Repository navigation
Conversation
Convert every version 1 .decls and .dtrace test input to the version 2 format, in preparation for dropping support for version 1 files. The converted inputs produce the same output, with two exceptions: * pine_mdiep loses one invariant over a two-dimensional variable, which version 2 cannot represent. * Nine more splitters compile, because the version 2 declarations let the splitter generator find array variables. Also delete test inputs that no test uses, and delete tests/diff-tests/StackAr, which has been excluded from testing (and broken) since 2010.
Daikon no longer reads version 1 .decls and .dtrace files, and reports an error when it encounters one. Remove FileIO.new_decl_format and all code that was used only for version 1 files, including the name-based program point hierarchy (MergeInvariants now uses the declaration-based one) and the remove_post_vars configuration option. Remove the Perl front end (dfepl), which produced version 1 files. Port convertcsv.pl and the cluster analysis scripts to version 2, and remove scripts that processed only version 1 files.
…t-support-old-format
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
….pl unescaping, remove version 1 leftovers * developer.texinfo: release snippets no longer test for the deleted front-end/ directory; example .decls is now in version 2 format; clarify that a file with no decl-version record is read as version 2. * CHANGELOG.md: accurately describe what identifies a version 1 file. * convertcsv.pl: unescapeDecl now handles \n and \r like FileIO.unescape_decl. * Remove lwpp-init.pl (only used with the removed lwpp.pl) and unused version 1 Kvasir goal files for SimpleDisambigStructTest. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…trace, cluster scripts - ppt_perc: never exclude GLOBAL/OBJECT/CLASS ppts (they are parents of included ppts); read ppt names via FileIO.declared_ppt_name, which applies user_mod_ppt_name. - MergeInvariants with a .decls file: create combined exits before init_hierarchy, and clean the ppt map for merging on both paths. - SplitDtrace: handle traces without declarations; pass header records through. - decls-add-cluster.pl: link the cluster variable to its parents. - dtrace-count.pl, dtrace-indent.pl: match declaration records exactly. - Remove simplify_expression/simplify_variable_names and the unused return value of read_decl_version. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…tDtrace comments - FileIO.is_parent_only_ppt_name: use PptName rather than a substring match on "GLOBAL". - read_ppt_decl: reuse declared_ppt_name to parse the ppt name. - SplitDtrace: use new FileIO.is_header_record, which also recognizes comment records. - MergeInvariants: clean ppts for merge only when reading .inv files. - developer manual: there is no longer a Perl front end. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ion 2 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This was referenced Oct 6, 2026
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…tants - extract_vars.pl: exclude variables by rep type rather than by name; drop redundant check - dtrace-count.pl, dtrace-indent.pl, sort-dtrace.pl, dtrace-diff.pl, convertcsv.pl: reject version 1 declarations explicitly - convertcsv.pl: skip constant variables; escape newlines and carriage returns - dtrace-add-cluster.pl: handle EOF in skip_till_next - decls-add-cluster.pl: link cluster variable for all parent relation types - CHANGELOG: note that version 1 input is no longer accepted
… tests only in directories with a Makefile
… update comments, delete version 1 goal file
Graphviz and netpbm were used only to build the dfepl-flow image. Also restore an attributed historical note, format the changelog entry as a bullet, and clarify that the input-language string is arbitrary.
…w: faster ppt_included, share is_declaration_header, sort-dtrace.pl header paragraphs
…iff test * In pine_mdiep, every "[]" variable has a scalar rep-type, so it is a pointer dereference, not an array. Declare it as "p[0]" with "var-kind field [0]", as Kvasir does, rather than as an array. This keeps the 613 doubly-dereferenced variables, and the output matches the version 1 output except for variable names. * Convert and restore tests/diff-tests/StackAr, with ENTER declarations moved before their EXIT declarations, and regenerate its goal files. * tests/diff-tests/Makefile: run every subdirectory, and stop if a subdirectory's make fails. * SplitterFactoryTestUpdater: read each decls file once, with read_data_trace_file, which throws IOException.
