This is a notice of intent to maintain this port (request maintainership). I will be maintaining the source, as well (new MASTER_SITES). As it is; the source hasn't been touched in ~10yrs. The original [authors] have since lost their domain, and interest. The currently stated license (MIT). Will not conflict with my intention(s), or their [lack of] interest(s). svn diff && reports to follow shortly. --Chris
Created attachment 146355 [details] databases/pgaccess [maintainer] STAGED SVN DIFF Here's the patch (svn diff) I promised. Adds STAGE MAINTAINER MASTER_SITES BUMPED REVISION MODIFIED VERSION NUMBER Additional modifications to source, docs, and pkg-plist (see svn diff, attached). Reports: https://redports.org/~portmaster/20140827042401-39847-237527/pgaccess-1.00.20140827_4.log https://redports.org/~portmaster/20140827042401-39847-237526/pgaccess-1.00.20140827_4.log https://redports.org/~portmaster/20140827042401-39847-237525/pgaccess-1.00.20140827_4.log https://redports.org/~portmaster/20140827042401-39847-237524/pgaccess-1.00.20140827_4.log Thank you for all your time, and consideration. --Chris
Chris, I really need to insist on poudriere logs. You estimated the new dev box would be up by now. Did that happen?
Created attachment 146433 [details] databases/pgaccess build,stage,clean output (as requested) OK. I think this should provide the information, you're looking for. :) Thanks, John. --Chris out...
well, no, I asked for poudriere and this isn't poudriere. It's a live system build. And it has warnings: Warning: Possibly owned by dependency: @dirrmtry share/doc/postgresql Warning: Possibly owned by dependency: @dirrmtry share/postgresql ^^ that means pkg-plist has unneeded lines The PR exists, the port is not going to be cut in 2 days. We can wait for poudriere logs. I thought you were going to have poudriere running this week?
Created attachment 146459 [details] [maintainer] MASTER_SITES LICENSE STAGE SVN DIFF [UPDATE] UPDATED Makefile && pkg-plist This change obsoletes the previous svn diff (2014-08-26) AND Fixes LICENSE from that diff Fixes pkg-plist from that diff This completes the work required to 1) STAGE this port (why it was DEPRECATED) 2) adds MAINTAINER (why MAINTAINER was reset) 3) adds MASTER_SITES (as source is no longer maintained upstream) 4) adds LICENSE (as none was declared) 5) makes requisite modifications to source, to enable the port to build, and build install CORRECTLY Honestly; this IS patch-ready. --Chris
Created attachment 146460 [details] databases/pgaccess Most recent output from requisite tests Most recent output from the testing session (databases-pgaccess-output) This clearly shows that the svn diff also provided, meets all the requirements to call this port STAGE ready. It also provides for the correction of errors indicated in the previous test session. This IS patch-ready. :) --Chris
<sigh> okay, poudriere logs are not coming (learned elsewhere). These logs don't include "make stage-qa" checks. There are no obvious errors in the provided diff. Moving this to patch-ready with an "take at your own risk" flag.
(In reply to John Marino from comment #7) > <sigh> > > okay, poudriere logs are not coming (learned elsewhere). > > These logs don't include "make stage-qa" checks. > > There are no obvious errors in the provided diff. > > Moving this to patch-ready with an "take at your own risk" flag. Sorry. While "wordy", it was probably not very _concise_. I jest went to the store and picked up a cheap SSD to experiment with. I should have 0 trouble getting poudriere running well enough to send you your desired logs by tonight, or tomorrow. :) All the best. --Chris
(In reply to C Hutchinson from comment #8) > (In reply to John Marino from comment #7) > > <sigh> > > > > okay, poudriere logs are not coming (learned elsewhere). > > > > These logs don't include "make stage-qa" checks. > > > > There are no obvious errors in the provided diff. > > > > Moving this to patch-ready with an "take at your own risk" flag. > > Sorry. While "wordy", it was probably not very _concise_. > I jest went to the store and picked up a cheap SSD to experiment > with. I should have 0 trouble getting poudriere running well > enough to send you your desired logs by tonight, or tomorrow. :) > > All the best. > > --Chris well, 2 of 3 PRs were moved to patch-ready with a "your risk" flag. Getting poudriere logs in will definitely remove those flags. The third was kicked back for technical reasons.
