commons-parent maven pom highly broken?
Hi! I've never seen any other ASF project where it is such a torture to release. This is partly because the quality level is really high, but a big part of it is that we don't have a mature parent pom. I have no clue why we don't just use the common apache parent pom. I've NEVER experienced such issues like missing NOTICE and LICENSE with it. It's really much more solid than our own one. LieGrue, strub - To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
Re: commons-parent maven pom highly broken?
On 13 December 2014 at 12:26, Mark Struberg wrote: > Hi! > > I've never seen any other ASF project where it is such a torture to release. > This is partly because the quality level is really high, but a big part of it > is that we don't have a mature parent pom. Sorry, but I don't think that is at all relevant. Releasing is tedious because Maven does not really support non-Maven release directories (i.e. www.apache.org/dist/commons) so work-rounds are needed. This is true whatever the parent pom is used. It's partly also that Maven does some things well, but when adjustments are needed, it can be all but impossible to work out how to coax it to do what's needed. > > I have no clue why we don't just use the common apache parent pom. I've NEVER > experienced such issues like missing NOTICE and LICENSE with it. It's really > much more solid than our own one. Until very recently the ASF parent pom was broken, and with no sign of any fixes. For example, it did not allow override of the compiler plugin version, and there were a few other issues with it. There are stil issues with LICENSE and NOTICE, because the automatically created ones are not always correct. I'm not sure it creates the appropriate source and javadoc jars either. And it does not create decent manifests. The Commons Pom includes a lot of common setup that would otherwise have to be added to each component's parent pom. That is not an efficient way of proceeding. There are various other additional optional features in the Commons pom which are useful for testing. I think the CP pom works fine for single module projects. I'm not sure about multimodule projects, but is the ASF pom any better? Rather than raise unsubstantiated criticisms of the CP pom, file bugs for any problems and/or try fixing them. > LieGrue, > strub > > - > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > - To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
Re: commons-parent maven pom highly broken?
At least some skip should be removed since it breaks apache parent behavior. Le 13 déc. 2014 18:42, "sebb" a écrit : > On 13 December 2014 at 12:26, Mark Struberg wrote: > > Hi! > > > > I've never seen any other ASF project where it is such a torture to > release. > > This is partly because the quality level is really high, but a big part > of it is that we don't have a mature parent pom. > > Sorry, but I don't think that is at all relevant. > Releasing is tedious because Maven does not really support non-Maven > release directories (i.e. www.apache.org/dist/commons) so work-rounds > are needed. > This is true whatever the parent pom is used. > > It's partly also that Maven does some things well, but when > adjustments are needed, it can be all but impossible to work out how > to coax it to do what's needed. > > > > > I have no clue why we don't just use the common apache parent pom. I've > NEVER experienced such issues like missing NOTICE and LICENSE with it. It's > really much more solid than our own one. > > Until very recently the ASF parent pom was broken, and with no sign of > any fixes. > For example, it did not allow override of the compiler plugin version, > and there were a few other issues with it. > > There are stil issues with LICENSE and NOTICE, because the > automatically created ones are not always correct. > I'm not sure it creates the appropriate source and javadoc jars either. > > And it does not create decent manifests. > > The Commons Pom includes a lot of common setup that would otherwise > have to be added to each component's parent pom. > That is not an efficient way of proceeding. > > There are various other additional optional features in the Commons > pom which are useful for testing. > > I think the CP pom works fine for single module projects. > I'm not sure about multimodule projects, but is the ASF pom any better? > > Rather than raise unsubstantiated criticisms of the CP pom, file bugs > for any problems and/or try fixing them. > > > LieGrue, > > strub > > > > - > > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > > For additional commands, e-mail: dev-h...@commons.apache.org > > > > - > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > >
Re: commons-parent maven pom highly broken?