… dead Perl references and spellcheck words
…zip output and dropped line; dtrace-add-cluster.pl: do not overwrite input
- util_daikon.pm: add die_if_version_1_decl (checks only a paragraph's first line), is_declaration_paragraph, and skip_till_next - dtrace-count.pl, dtrace-indent.pl, sort-dtrace.pl, dtrace-diff.pl, convertcsv.pl: use the shared helpers - dtrace-diff.pl: skip declaration paragraphs in combined dtrace files; skip constant variables - sort-dtrace.pl: pass through comment paragraphs - decls-add-cluster.pl: link the cluster variable only for "parent" relations, not "user" relations - extract_vars.pl: check for constants before hashcodes; fix null-object assignment; handle end of file without a trailing blank line - convertcsv.pl, extract_vars.pl: replace closures with simpler code - decls2comp.py: implement the no-hashcodes option; remove Lackwit note
… checks - Reject version 1 input in the cluster scripts. - Accept # comments, and comments not followed by a blank line. - Detect version 1 files with CRLF line endings or trailing whitespace. - Detect an existing variable named "cluster". - In dtrace-diff.pl, explain mismatches involving constants. - Move escape_decl, unescape_decl, and a version 2 decls parser to util_daikon.pm. - Classify each record with a single regular expression. - Remove decl-input.
…ords
* util_daikon.pm: new read_record reads one record from a file the way
Daikon does: decl-version, var-comparability, and input-language are a
single line, and leading comments are returned with the record. All
the scripts use it instead of paragraph mode or skip_till_next.
* record_kind matches Daikon's tests ("ppt " and an exact
"ListImplementors" line).
* read_ppt_decls rejects data trace records in a declaration file.
* decls-add-cluster.pl: name the output after the file's basename, and
die if two inputs would map to the same output.
* sort-dtrace.pl: comments stay with the declaration or variable that
follows them.
* extract_vars.pl: exclude variables by name as well as by rep type.
* dtrace-diff.pl: compare variable lists element by element.
* Share the "cluster" variable check as check_no_cluster_var.
Pull request codespecs#856 rewrites runcluster.pl and the scripts it calls (extract_vars.pl, decls-add-cluster.pl, and dtrace-add-cluster.pl), including their support for version 2 files.
…pull requests Pull request codespecs#867 fixes DtraceNonceFixer, including its handling of header records, declarations, compressed files, and samples without nonces. Pull request codespecs#856 rewrites runcluster.pl and the scripts it calls, including their support for version 2 files.
This was referenced Oct 7, 2026
mernst
added a commit
to mernst/daikon
that referenced
this pull request
Oct 7, 2026
Daikon reads version 2 declarations both now and after codespecs#880 removes support for version 1. In version 2, array variables are named "this.a[..]" rather than "this.a[]", so update the goal files and the test's expected values accordingly.
The tests run `perl -Si sort-dtrace.pl FILE`, but -i has no effect on files that a script opens itself, so the sorted output went to standard output and the file was left unsorted. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…te declarations Read records a paragraph at a time, which is faster; end a record only at an empty line, as Daikon does. Share the record loop via for_each_record, which reads through ARGV so that perl -i works. Give ListImplementors its own record kind. Keep indented comments with the following variable in sort-dtrace.pl. Reject unknown keys in program point declarations. Remove the unused check_no_cluster_var.
…arations and escaping * util_daikon.pm: accept \r\n line endings; process paragraphs by offset (linear time); reject a variable declared twice or lacking var-kind, dec-type, or rep-type; build error messages only when needed. * decls2comp.py: list variables with negative comparability on a "-1:" line. * dtrace-diff.pl: read declarations from the traces (decls file optional); escape names in all messages; strip comments with one regex; remove the no-op --ignore_exitno option. * convertcsv.pl: match -decl variables to columns by either form of the name, die if none matches, and write the declared names.