Created attachment 146589 [details] Updated patch addressing most issues with the previous submission There were some warnings from both portlint and check-sanity during build. I have corrected and simplified wherever possible. Still there is at least one thing left for the maintainer to determine. You are setting TCLVERSION?=8.4 and do some reinplacement during build. What is the relevance of this? On a clean system as of today, the default tcl installation is 8.6. Does the port still work then? Is the reinplace with TCLVERSION really necessary? Could you use this patch as a starting point for resolving the tcl issue, we should really try to get this committed today before the expiration.
(In reply to Thomas Zander from comment #10) > Created attachment 146589 [details] > Updated patch addressing most issues with the previous submission > > There were some warnings from both portlint and check-sanity during build. I > have corrected and simplified wherever possible. > Still there is at least one thing left for the maintainer to determine. You > are setting TCLVERSION?=8.4 and do some reinplacement during build. > What is the relevance of this? On a clean system as of today, the default > tcl installation is 8.6. Does the port still work then? Is the reinplace > with TCLVERSION really necessary? > Could you use this patch as a starting point for resolving the tcl issue, we > should really try to get this committed today before the expiration. Good call, and thank you for the changes you made. I had intended to come back to it after getting a poudriere install properly setup, to better manage all of my submissions. But, as yet, I am unable to get poudriere to use MY locally maintained ports tree. All the questions I've asked elsewhere to accomplish this. Have gone unanswered. Looks like I'll be better off using dump(8), and restore(8) between all of my make sessions. To get a fresh copy of my current 11-CURRENT new world && kernel. Thanks again, and I'll submit something more coherent ASAP. --Chris
(In reply to C Hutchinson from comment #11) > But, as yet, > I am unable to get poudriere to use MY locally maintained ports tree. > All the questions I've asked elsewhere to accomplish this. Have gone > unanswered. Why on earth are you trying to do that? Nobody wants you do to start with a contaminated tree. That would be unreproducible for us. We would want you to use a standard tree and replace / update the port you want to test with the shar/diff that you are submitting.
(In reply to John Marino from comment #12) > (In reply to C Hutchinson from comment #11) > > But, as yet, > > I am unable to get poudriere to use MY locally maintained ports tree. > > All the questions I've asked elsewhere to accomplish this. Have gone > > unanswered. > > > Why on earth are you trying to do that? > Nobody wants you do to start with a contaminated tree. That would be > unreproducible for us. > > We would want you to use a standard tree and replace / update the port you > want to test with the shar/diff that you are submitting. Understood. I guess I could have better explained that as; use MY version (revision) of the tree. In other words, the revision I have checked out. Where I can then batch the builds of my modifications against it. --Chris
(In reply to C Hutchinson from comment #13) > (In reply to John Marino from comment #12) > > (In reply to C Hutchinson from comment #11) > > > But, as yet, > > > I am unable to get poudriere to use MY locally maintained ports tree. > > > All the questions I've asked elsewhere to accomplish this. Have gone > > > unanswered. > > > > > > Why on earth are you trying to do that? > > Nobody wants you do to start with a contaminated tree. That would be > > unreproducible for us. > > > > We would want you to use a standard tree and replace / update the port you > > want to test with the shar/diff that you are submitting. > > Understood. I guess I could have better explained that as; > use MY version (revision) of the tree. In other words, the revision > I have checked out. Where I can then batch the builds of my modifications > against it. > > --Chris Oh, and if I [now] understand it correctly. It's done thusly: poudriere ports -c -F -f none -M /ports/custom -p custom where /ports/custom is the revision I have checked out [untouched] Then I just feed poudriere the list of ports I have modified, and it re-runs the build within that tree. Only building the ports list I feed it. Sorry. But I'm having a tough time w/it. For example; the docs/ examples keep using terms incorrectly -- jail(s), for example. When what it should _actually_ say, is chroot(8). There are other examples. But I'll just leave it at that, for now. :) --Chris
