summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
...
| * | fix(skel): Init state buffer correctlyJoseph Langley2021-07-023-57/+40
| | | | | | | | | | | | State buffer initialization in cpp-flex.skl got lost during a rebase. Added it back in from c99-flex.skl and made improvements suggested by smcpeak.
| * | chore(perm): Fix permissions 755->644Joseph Langley2021-07-027-0/+0
| | |
| * | chore(make): Ensure distcheck cleans up directories.Joseph Langley2021-06-092-3/+4
| | |
| * | fix(skel): Fix porting errors from the cpp to the c99 and go skeletons.Joseph Langley2021-06-092-2/+5
| | |
| * | fix(skel): Use yypanic in place of YY_FATAL_ERROR.Joseph Langley2021-06-092-2/+2
| | | | | | | | | | | | yypanic() is new in the c99 and go skeletons.
| * | fix(skel): Add state buffer size handling to reentrant scanner skeletons.Joseph Langley2021-06-093-12/+79
| | |
| * | chore(test): Declare script variables before use.Joseph Langley2021-06-091-0/+2
| | | | | | | | | | | | For set -euvx
| * | fix(skel): Add state buffer size handling to non-reentrant scanner skeletons.Joseph Langley2021-06-093-5/+109
| | |
| * | Ensure size of Reject state buffer is sufficient each time ↵Mightyjo2021-06-095-1/+556
| | | | | | | | | | | | yy_switch_to_buffer is called.
| * | Don't use C++ auto, it angers TravisCI.Mightyjo2021-06-091-1/+1
| | |
| * | Fix test failure on TravisCI.Mightyjo2021-06-091-1/+1
| | |
| * | Test for reversion of CVE-2006-0459.Mightyjo2021-06-091-1/+5
| | |
| * | Test for reversion of CVE-2006-0459.Mightyjo2021-06-092-0/+544
| | |
* | | chore(action): Clean up dead codeJoseph Langley2021-08-131-4/+1
| | | | | | | | | Leave clear instructions on what to do if the Autobuild step breaks.
* | | fix(action): Let Autobuild do most of the workJoseph Langley2021-08-131-12/+18
| | | | | | | | | Try preparing configure with autogen.sh, then letting codeql-action/autobuild do it's work prior to analysis.
* | | Update codeql-analysis.ymlJoseph Langley2021-08-131-6/+11
| | |
* | | fix compile warning originating in cpp skeletonSimon Sobisch2021-06-221-1/+1
|/ / | | | | `gcount()` is of type `streamsize` which has a system-dependent size with at least `int`. Because of the `int max_size` limit no bigger value will be returned, the explicit cast fixes warnings from compilers/static analyzers.
* | Merge branch 'yy-init-buffer-nonnull' of git://github.com/Explorer09/flex ↵Will Estes2021-06-083-3/+3
|\ \ | | | | | | | | | into Explorer09-yy-init-buffer-nonnull
| * | scanner: Fix compiler confusion in yy_init_buffer()Explorer092021-06-073-3/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When complied with 'gcc -O3', the yy_init_buffer call can confuse gcc, thinking the 'b' pointer may be NULL. (gcc would warn that if '-Wnull-dereference' is used). Fix the confusion by never pass a NULL constant to the function. Fixes: #377 Signed-off-by: Kang-Che Sung <explorer09@gmail.com>
* | | Remove manual m4 search code. No more stat() dependency.Explorer092021-06-073-39/+1
|/ / | | | | | | | | | | | | | | | | | | | | | | | | | | This reverts commit c34590c4bf067d4e749e5298a8ce9f90072d6332 and all search code patches that followed it. The execvp() would do the $PATH searching of m4 for flex. There's no need to duplicate the effort. Another reason for the removal is to remove dependency on stat() system call, which would cause a problem in a 32-bit OS accessing a filesystem with 64-bit inode number (see #413). Signed-off-by: Kang-Che Sung <explorer09@gmail.com>
* | Merge branch 'autogen-sh' of git://github.com/Explorer09/flex into ↵Will Estes2021-06-051-12/+13
|\ \ | | | | | | | | | Explorer09-autogen-sh
| * | build: Fix autogen.sh $LIBTOOLIZE detection logicExplorer092018-09-221-11/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * Don't silence `libtoolize --version` error message. It would be useful for diagnosing any problem when `libtoolize --version` fails. (libtoolize not found? No permission to execute? Now you know which error it is.) * Change the guessing code of libtoolize program name to a loop. It'll perform silently. * Silence the `libtoolize --version` output when it succeeds. Signed-off-by: Kang-che Sung <explorer09@gmail.com>
| * | build: autogen.sh libtoolize invocation minor fixes.Explorer092018-09-221-3/+3
| | | | | | | | | | | | | | | | | | | | | | | | * Let autogen.sh error message go to stderr. * Don't quote the variable $LIBTOOLIZE when invoking it. This would allow invocation when LIBTOOLIZE='a-future-libtoolize --gnu-compat' Signed-off-by: Kang-che Sung <explorer09@gmail.com>
* | | Merge branch 'doc_fix_typo' of git://github.com/houtanb/flex into ↵Will Estes2021-06-051-1/+1
|\ \ \ | | | | | | | | | | | | houtanb-doc_fix_typo
| * | | doc: fix typoHoutan Bastani2018-11-261-1/+1
| |/ /
* | | Merge branch 'westes:master' into esr_retargetingJoseph Langley2021-06-051-0/+71
|\ \ \
| * | | Create codeql-analysis.ymlWill Estes2021-06-051-0/+71
| | | |
* | | | chore(test): Clean up debugging in testsJoseph Langley2021-06-053-6/+1
| | | |
* | | | fix(test): Re-add posix test input to repo.Joseph Langley2021-06-052-0/+2
| | | |
* | | | chore(test): Enable shell '-eu' strictnessJoseph Langley2021-06-052-6/+9
| | | |
* | | | chore(test): Regenerate ruleset.amJoseph Langley2021-06-051-1/+1
| | | |
* | | | fix(test): Make test generating scripts work in VPATH buildsJoseph Langley2021-06-055-250/+324
| | | |
* | | | fix(test): Re-add test inputs to tests/Joseph Langley2021-06-0522-0/+109
| | | | | | | | | | | | | | | | | | | | Generating these test inputs on the fly breaks VPATH builds and 'make distcheck'. Glad we have a lot fewer of them now, at least.
