Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Jun 23 01:47

    patrickhuber on nuget

    (compare)

  • Jun 23 01:47

    patrickhuber on main

    Bump Newtonsoft.Json in /tests/… Merge pull request #110 from pa… (compare)

  • Jun 23 01:47
    patrickhuber closed #110
  • Jun 22 17:12
    dependabot[bot] labeled #110
  • Jun 22 17:12
    dependabot[bot] opened #110
  • Jun 22 17:12

    dependabot[bot] on nuget

    Bump Newtonsoft.Json in /tests/… (compare)

  • Jun 13 17:53

    patrickhuber on main

    updates sample with ast and int… (compare)

  • Jun 13 16:28

    patrickhuber on v0.10.0

    (compare)

  • Jun 13 16:25

    patrickhuber on main

    updates pipeline to use proper … (compare)

  • Jun 13 16:05

    patrickhuber on main

    updates version string (compare)

  • Jun 13 15:58

    patrickhuber on main

    updates artifact paths to relat… (compare)

  • Jun 13 15:55

    patrickhuber on v0.10.0

    (compare)

  • Jun 13 15:54

    patrickhuber on main

    updates build workflow (compare)

  • Jun 13 15:15

    patrickhuber on v0.10.0

    (compare)

  • Jun 13 15:14

    patrickhuber on v0.10.0

    (compare)

  • Jun 13 15:09

    patrickhuber on main

    updates artifact path (compare)

  • Jun 13 15:05

    patrickhuber on v0.10.0

    (compare)

  • Jun 13 15:03

    patrickhuber on master

    (compare)

  • Jun 13 15:01

    patrickhuber on main

    (compare)

  • Jun 13 14:25

    patrickhuber on master

    changes IAndNode to IPackedNode… (compare)

Patrick Huber
@patrickhuber
Once you have the current root, a depth first search should give you the forest. It will appear as a graph not a tree because there can be loops.
Same with the Parse Tree
graph not a 'hierarchy'
bilsaboob
@bilsaboob
Thanks for the input, I'll look into this tomorrow again :), as you say I'm iterating the parse structure for every Pulse... but probably messed up the Msagl graph building part due to lack of insight on the different node types - I started looking into how the ParseTreeEnumerator works
that should probably make things clear
btw, what's the difference between the ITreeNode and IForestNode?
both seem to be very similar - the IForetNode are the actual parse nodes that are created during the parse
then you do transformation to ITreeNode during the ParseTreeEnumerator enumeration
is it in order to "hide" some "internals" from the parse phase?
Patrick Huber
@patrickhuber
tree node is unambiguous and the forest node is ambiguous
the and nodes represent each ambiguous branch
when you have ambiguity more than one and node will exist at the site of the ambiguity
i implemented a couple traversal algorithms, need to check if the parse tree enumerator has any bugs
bilsaboob
@bilsaboob
ahaa, thanks, that should make things clear :)
Patrick Huber
@patrickhuber
not sure if you saw issue #66 but ArsenShnurkov needs unicode support and ebnf text parsing.
bilsaboob
@bilsaboob
no, I just noticed :)
though I don't like the Ebnf syntax at all :/ - thats why I started on a custom Bnf like syntax closer to the one in Antlr
I find that one much easier to follow / understand
the RapidBnf in my RapidPliant repo
part of that one I want to have things such as "parameterized" rules
for things similar to "list" and what not
though the Unicode support would mean the lexing would need to be extended to support full unicode set
Patrick Huber
@patrickhuber
I think the current implementation supports unicode, but you have to use code style grammars
I could be wrong on that though
bilsaboob
@bilsaboob
aha, I see, thats the issue
Patrick Huber
@patrickhuber
I remember your bug fix had escape syntax but I don't think we supported \uXXXX
bilsaboob
@bilsaboob
cool
no, it didn't - it was only for the default character classes
and it was for the regex parser, is that what you use in the Ebnf?
Patrick Huber
@patrickhuber
good point
I think so
I put the regex grammar in the ebnf grammar
bilsaboob
@bilsaboob
cool
well, If you want I'll look into that later today
I'm using the Regex grammar too ;)
though it will most likely require rewriting the grammar?
i I remember correctly it was designed to only take a single char after the "\"
on grammar level
Patrick Huber
@patrickhuber
I think the AST classes take one character, but because a char in .NET is unicode 16 it should be able to store the value
like you said the grammar may have some issues
EscapeSequence lex rule is the one I think needs to be looked at
and its just a dfa
bilsaboob
@bilsaboob
aha, ok, so just find a regex for unicode sequence and create the dfa "manually" right?
Patrick Huber
@patrickhuber
yep, that should be it
Patrick Huber
@patrickhuber
Think the MarpaParseEngine optimization worked
I finally found the bug in my implementation
214 ms to parse the large json file
the benchmark was even better @ 186.244 ms
bilsaboob
@bilsaboob
Hey, cool :) - I'm finally having Christmas vacation - will check out what you've done there, but having it below 200ms sounds pretty good :)
osipenkokat
@osipenkokat
Hello! Can you please tell me, I want to use this project here: https://github.com/uxmal/pytocs But I don't know how to insert a project in Python. If anyone has worked with this please show.
1 reply