(In reply to C Hutchinson from comment #14) > Oh, and if I [now] understand it correctly. It's done thusly: > poudriere ports -c -F -f none -M /ports/custom -p custom > where /ports/custom is the revision I have checked out [untouched] > Then I just feed poudriere the list of ports I have modified, and > it re-runs the build within that tree. Only building the ports list > I feed it. I was actually saying to create a ports tree with the defaults, and then individually update the port in question. You created an entry for a ports tree that points to an existing ports tree. I also have no idea how you update that custom ports tree -- manually I suppose. > Sorry. But I'm having a tough time w/it. For example; the docs/ > examples keep using terms incorrectly -- jail(s), for example. > When what it should _actually_ say, is chroot(8). There are other > examples. But I'll just leave it at that, for now. :) Documentation sucks, but it doesn't use chroot, it uses jails.
(In reply to C Hutchinson from comment #14) > Oh, and if I [now] understand it correctly. It's done thusly: > poudriere ports -c -F -f none -M /ports/custom -p custom > where /ports/custom is the revision I have checked out [untouched] > Then I just feed poudriere the list of ports I have modified, and > it re-runs the build within that tree. Only building the ports list > I feed it. No. /ports/custom is the ports tree including all the modification that you want to test. It's simple actually. Poudriere does not do any merge/modification of the tree whatsoever. It just uses the ports tree rooted at /ports/custom in your example to build ports from. So whatever deviations you made there are used. > Sorry. But I'm having a tough time w/it. For example; the docs/ > examples keep using terms incorrectly -- jail(s), for example. > When what it should _actually_ say, is chroot(8). There are other > examples. But I'll just leave it at that, for now. :) I am not sure I understand what you mean. poudriere does build ports inside jails, they are not just chrooted. The ports tree you are giving to it for the build (by using e.g. 'poudriere testport category/portname -p custom ...') is nullfs-mounted into the jail to /usr/ports and then inside the jail everything runs normal. You can jexec into it and investigate build results and logs and whatever you need.
(In reply to John Marino from comment #15) > I also have no idea how you > update that custom ports tree -- manually I suppose. Yep. When you create custom trees, poudriere does not assume anything. If no method is specified, it does not do anything and leaves the management/update etc. to the user. This works great combined with zfs clone to test multiple mutually incompatible modifications to the ports tree at the same time.
Thomas, on a barely related note, the new documentation has a poudriere section but it's only a link to the poudriere web site, which is not only a self-signed cert site (meaning red/yellow warnings - do not enter) but it's also terrible. mat@ told me "contributions welcome" What that section needs is to explain what poudriere is and have a mini "how-to" that would build a jail, create a ports tree, and run through a couple of examples of "poudriere build" and "poudriere testport". Do you have any interest in writing that page? It's hard to tell people to provide poudriere logs when such a simple howto isn't available anywhere (including the poudriere site)
(In reply to John Marino from comment #18) > Thomas, on a barely related note, the new documentation has a poudriere > section but it's only a link to the poudriere web site, which is not only a > self-signed cert site (meaning red/yellow warnings - do not enter) but it's > also terrible. > > mat@ told me "contributions welcome" > > What that section needs is to explain what poudriere is and have a mini > "how-to" that would build a jail, create a ports tree, and run through a > couple of examples of "poudriere build" and "poudriere testport". > > Do you have any interest in writing that page? It's hard to tell people to > provide poudriere logs when such a simple howto isn't available anywhere > (including the poudriere site) You mean this one?: https://www.freebsd.org/doc/en/books/porters-handbook/testing-poudriere.html In general, yes, I'd be happy to write that page. It is just a question of when. Let's take this thread offline before we get off-topic.
