Re: [NEW/WIP] Qflow porting // [7/7] opensta-20190320
On 2019/03/27 07:37, Alessandro DE LAURENZIS wrote: > Hello Stuart, > > On 26/03/2019 11:59, Stuart Henderson wrote: > > [...] > > I suspect you probably have a symlink for tclsh -> tclsh8.5. The following > > is needed to let it build otherwise: > > > > --- Makefile.orig Tue Mar 26 10:57:14 2019 > > +++ MakefileTue Mar 26 10:56:23 2019 > > @@ -32,6 +32,9 @@ CONFIGURE_ARGS = -DTCL_HEADER=${MODTCL_INCDIR}/tcl.h \ > > NO_TEST = Yes > > +pre-configure: > > + cd ${WRKSRC}/etc && ${MODTCL_WISH_ADJ} TclEncode.tcl SwigCleanup.tcl > > + > > post-install: > > ${INSTALL_DATA_DIR} ${PREFIX}/share/doc/opensta > > ${INSTALL_DATA} ${WRKSRC}/doc/OpenSTA.pdf ${PREFIX}/share/doc/opensta > > > > With that added, it's OK sthen@ to import. > > > > You're right, I removed tclsh and wish links to avoid similar issues for the > future. But while there, maybe it's worth noting that, being both > TclEncode.tcl and SwigCleanup.tcl run during build, and since they launch > wish, the following message is displayed: > > Application initialization failed: no display name and no $DISPLAY > > environment variable Sorry my mistake, I should have suggested MODTCL_TCLSH_ADJ rather than MODTCL_WISH_ADJ. I've fixed that and put MODULES back to lang/tcl.
Re: [NEW/WIP] Qflow porting // [7/7] opensta-20190320
Alessandro DE LAURENZIS writes: > Hello Stuart, > > On 26/03/2019 11:59, Stuart Henderson wrote: > > [...] > > I suspect you probably have a symlink for tclsh -> tclsh8.5. The following > > is needed to let it build otherwise: > > > > --- Makefile.orig Tue Mar 26 10:57:14 2019 > > +++ MakefileTue Mar 26 10:56:23 2019 > > @@ -32,6 +32,9 @@ CONFIGURE_ARGS = -DTCL_HEADER=${MODTCL_INCDIR}/tcl.h \ > > > > NO_TEST = Yes > > > > +pre-configure: > > + cd ${WRKSRC}/etc && ${MODTCL_WISH_ADJ} TclEncode.tcl SwigCleanup.tcl > > + > > post-install: > > ${INSTALL_DATA_DIR} ${PREFIX}/share/doc/opensta > > ${INSTALL_DATA} ${WRKSRC}/doc/OpenSTA.pdf ${PREFIX}/share/doc/opensta > > > > With that added, it's OK sthen@ to import. > > > > You're right, I removed tclsh and wish links to avoid similar issues for > the future. But while there, maybe it's worth noting that, being both > TclEncode.tcl and SwigCleanup.tcl run during build, and since they > launch wish, the following message is displayed: > > > Application initialization failed: no display name and no $DISPLAY environm > ent variable > > I don't understand if anything is actually going wrong, since the > compilation reach the end without problems... > > Meanwhile, upstream accepted our patch: patch-etc_SwigCleanup_tcl no > needed anymore. I've imported this with two changes. First, it needs the tk module instead of tcl, and MODTK_BUILD_DEPENDS in BUILD_DEPENDS, because tk is what contains wish. Second, I changed the version number from the date to the version number in git log. Since this commit is later than the latest version (2.0.11), but the new commit is worth keeping (corrected license text), I gave it the number 2.0.11.20190327. -- Anthony J. Bentley
Re: [NEW/WIP] Qflow porting // [7/7] opensta-20190320
