Re: [UrJTAG-dev] [PATCH 0/4] provide reverse-engineered jamexp.y to generate jamexp.c/jamytab.h
Hi, I agree with Benjamin, SourceForge is dying and a blocker for potential newcomers to the project. I suggest migrating to whatever modern Git forge you prefer (GitHub, GitLab, SourceHut) and having a mirror on at least GitHub so that people can contribute there. I have seen many people creating GitHub PR for UrJTAG. I can help with the migration, and arrage for a "premium" SourceHut account. Jean Le mardi 20 septembre 2022 à 21:20 +0200, Geert Stappers a écrit : > On Wed, Aug 31, 2022 at 08:33:50AM +0200, Geert Stappers wrote: > > On Sat, Aug 13, 2022 at 03:13:27PM +0200, Peter Pöschl wrote: > > > Hi all, > > > > > > I have reverse-engineered a jamexp.y from > > > urjtag/src/stapl/jamexp.c. > > > The following patch series will provide this file and the unit > > > tests I wrote > > > to verify the grammar in the file. > > > > > > Patch 1 instruments the existing jamexp.c to report shift/reduce > > > actions. > > > Patch 2 adds files to allow writing unit tests with C TAP Harness > > > [1]. > > > Patch 3 adds a unit test program to exercise all productions in > > > the grammar > > > encoded in the tables of the current jamexp.c. > > > Patch 4 provides the reverse-engineered jamexp.y and another test > > > program > > > to verify the grammar in jamexp.y with the test vectors > > > introduced > > > in patch 3. > > > In jamexp.y I added an Altera copyright header, because > > > the Lexer- > > > and calculator code is theirs (the headers are reuse [2] > > > compliant). > > > > > > Unfortunately, I have zero experience with Automake, so the > > > integration in the > > > build system is missing. Hopefully somebody from the list will be > > > able to > > > achieve that. Integration hints are in the commit messages of > > > patches 3 and 4. > > > > > > Notes: > > > > > > * jamexp.y currently uses GNU extensions of the YACC syntax, > > > thus requires > > > bison. > > > The reason was to keep the differences between the old jamexp.c > > > and the new > > > jamexp.y as small as possible. > > > > > > * The ordering of the productions in the grammar is suboptimal > > > (the grammar > > > would be much clearer if the productions were ordered > > > differently). > > > The reason is to replicate the production order of the old > > > jamexp.c grammar. > > > > > > * For integration I propose to keep the old jamexp.c and > > > jamytab.h > > > by moving them to tests/stapl/nongen/. This would allow to run > > > the test with both the old and new parser. > > > > > > > I'm gonna review the patches. > > Short: Applied > > > Long: > - patch 1 and 2 were fine > - patch 3 had two trailing spaces, they are removed > - patch 4 had a trailing space, it is removed > - all applied > - Source Forge is the leading git repository > > > Something else: > > Those who want the urjtag project moving faster, > please express yourself. > > > Regards > Geert Stappers > Not the same person as when he came to this project. ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] [PATCH 0/4] provide reverse-engineered jamexp.y to generate jamexp.c/jamytab.h
Le mar. 20 sept. 2022 à 21:21, Geert Stappers a écrit : > On Wed, Aug 31, 2022 at 08:33:50AM +0200, Geert Stappers wrote: > > On Sat, Aug 13, 2022 at 03:13:27PM +0200, Peter Pöschl wrote: > > > Hi all, > > > > > > I have reverse-engineered a jamexp.y from urjtag/src/stapl/jamexp.c. > > > The following patch series will provide this file and the unit tests I > wrote > > > to verify the grammar in the file. > > > > > > Patch 1 instruments the existing jamexp.c to report shift/reduce > actions. > > > Patch 2 adds files to allow writing unit tests with C TAP Harness [1]. > > > Patch 3 adds a unit test program to exercise all productions in the > grammar > > > encoded in the tables of the current jamexp.c. > > > Patch 4 provides the reverse-engineered jamexp.y and another test > program > > > to verify the grammar in jamexp.y with the test vectors > introduced > > > in patch 3. > > > In jamexp.y I added an Altera copyright header, because the > Lexer- > > > and calculator code is theirs (the headers are reuse [2] > compliant). > > > > > > Unfortunately, I have zero experience with Automake, so the > integration in the > > > build system is missing. Hopefully somebody from the list will be able > to > > > achieve that. Integration hints are in the commit messages of patches > 3 and 4. > > > > > > Notes: > > > > > > * jamexp.y currently uses GNU extensions of the YACC syntax, thus > requires > > > bison. > > > The reason was to keep the differences between the old jamexp.c and > the new > > > jamexp.y as small as possible. > > > > > > * The ordering of the productions in the grammar is suboptimal (the > grammar > > > would be much clearer if the productions were ordered differently). > > > The reason is to replicate the production order of the old jamexp.c > grammar. > > > > > > * For integration I propose to keep the old jamexp.c and jamytab.h > > > by moving them to tests/stapl/nongen/. This would allow to run > > > the test with both the old and new parser. > > > > > > > I'm gonna review the patches. > > Short: Applied > > > Long: > - patch 1 and 2 were fine > - patch 3 had two trailing spaces, they are removed > - patch 4 had a trailing space, it is removed > - all applied > - Source Forge is the leading git repository > > > Something else: > > Those who want the urjtag project moving faster, > please express yourself. > Move to Github, Gitlab or Codeberg? Best, ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] [PATCH 0/4] provide reverse-engineered jamexp.y to generate jamexp.c/jamytab.h
On Wed, Aug 31, 2022 at 08:33:50AM +0200, Geert Stappers wrote: > On Sat, Aug 13, 2022 at 03:13:27PM +0200, Peter Pöschl wrote: > > Hi all, > > > > I have reverse-engineered a jamexp.y from urjtag/src/stapl/jamexp.c. > > The following patch series will provide this file and the unit tests I > > wrote > > to verify the grammar in the file. > > > > Patch 1 instruments the existing jamexp.c to report shift/reduce actions. > > Patch 2 adds files to allow writing unit tests with C TAP Harness [1]. > > Patch 3 adds a unit test program to exercise all productions in the grammar > > encoded in the tables of the current jamexp.c. > > Patch 4 provides the reverse-engineered jamexp.y and another test program > > to verify the grammar in jamexp.y with the test vectors introduced > > in patch 3. > > In jamexp.y I added an Altera copyright header, because the Lexer- > > and calculator code is theirs (the headers are reuse [2] compliant). > > > > Unfortunately, I have zero experience with Automake, so the integration in > > the > > build system is missing. Hopefully somebody from the list will be able to > > achieve that. Integration hints are in the commit messages of patches 3 and > > 4. > > > > Notes: > > > > * jamexp.y currently uses GNU extensions of the YACC syntax, thus requires > > bison. > > The reason was to keep the differences between the old jamexp.c and the > > new > > jamexp.y as small as possible. > > > > * The ordering of the productions in the grammar is suboptimal (the grammar > > would be much clearer if the productions were ordered differently). > > The reason is to replicate the production order of the old jamexp.c > > grammar. > > > > * For integration I propose to keep the old jamexp.c and jamytab.h > > by moving them to tests/stapl/nongen/. This would allow to run > > the test with both the old and new parser. > > > > I'm gonna review the patches. Short: Applied Long: - patch 1 and 2 were fine - patch 3 had two trailing spaces, they are removed - patch 4 had a trailing space, it is removed - all applied - Source Forge is the leading git repository Something else: Those who want the urjtag project moving faster, please express yourself. Regards Geert Stappers Not the same person as when he came to this project. -- Silence is hard to parse ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] [PATCH 0/4] provide reverse-engineered jamexp.y to generate jamexp.c/jamytab.h