(In reply to John Marino from comment #18) > Thomas, on a barely related note, the new documentation has a poudriere > section but it's only a link to the poudriere web site, which is not only a > self-signed cert site (meaning red/yellow warnings - do not enter) but it's > also terrible. > > mat@ told me "contributions welcome" > > What that section needs is to explain what poudriere is and have a mini > "how-to" that would build a jail, create a ports tree, and run through a > couple of examples of "poudriere build" and "poudriere testport". > > Do you have any interest in writing that page? It's hard to tell people to > provide poudriere logs when such a simple howto isn't available anywhere > (including the poudriere site) FWIW -- keeping OT ;) I've tried to draw from the following: https://forums.freebsd.org/viewtopic.php?f=39&t=38859 http://negativo17.org/freebsd-10-new-x-org-kms-pkgng-poudriere/ https://wiki.freebsd.org/PkgPrimer https://www.freebsd.org/doc/handbook/ports-poudriere.html http://fossil.etoilebsd.net/poudriere/doc/trunk/doc/index.wiki http://fossil.etoilebsd.net/poudriere/doc/trunk/doc/use_system_ports_tree.wiki https://www.freebsd.org/cgi/man.cgi?query=poudriere&apropos=0&sektion=0&manpath=FreeBSD+9.2-RELEASE+and+Ports&arch=default&format=html http://www.neant.ro/page/2/ http://blog.feld.me/tag/pkg/ http://blather.michaelwlucas.com/archives/1941 http://dan.langille.org/2014/04/04/using-poudriere-to-create-a-custom-freebsd-repository-for-package-installation/ https://www.mywushublog.com/2013/04/building-packages-for-freebsd/ Maybe there's enough for someone already familiar with poo-dree-A to assemble something useful, as a sort 0f, how-to. Hope this helps, and thanks. --Chris
(In reply to Thomas Zander from comment #19) > You mean this one?: > https://www.freebsd.org/doc/en/books/porters-handbook/testing-poudriere.html yes, I thought it was new with the draft I saw, but it seems that particular page is already publish > In general, yes, I'd be happy to write that page. It is just a question of > when. Let's take this thread offline before we get off-topic. Okay. I think mat@ would have those answers anyway. I just brought up the subject here because it seemed like you'd do a good job on it and I haven't the time in the very near future.
Created attachment 146709 [details] databases/pgaccess updated patch new distfile version other changes databases/pgaccess This patch draws upon Thomas Zander's previous patch, and removes requirement for TCL version being exactly 8.4. I made some grammatical modifications to pkg-descr, as well as some minor re-formatting of the same. I was also required to make some changes to the source, which required changing distinfo, and Makefile to reflect same. I am also attaching output from requisite testing (pgaccess-test-logs). Thank you for all your time, and consideration. --Chris
Created attachment 146710 [details] databases/pgaccess output from requisite testing against 2014-09-02.diff databases/pgaccess This attachment contains the output from requisite testing for the 2014-09-02.diff, also attached. Thank you for all your time, and consideration. --Chris
Portlint is squawking: /!\ portlint-2.15.4: Makefile warnings, please consider fixing /!\ LICENSE must not contain BSD, instead use BSD[234]CLAUSE That means the LICENCE= line needs to be fixed. It tells you how in the error.
LICENSE looks correct here. After patching I see LICENSE= BSD4CLAUSE in Makefile. But the tcl part is not fully resolved. In the do-build target there is this reinplace: @ ${SED} -e "s,%%LOCALBASE%%,${LOCALBASE},g; \ s,%%TCL_VERSION%%,${TCLVERSION},g; \ s,%%PREFIX%%,${PREFIX},g; \ s,%%PROGRAM%%,${targetfile},g" \ < ${FILESDIR}/run-tcl-file > ${WRKDIR}/${targetfile:R}.sh which wants to implant TCLVERSION in the script, but now TCLVERSION is not defined anymore. Could you take a look? Thanks!