…t-support-old-format # Conflicts: # java/daikon/test/split/SplitterFactoryTestUpdater.java # tests/daikon-tests/pine_mdiep/pine_mdiep.decls.gz # tests/daikon-tests/pine_mdiep/pine_mdiep.dtrace.gz # tests/daikon-tests/pine_mdiep/pine_mdiep.txt-daikon.goal # tests/diff-tests/Makefile # tests/diff-tests/StackAr/Makefile # tests/diff-tests/StackAr/StackAr-1.dtrace.gz # tests/diff-tests/StackAr/StackAr-1m2.txt.goal # tests/diff-tests/StackAr/StackAr-2.dtrace.gz # tests/diff-tests/StackAr/StackAr-2m1.txt.goal # tests/diff-tests/StackAr/StackAr-e.txt.goal # tests/diff-tests/StackAr/StackAr-jt.txt.goal # tests/diff-tests/StackAr/StackAr-t.txt.goal # tests/diff-tests/StackAr/StackAr-u.txt.goal # tests/diff-tests/StackAr/StackAr-union.txt.goal # tests/diff-tests/StackAr/StackAr-xor.txt.goal # tests/diff-tests/StackAr/StackAr.decls # tests/diff-tests/StackAr/StackAr.txt.goal
…-support-old-format # Conflicts: # Makefile # doc/CHANGELOG.md # doc/Makefile
…ont-support-old-format # Conflicts: # doc/CHANGELOG.md # scripts/convertcsv.pl # scripts/dtrace-count.pl # scripts/dtrace-diff.pl # scripts/dtrace-indent.pl # scripts/sort-dtrace.pl
…ading decls files
…arability, truncated records * dtrace-diff.pl reports a difference when two declarations of a program point have different variables, and reuses read_ppt_decls. * sort-dtrace.pl writes the input file's line endings. * decls2comp.py recognizes arrays by their rep-type. * dtrace-count.pl honors a "// EOF" line among a record's leading comments. * util_daikon.pm and dtrace-diff.pl use English without match variables.
…file in dtrace-diff-fast.pl, dead code
… for variables without comparability
Member
Author
|
@coderabbitai review |
|
…-front-end # Conflicts: # doc/CHANGELOG.md
…ion-1-scripts # Conflicts: # doc/CHANGELOG.md
…test-inputs # Conflicts: # tests/diff-tests/Makefile.common # tests/diff-tests/StackAr/Makefile # tests/diff-tests/StackAr/StackAr-u.txt.goal
# Conflicts: # doc/CHANGELOG.md
# Conflicts: # doc/CHANGELOG.md
# Conflicts: # doc/CHANGELOG.md
(cherry picked from commit b694105)
…rmat # Conflicts: # java/daikon/SplitDtrace.java
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR has been split into smaller PRs that can be reviewed and merged independently. This PR includes their changes, so it will shrink as they are merged:
What remains here is the change to Daikon itself (plus the related documentation). This PR overlaps #875, much of which it makes unnecessary.
The changelog entry for version 5.8.6 says "Removed support for the obsolete version 1 file format", but that change only stopped Chicory and DynComp from writing version 1 files. Daikon still read them. Now it doesn't: a version 1 file (one that starts with a
DECLAREorVarComparabilityrecord, or that saysdecl-version 1.0) produces an error that says to re-generate the file with a current front end.Daikon:
FileIO.new_decl_formatand the code that ran only for version 1 files: the version 1 parser, the name-based program point hierarchy, and manyVarInfomethods'VarInfoNamebranches.init_hierarchy_newis renamedinit_hierarchy.MergeInvariantsused the version 1 hierarchy code even for version 2 decls; now it uses the declaration-based code.remove_post_vars. It had an effect only for version 1 files (with version 2 files it failed an assertion).serialVersionUIDofFileIO.SerialFormatandInvMap, whose serialized form changed, so that old.invfiles produce the existing "regenerate your .inv file" message.Front ends and scripts:
convertcsv.plto version 2. I testedconvertcsv.plthrough Daikon. The cluster analysis scripts (runcluster.pl,extract_vars.pl,decls-add-cluster.pl,dtrace-add-cluster.pl) are ported to version 2 by Fix runcluster.pl #856, andDtraceNonceFixeris fixed by DtraceNonceFixer: fix lost variables, declarations, and compressed files #867.Documentation: remove the "Version 1 Declarations" chapter of the developer manual and the dfepl sections of the user manual, and rewrite the troubleshooting entry for this error.