On Sat, Aug 13, 2022 at 03:13:27PM +0200, Peter Pöschl wrote: > Hi all, > > I have reverse-engineered a jamexp.y from urjtag/src/stapl/jamexp.c. > The following patch series will provide this file and the unit tests I wrote > to verify the grammar in the file. > > Patch 1 instruments the existing jamexp.c to report shift/reduce actions. > Patch 2 adds files to allow writing unit tests with C TAP Harness [1]. > Patch 3 adds a unit test program to exercise all productions in the grammar > encoded in the tables of the current jamexp.c. > Patch 4 provides the reverse-engineered jamexp.y and another test program > to verify the grammar in jamexp.y with the test vectors introduced > in patch 3. > In jamexp.y I added an Altera copyright header, because the Lexer- > and calculator code is theirs (the headers are reuse [2] compliant). > > Unfortunately, I have zero experience with Automake, so the integration in > the > build system is missing. Hopefully somebody from the list will be able to > achieve that. Integration hints are in the commit messages of patches 3 and 4. > > Notes: > > * jamexp.y currently uses GNU extensions of the YACC syntax, thus requires > bison. > The reason was to keep the differences between the old jamexp.c and the new > jamexp.y as small as possible. > > * The ordering of the productions in the grammar is suboptimal (the grammar > would be much clearer if the productions were ordered differently). > The reason is to replicate the production order of the old jamexp.c grammar. > > * For integration I propose to keep the old jamexp.c and jamytab.h > by moving them to tests/stapl/nongen/. This would allow to run > the test with both the old and new parser. > > > Can anybody tell me > > * which packets to install and > * how to call ./configure ./configure It has sane defaults. > to generate 32-bit executables on an amd64 Debian Linux system? Be inside such 32-bit environment. > This is my first foray into the multiarch/multilib area. Acknowledge > I developed the tests with a home-grown build system and managed to create 32- > bit binaries by linking with files from libc6-dev-i386-x32-cross and > lib32gcc-11-dev. > > After './configure --host x86_64-linux-gnux32 --enable-stapl', however, a > make > does not produce the desired 32-bit binaries. Sorry for the late response. I'm gonna review the patches. The wish for some special binaries have I seen, it won't get further attention from me. > Best Regards, > >Peter Pöschl Groeten Geert Stappers [1] https://www.eyrie.org/~eagle/software/c-tap-harness [2] https://reuse.software/tutorial -- Silence is hard to parse ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
[UrJTAG-dev] [PATCH 0/4] provide reverse-engineered jamexp.y to generate jamexp.c/jamytab.h
Hi all, I have reverse-engineered a jamexp.y from urjtag/src/stapl/jamexp.c. The following patch series will provide this file and the unit tests I wrote to verify the grammar in the file. Patch 1 instruments the existing jamexp.c to report shift/reduce actions. Patch 2 adds files to allow writing unit tests with C TAP Harness [1]. Patch 3 adds a unit test program to exercise all productions in the grammar encoded in the tables of the current jamexp.c. Patch 4 provides the reverse-engineered jamexp.y and another test program to verify the grammar in jamexp.y with the test vectors introduced in patch 3. In jamexp.y I added an Altera copyright header, because the Lexer- and calculator code is theirs (the headers are reuse [2] compliant). Unfortunately, I have zero experience with Automake, so the integration in the build system is missing. Hopefully somebody from the list will be able to achieve that. Integration hints are in the commit messages of patches 3 and 4. Notes: * jamexp.y currently uses GNU extensions of the YACC syntax, thus requires bison. The reason was to keep the differences between the old jamexp.c and the new jamexp.y as small as possible. * The ordering of the productions in the grammar is suboptimal (the grammar would be much clearer if the productions were ordered differently). The reason is to replicate the production order of the old jamexp.c grammar. * For integration I propose to keep the old jamexp.c and jamytab.h by moving them to tests/stapl/nongen/. This would allow to run the test with both the old and new parser. Can anybody tell me * which packets to install and * how to call ./configure to generate 32-bit executables on an amd64 Debian Linux system (bookworm/testing)? This is my first foray into the multiarch/multilib area. I developed the tests with a home-grown build system and managed to create 32- bit binaries by linking with files from libc6-dev-i386-x32-cross and lib32gcc-11-dev. After './configure --host x86_64-linux-gnux32 --enable-stapl', however, a make does not produce the desired 32-bit binaries. Best Regards, Peter Pöschl [1] https://www.eyrie.org/~eagle/software/c-tap-harness [2] https://reuse.software/tutorial ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