Hello Stuart, On 26/03/2019 11:59, Stuart Henderson wrote: [...] I suspect you probably have a symlink for tclsh -> tclsh8.5. The following is needed to let it build otherwise: --- Makefile.orig Tue Mar 26 10:57:14 2019 +++ MakefileTue Mar 26 10:56:23 2019 @@ -32,6 +32,9 @@ CONFIGURE_ARGS = -DTCL_HEADER=${MODTCL_INCDIR}/tcl.h \ NO_TEST = Yes +pre-configure: + cd ${WRKSRC}/etc && ${MODTCL_WISH_ADJ} TclEncode.tcl SwigCleanup.tcl + post-install: ${INSTALL_DATA_DIR} ${PREFIX}/share/doc/opensta ${INSTALL_DATA} ${WRKSRC}/doc/OpenSTA.pdf ${PREFIX}/share/doc/opensta With that added, it's OK sthen@ to import. You're right, I removed tclsh and wish links to avoid similar issues for the future. But while there, maybe it's worth noting that, being both TclEncode.tcl and SwigCleanup.tcl run during build, and since they launch wish, the following message is displayed: Application initialization failed: no display name and no $DISPLAY environment variable I don't understand if anything is actually going wrong, since the compilation reach the end without problems... Meanwhile, upstream accepted our patch: patch-etc_SwigCleanup_tcl no needed anymore. Updated tarball attached. -- Alessandro DE LAURENZIS [mailto:jus...@atlantide.t28.net] Web: http://www.atlantide.t28.net LinkedIn: https://www.linkedin.com/in/delaurenzis/ opensta.tar.gz Description: application/gzip
Re: [NEW/WIP] Qflow porting // [7/7] opensta-20190320
On 2019/03/26 08:22, Alessandro DE LAURENZIS wrote: > Hello Stuart, > > On 22/03/2019 19:26, Stuart Henderson wrote: > > On 2019/03/22 16:29, Alessandro DE LAURENZIS wrote: > > > My main concern with the port is related to the license; although it is > > > clear that the software is released under the GPLv3, as confirmed by all > > > source code file headers, I'm puzzled; the README.md ends as follows: > > > > > > > ## License > > > > > > > > Copyright (c) 2019, Parallax Software, Inc. > > > > All rights reserved. > > > > > > > > No part of this document may be copied, transmitted or > > > > disclosed in any form or fashion without the express > > > > written consent of Parallax Software, Inc. > > > > > > What do you think? > > > > odd. can you ask upstream about that please? > > > > I flagged the problem upstream and they promptly amended the Copyright > notice (they use some filter to extract the open source version from the > proprietary one, see [1]). Great! > Updated tarball attached; note that I have added: > > COMPILER = base-clang ports-gcc base-gcc > > since this is C++11 > > Also, reported the reinterpret_cast issue upstream [2] to see if it is > possible to remove the only patch we currently require (this should not > prevent to import, though). > > [1] https://github.com/abk-openroad/OpenSTA/issues/17 > > [2] https://github.com/abk-openroad/OpenSTA/issues/20 > > -- > Alessandro DE LAURENZIS > [mailto:jus...@atlantide.t28.net] > Web: http://www.atlantide.t28.net > LinkedIn: https://www.linkedin.com/in/delaurenzis/ I suspect you probably have a symlink for tclsh -> tclsh8.5. The following is needed to let it build otherwise: --- Makefile.orig Tue Mar 26 10:57:14 2019 +++ MakefileTue Mar 26 10:56:23 2019 @@ -32,6 +32,9 @@ CONFIGURE_ARGS = -DTCL_HEADER=${MODTCL_INCDIR}/tcl.h \ NO_TEST = Yes +pre-configure: + cd ${WRKSRC}/etc && ${MODTCL_WISH_ADJ} TclEncode.tcl SwigCleanup.tcl + post-install: ${INSTALL_DATA_DIR} ${PREFIX}/share/doc/opensta ${INSTALL_DATA} ${WRKSRC}/doc/OpenSTA.pdf ${PREFIX}/share/doc/opensta With that added, it's OK sthen@ to import.