On 13 December 2014 at 17:51, Romain Manni-Bucau wrote: > At least some skip should be removed since it breaks apache parent behavior. No idea what you mean here - if you think there is a bug, file a JIRA issue. > Le 13 déc. 2014 18:42, "sebb" a écrit : > >> On 13 December 2014 at 12:26, Mark Struberg wrote: >> > Hi! >> > >> > I've never seen any other ASF project where it is such a torture to >> release. >> > This is partly because the quality level is really high, but a big part >> of it is that we don't have a mature parent pom. >> >> Sorry, but I don't think that is at all relevant. >> Releasing is tedious because Maven does not really support non-Maven >> release directories (i.e. www.apache.org/dist/commons) so work-rounds >> are needed. >> This is true whatever the parent pom is used. >> >> It's partly also that Maven does some things well, but when >> adjustments are needed, it can be all but impossible to work out how >> to coax it to do what's needed. >> >> > >> > I have no clue why we don't just use the common apache parent pom. I've >> NEVER experienced such issues like missing NOTICE and LICENSE with it. It's >> really much more solid than our own one. >> >> Until very recently the ASF parent pom was broken, and with no sign of >> any fixes. >> For example, it did not allow override of the compiler plugin version, >> and there were a few other issues with it. >> >> There are stil issues with LICENSE and NOTICE, because the >> automatically created ones are not always correct. >> I'm not sure it creates the appropriate source and javadoc jars either. >> >> And it does not create decent manifests. >> >> The Commons Pom includes a lot of common setup that would otherwise >> have to be added to each component's parent pom. >> That is not an efficient way of proceeding. >> >> There are various other additional optional features in the Commons >> pom which are useful for testing. >> >> I think the CP pom works fine for single module projects. >> I'm not sure about multimodule projects, but is the ASF pom any better? >> >> Rather than raise unsubstantiated criticisms of the CP pom, file bugs >> for any problems and/or try fixing them. >> >> > LieGrue, >> > strub >> > >> > - >> > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >> > For additional commands, e-mail: dev-h...@commons.apache.org >> > >> >> - >> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >> For additional commands, e-mail: dev-h...@commons.apache.org >> >> - To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
Re: commons-parent maven pom highly broken?
@Sebb: when trying to release jcs I had a lot of trouble just because [commons] parent pom doesn't follow asf rules (remote resource plugin is deactivated where it is correct in apache pom for instance) Romain Manni-Bucau @rmannibucau http://www.tomitribe.com http://rmannibucau.wordpress.com https://github.com/rmannibucau 2014-12-13 18:56 GMT+01:00 sebb : > On 13 December 2014 at 17:51, Romain Manni-Bucau > wrote: >> At least some skip should be removed since it breaks apache parent behavior. > > No idea what you mean here - if you think there is a bug, file a JIRA issue. > >> Le 13 déc. 2014 18:42, "sebb" a écrit : >> >>> On 13 December 2014 at 12:26, Mark Struberg wrote: >>> > Hi! >>> > >>> > I've never seen any other ASF project where it is such a torture to >>> release. >>> > This is partly because the quality level is really high, but a big part >>> of it is that we don't have a mature parent pom. >>> >>> Sorry, but I don't think that is at all relevant. >>> Releasing is tedious because Maven does not really support non-Maven >>> release directories (i.e. www.apache.org/dist/commons) so work-rounds >>> are needed. >>> This is true whatever the parent pom is used. >>> >>> It's partly also that Maven does some things well, but when >>> adjustments are needed, it can be all but impossible to work out how >>> to coax it to do what's needed. >>> >>> > >>> > I have no clue why we don't just use the common apache parent pom. I've >>> NEVER experienced such issues like missing NOTICE and LICENSE with it. It's >>> really much more solid than our own one. >>> >>> Until very recently the ASF parent pom was broken, and with no sign of >>> any fixes. >>> For example, it did not allow override of the compiler plugin version, >>> and there were a few other issues with it. >>> >>> There are stil issues with LICENSE and NOTICE, because the >>> automatically created ones are not always correct. >>> I'm not sure it creates the appropriate source and javadoc jars either. >>> >>> And it does not create decent manifests. >>> >>> The Commons Pom includes a lot of common setup that would otherwise >>> have to be added to each component's parent pom. >>> That is not an efficient way of proceeding. >>> >>> There are various other additional optional features in the Commons >>> pom which are useful for testing. >>> >>> I think the CP pom works fine for single module projects. >>> I'm not sure about multimodule projects, but is the ASF pom any better? >>> >>> Rather than raise unsubstantiated criticisms of the CP pom, file bugs >>> for any problems and/or try fixing them. >>> >>> > LieGrue, >>> > strub >>> > >>> > - >>> > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>> > For additional commands, e-mail: dev-h...@commons.apache.org >>> > >>> >>> - >>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>> For additional commands, e-mail: dev-h...@commons.apache.org >>> >>> > > - > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > - To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
Re: commons-parent maven pom highly broken?
