables that are assigned!!! see Telor It seems to me that this is a Q'Nial interpreter error, as I should be able to declare a variable as an EXTERNAL, and yet still assign values or use it at will. Work-around: I had to assign the variables within the Tlor operation that used them. This isn't going to help with more general situations! #**************************************** ?eof encountered 16Feb07 building Laskar project It seems that the host command doesn't work at all. Very invonvenient (eg setting default directory). Work-around: None so far. #**************************************** ?missing_op triggered in: 28May07 building Paillard project The right number of parameters with the right values were input. I forget what the problem was (arguments to a function out of whack). enddoc