* | | | fix(build): Set texi2* include paths for VPATH.Joseph Langley2021-06-042-3/+6
| | | | | | | | | | | | | | | | | | | | | | | | Override the TEXI2DVI and TEXI2PDF automake variables to add an include directory so VPATH builds work. I'm not thrilled with doing this. The right thing is probably to move the included examples to doc/ and do the VPATH magic in examples/manual/ where LaTeX is better able to deal with it.
* | | | fix(build): Expand wildcards in Makefile.am.Joseph Langley2021-06-041-6/+27
| | | | | | | | | | | | | | | | Wildcards were causing automake to generate spurious build targets.
* | | | fix(build): Replace wildcards in Makefile.am.Joseph Langley2021-06-041-3/+11
| | | | | | | | | | | | | | | | Wildcards were causing automake to generate build rules for the *-filename dummies. They're not generally portable.
* | | | fix(build): Prevent automake computing default _SOURCES for tests.Joseph Langley2021-06-033-35/+34
| | | | | | | | | | | | | | | | | | | | | | | | Automake attempts to compute default _SOURCES for the test-yydecl-*.sh runners generated by test-yydecl-gen.sh. Providing an intentionally empty _SOURCES variable for each script fixes the attendant errors. Also removed EXTRA_DIST lines for deleted test inputs.
* | | | Merge pull request #16 from eric-s-raymond/retargetingJoseph Langley2021-06-02134-9369/+15868
|\ \ \ \ | |/ / / |/| | | Reviewing ESR's Retargeting branch from westes/flex
| * | | Remove prrequisite from .l.go suffix ruleEric S. Raymond2021-02-051-1/+1
| | | | | | | | | | | | | | | | | | | | autogen gets indigestion iv you give a suffix rule a prerequisite, so don't do that.
| * | | Go back end: More Goification of names.Eric S. Raymond2020-12-042-28/+39
| | | |
| * | | Portability fix in tests - not all languages have ++ and --.Eric S. Raymond2020-12-042-2/+10
| | | |
| * | | check that each skeleton we build has the correct number of sections (#7)Daniel Brooks2020-11-262-0/+36
| | | | | | | | | | | | | | | | Probably we could turn the section markers into m4 macros, but until then this will prevent accidents.
| * | | Mollify picky C compilers some more (#8)Daniel Brooks2020-11-262-2/+2
| | | |
| * | | Improve standards conformance in C99 and Go skeletons.Eric S. Raymond2020-11-262-14/+24
| | | |
| * | | Deal with a GCC warning elicited by upgrade to GCC 10.2.0.Eric S. Raymond2020-11-201-2/+2
| | | |
| * | | In the Go back end, more moves towards Go conventions.Eric S. Raymond2020-11-202-88/+90
| | | | | | | | | | | | | | | | All documented in the go-flex.skl header comment.
| * | | Improve information hiding in C99 and Go back ends.Eric S. Raymond2020-11-204-241/+238
| | | | | | | | | | | | | | | | | | | | | | | | Things that don't need to be declared before that DFA tables shouldn't be.
| * | | In the Go back end, begin Go-ifying the public interface.Eric S. Raymond2020-11-201-68/+68
| | | | | | | | | | | | | | | | struct yyguts_t becomes FlexLexer typedef.
| * | | Eliminate forwards in the C99 back end.Eric S. Raymond2020-11-193-47/+43
| | | |
| * | | In the Go back end, elimate some forwards.Eric S. Raymond2020-11-191-38/+30
| | | |