The newest portupgrade does a 'make $MAKE_ARGS -V PKGNAME -V IGNORE' before trying to build the port. The effect of this is that if you pass -DNO_IGNORE in make args, you still cannot build the port because portupgrade cuts you off at the knees first. (NO_IGNORE does not *unset* IGNORE, it merely negates its effect.) It is not portupgrade's business to stop me from making a port with IGNORE set. The code that does this should be removed. Checking the make args for NO_IGNORE would not be sufficient, because it could be in /etc/make.conf, too. FWIW, the -v option should show matches in pkgtools.conf for things, and should also show make invocations. It would have made finding out why portupgrade was broken a lot easier. Fix: Remove the offending code. Let the port tell you if it isn't going to build. Please do not try to protect the user by preventing an action that should (and used to) work correctly. How-To-Repeat: Try to rebuild an installed java/jdk13 with: portupgrade -f -m '-DNO_IGNORE -DBATCH' jdk And please don't tell me to just remove the '-DBATCH'; in reality, it's in my /etc/make.conf, for a reason, and will stay there.
Responsible Changed From-To: freebsd-ports->knu Over to maintainer
At Fri, 9 Aug 2002 03:55:04 -0400 (EDT), Alan Eldridge wrote: > It is not portupgrade's business to stop me from making a port with > IGNORE set. The code that does this should be removed. You could try modifying portupgrade and see what would happen. For ports that are marked as IGNORE, "make build/install" doesn't fail in error, so portupgrade tries to proceed anyway; and due misery follows. For the record, Jimmy Olgeni was the one who informed me of that. > >Fix: > > Remove the offending code. Let the port tell you if it isn't going to build. > Please do not try to protect the user by preventing an action that should > (and used to) work correctly. As I explained above, there is a case where user is terribly hurt and that's why the check was introduced. I'll alter the code to check NO_IGNORE also to solve your problem. -- / /__ __ Akinori.org / MUSHA.org / ) ) ) ) / FreeBSD.org / Ruby-lang.org Akinori MUSHA aka / (_ / ( (__( @ iDaemons.org / and.or.jp "When I leave I don't know what I'm hoping to find When I leave I don't know what I'm leaving behind.."
At Fri, 09 Aug 2002 20:07:12 +0900, I wrote: > As I explained above, there is a case where user is terribly hurt and > that's why the check was introduced. I'll alter the code to check > NO_IGNORE also to solve your problem. Here's a patch, but beware, defining NO_IGNORE could be dangerous in some cases (ports). Index: portupgrade =================================================================== RCS file: /home/cvs/pkgtools/bin/portupgrade,v retrieving revision 1.182 diff -u -1 -r1.182 portupgrade --- portupgrade 4 Aug 2002 22:46:15 -0000 1.182 +++ portupgrade 9 Aug 2002 11:20:30 -0000 @@ -1039,5 +1039,5 @@ - output = `cd #{portdir} && #{shelljoin(*cmdargs)} -V PKGNAME -V IGNORE 2>&1`.to_a + output = `cd #{portdir} && #{shelljoin(*cmdargs)} -V PKGNAME -V IGNORE -V NO_IGNORE 2>&1`.to_a - if output.size != 2 + if output.size != 3 warning_message "Makefile of '#{origin}' is possibly broken:" @@ -1048,2 +1048,3 @@ ignore = output[1].chomp + no_ignore = !output[2].chomp.empty? @@ -1052,3 +1053,5 @@ STDERR.puts "\t" + ignore - raise IgnoreMarkError + raise IgnoreMarkError unless no_ignore + + warning_message "Proceeding anyway since NO_IGNORE is defined" end -- / /__ __ Akinori.org / MUSHA.org / ) ) ) ) / FreeBSD.org / Ruby-lang.org Akinori MUSHA aka / (_ / ( (__( @ iDaemons.org / and.or.jp "When I leave I don't know what I'm hoping to find When I leave I don't know what I'm leaving behind.."
On Fri, Aug 09, 2002 at 08:29:24PM +0900, Akinori MUSHA wrote: >At Fri, 09 Aug 2002 20:07:12 +0900, >I wrote: >> As I explained above, there is a case where user is terribly hurt and >> that's why the check was introduced. I'll alter the code to check >> NO_IGNORE also to solve your problem. > >Here's a patch, but beware, defining NO_IGNORE could be dangerous in >some cases (ports). Thank you very much. I have applied the patch. I assume we can expect it to show up in the next rev of portupgrade, right? -- AlanE KDE-FreeBSD Team
At Fri, 9 Aug 2002 13:28:00 -0400, Alan E wrote: > On Fri, Aug 09, 2002 at 08:29:24PM +0900, Akinori MUSHA wrote: > >At Fri, 09 Aug 2002 20:07:12 +0900, > >I wrote: > >> As I explained above, there is a case where user is terribly hurt and > >> that's why the check was introduced. I'll alter the code to check > >> NO_IGNORE also to solve your problem. > > > >Here's a patch, but beware, defining NO_IGNORE could be dangerous in > >some cases (ports). > > Thank you very much. I have applied the patch. I assume we can expect it > to show up in the next rev of portupgrade, right? Well, actually I attached the patch because I wanted you to test. :) If it works for you and satisfies your need, it will appear in the next release. Thanks, -- / /__ __ Akinori.org / MUSHA.org / ) ) ) ) / FreeBSD.org / Ruby-lang.org Akinori MUSHA aka / (_ / ( (__( @ iDaemons.org / and.or.jp "When I leave I don't know what I'm hoping to find When I leave I don't know what I'm leaving behind.."
On Sun, Aug 11, 2002 at 06:51:27PM +0900, Akinori MUSHA wrote: >At Fri, 9 Aug 2002 13:28:00 -0400, >Alan E wrote: >> On Fri, Aug 09, 2002 at 08:29:24PM +0900, Akinori MUSHA wrote: >> >At Fri, 09 Aug 2002 20:07:12 +0900, >> >I wrote: >> >> As I explained above, there is a case where user is terribly hurt and >> >> that's why the check was introduced. I'll alter the code to check >> >> NO_IGNORE also to solve your problem. >> > >> >Here's a patch, but beware, defining NO_IGNORE could be dangerous in >> >some cases (ports). >> >> Thank you very much. I have applied the patch. I assume we can expect it >> to show up in the next rev of portupgrade, right? > >Well, actually I attached the patch because I wanted you to test. :) >If it works for you and satisfies your need, it will appear in the >next release. Works perfectly. Thanks. Please add to release. 8=) -- AlanE KDE-FreeBSD Team
State Changed From-To: open->patched A fix will be included in the next release.
Please commit NO_IGNORE patch. It has been 3 weeks now.
State Changed From-To: patched->closed The port was updated with the patch integrated.