(In reply to Thomas Zander from comment #25) > LICENSE looks correct here. After patching I see > > LICENSE= BSD4CLAUSE So it is. Then why is portlint complaining falsely? Unless it wasn't run on the latest patch?
(In reply to John Marino from comment #24) > Portlint is squawking: > > /!\ portlint-2.15.4: Makefile warnings, please consider fixing /!\ > > LICENSE must not contain BSD, instead use BSD[234]CLAUSE > > > > > > > That means the LICENCE= line needs to be fixed. It tells you how in the > error. Thanks, yes. I knew that. I wanted to include it, so that I would remember to fix it, and send the maintainer a pr with a patch for it. :) I'm on a recent 11-CURRENT. So I'm guessing this error still exists. --Chris
okay, that's from installing portlint. That shouldn't be part of the log. Too much information, and it mislead me. Ultimately it's my fault for not reading closer. Although it's interesting to now that portlint was never installed until now...
(In reply to Thomas Zander from comment #25) > LICENSE looks correct here. After patching I see > > LICENSE= BSD4CLAUSE > > in Makefile. But the tcl part is not fully resolved. In the do-build target > there is this reinplace: > > @ ${SED} -e "s,%%LOCALBASE%%,${LOCALBASE},g; \ > s,%%TCL_VERSION%%,${TCLVERSION},g; \ > s,%%PREFIX%%,${PREFIX},g; \ > s,%%PROGRAM%%,${targetfile},g" \ > < ${FILESDIR}/run-tcl-file > ${WRKDIR}/${targetfile:R}.sh > > which wants to implant TCLVERSION in the script, but now TCLVERSION is not > defined anymore. Could you take a look? Thanks! I know the target for this (files/run-tcl-file). I'm going to have to extrapolate what's ultimately chosen|already-installed and populate it with that version. it's 12:30 AM, my time. So unless one of you chooses to attack this. I'm going to get a little sleep before I do it myself. :) Thanks! --Chris
A commit references this bug: Author: riggs Date: Sat Sep 6 09:27:25 UTC 2014 New revision: 367409 URL: http://svnweb.freebsd.org/changeset/ports/367409 Log: - Update to new upstream version 1.00.20140902 - Assign maintainership to Chris Hutchinson <portmaster@bsdforge.com> - Cleanup Makefile, update pkg-descr - Stagify - Remove EXPIRATION_DATE PR: 193046 Submitted by: portmaster@bsdforge.com Reviewed by: riggs Changes: head/databases/pgaccess/Makefile head/databases/pgaccess/distinfo head/databases/pgaccess/files/run-tcl-file head/databases/pgaccess/pkg-descr head/databases/pgaccess/pkg-plist
(In reply to commit-hook from comment #30) > A commit references this bug: > > Author: riggs > Date: Sat Sep 6 09:27:25 UTC 2014 > New revision: 367409 > URL: http://svnweb.freebsd.org/changeset/ports/367409 > > Log: > - Update to new upstream version 1.00.20140902 > - Assign maintainership to Chris Hutchinson <portmaster@bsdforge.com> > - Cleanup Makefile, update pkg-descr > - Stagify > - Remove EXPIRATION_DATE > > PR: 193046 > Submitted by: portmaster@bsdforge.com > Reviewed by: riggs > > Changes: > head/databases/pgaccess/Makefile > head/databases/pgaccess/distinfo > head/databases/pgaccess/files/run-tcl-file > head/databases/pgaccess/pkg-descr > head/databases/pgaccess/pkg-plist Thank you, very much, riggs. Greatly appreciated. --Chris