Re: [NEW/WIP] Qflow porting // [7/7] opensta-20190320
Hello Stuart, On 22/03/2019 19:26, Stuart Henderson wrote: On 2019/03/22 16:29, Alessandro DE LAURENZIS wrote: My main concern with the port is related to the license; although it is clear that the software is released under the GPLv3, as confirmed by all source code file headers, I'm puzzled; the README.md ends as follows: ## License Copyright (c) 2019, Parallax Software, Inc. All rights reserved. No part of this document may be copied, transmitted or disclosed in any form or fashion without the express written consent of Parallax Software, Inc. What do you think? odd. can you ask upstream about that please? I flagged the problem upstream and they promptly amended the Copyright notice (they use some filter to extract the open source version from the proprietary one, see [1]). Updated tarball attached; note that I have added: COMPILER = base-clang ports-gcc base-gcc since this is C++11 Also, reported the reinterpret_cast issue upstream [2] to see if it is possible to remove the only patch we currently require (this should not prevent to import, though). [1] https://github.com/abk-openroad/OpenSTA/issues/17 [2] https://github.com/abk-openroad/OpenSTA/issues/20 -- Alessandro DE LAURENZIS [mailto:jus...@atlantide.t28.net] Web: http://www.atlantide.t28.net LinkedIn: https://www.linkedin.com/in/delaurenzis/ opensta.tar.gz Description: application/gzip
Re: [NEW/WIP] Qflow porting // [7/7] opensta-20190320
On 2019/03/22 16:29, Alessandro DE LAURENZIS wrote: > My main concern with the port is related to the license; although it is > clear that the software is released under the GPLv3, as confirmed by all > source code file headers, I'm puzzled; the README.md ends as follows: > > > ## License > > > > Copyright (c) 2019, Parallax Software, Inc. > > All rights reserved. > > > > No part of this document may be copied, transmitted or > > disclosed in any form or fashion without the express > > written consent of Parallax Software, Inc. > > What do you think? odd. can you ask upstream about that please?
[NEW/WIP] Qflow porting // [7/7] opensta-20190320
Dear ports@ readers, this is a proposal for a new port: OpenSTA, the Parallax Static Timing Analyzer. DESCR = OpenSTA is a gate level static timing verifier. As a stand-alone executable it can be used to verify the timing of a design using standard file formats: - Verilog netlist - Liberty library - SDC timing constraints - SDF delay annotation - SPEF parasitics OpenSTA uses a TCL command interpreter to read the design, specify timing constraints and print timing reports. I knew this project since a couple of years and its porting was in my TODO list; but recently I noticed that the internal Qflow tool for static timing analysis is partially broken (already reported upstream, see [1]), so I decided to treat it as a Qflow dependency. Please note that it needs cudd (see [2], port is already ok sthen). To compile it, I only patched a file in order to change reinterpret_cast to static_cast. I also added some documentation with the post-install target (those files are in the upstream tarball, but it seems there is no way to install them through their Makefile). My main concern with the port is related to the license; although it is clear that the software is released under the GPLv3, as confirmed by all source code file headers, I'm puzzled; the README.md ends as follows: ## License Copyright (c) 2019, Parallax Software, Inc. All rights reserved. No part of this document may be copied, transmitted or disclosed in any form or fashion without the express written consent of Parallax Software, Inc. What do you think? Taking MAINTAINER, port's tarball attached. [1] http://opencircuitdesign.com/pipermail/eda-dev/2019-March/000127.html [2] https://marc.info/?l=openbsd-ports=155319441906710=2 -- Alessandro DE LAURENZIS [mailto:jus...@atlantide.t28.net] Web: http://www.atlantide.t28.net LinkedIn: https://www.linkedin.com/in/delaurenzis/ opensta.tar.gz Description: application/gzip