File a JIRA with exact details. On 13 December 2014 at 18:03, Romain Manni-Bucau wrote: > @Sebb: when trying to release jcs I had a lot of trouble just because > [commons] parent pom doesn't follow asf rules (remote resource plugin > is deactivated where it is correct in apache pom for instance) > > > Romain Manni-Bucau > @rmannibucau > http://www.tomitribe.com > http://rmannibucau.wordpress.com > https://github.com/rmannibucau > > > 2014-12-13 18:56 GMT+01:00 sebb : >> On 13 December 2014 at 17:51, Romain Manni-Bucau >> wrote: >>> At least some skip should be removed since it breaks apache parent behavior. >> >> No idea what you mean here - if you think there is a bug, file a JIRA issue. >> >>> Le 13 déc. 2014 18:42, "sebb" a écrit : >>> On 13 December 2014 at 12:26, Mark Struberg wrote: > Hi! > > I've never seen any other ASF project where it is such a torture to release. > This is partly because the quality level is really high, but a big part of it is that we don't have a mature parent pom. Sorry, but I don't think that is at all relevant. Releasing is tedious because Maven does not really support non-Maven release directories (i.e. www.apache.org/dist/commons) so work-rounds are needed. This is true whatever the parent pom is used. It's partly also that Maven does some things well, but when adjustments are needed, it can be all but impossible to work out how to coax it to do what's needed. > > I have no clue why we don't just use the common apache parent pom. I've NEVER experienced such issues like missing NOTICE and LICENSE with it. It's really much more solid than our own one. Until very recently the ASF parent pom was broken, and with no sign of any fixes. For example, it did not allow override of the compiler plugin version, and there were a few other issues with it. There are stil issues with LICENSE and NOTICE, because the automatically created ones are not always correct. I'm not sure it creates the appropriate source and javadoc jars either. And it does not create decent manifests. The Commons Pom includes a lot of common setup that would otherwise have to be added to each component's parent pom. That is not an efficient way of proceeding. There are various other additional optional features in the Commons pom which are useful for testing. I think the CP pom works fine for single module projects. I'm not sure about multimodule projects, but is the ASF pom any better? Rather than raise unsubstantiated criticisms of the CP pom, file bugs for any problems and/or try fixing them. > LieGrue, > strub > > - > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > - To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org >> >> - >> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >> For additional commands, e-mail: dev-h...@commons.apache.org >> > > - > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > - To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
Re: commons-parent maven pom highly broken?
done https://issues.apache.org/jira/browse/COMMONSSITE-79 Romain Manni-Bucau @rmannibucau http://www.tomitribe.com http://rmannibucau.wordpress.com https://github.com/rmannibucau 2014-12-13 19:13 GMT+01:00 sebb : > File a JIRA with exact details. > > On 13 December 2014 at 18:03, Romain Manni-Bucau > wrote: >> @Sebb: when trying to release jcs I had a lot of trouble just because >> [commons] parent pom doesn't follow asf rules (remote resource plugin >> is deactivated where it is correct in apache pom for instance) >> >> >> Romain Manni-Bucau >> @rmannibucau >> http://www.tomitribe.com >> http://rmannibucau.wordpress.com >> https://github.com/rmannibucau >> >> >> 2014-12-13 18:56 GMT+01:00 sebb : >>> On 13 December 2014 at 17:51, Romain Manni-Bucau >>> wrote: At least some skip should be removed since it breaks apache parent behavior. >>> >>> No idea what you mean here - if you think there is a bug, file a JIRA issue. >>> Le 13 déc. 2014 18:42, "sebb" a écrit : > On 13 December 2014 at 12:26, Mark Struberg wrote: > > Hi! > > > > I've never seen any other ASF project where it is such a torture to > release. > > This is partly because the quality level is really high, but a big part > of it is that we don't have a mature parent pom. > > Sorry, but I don't think that is at all relevant. > Releasing is tedious because Maven does not really support non-Maven > release directories (i.e. www.apache.org/dist/commons) so work-rounds > are needed. > This is true whatever the parent pom is used. > > It's partly also that Maven does some things well, but when > adjustments are needed, it can be all but impossible to work out how > to coax it to do what's needed. > > > > > I have no clue why we don't just use the common apache parent pom. I've > NEVER experienced such issues like missing NOTICE and LICENSE with it. > It's > really much more solid than our own one. > > Until very recently the ASF parent pom was broken, and with no sign of > any fixes. > For example, it did not allow override of the compiler plugin version, > and there were a few other issues with it. > > There are stil issues with LICENSE and NOTICE, because the > automatically created ones are not always correct. > I'm not sure it creates the appropriate source and javadoc jars either. > > And it does not create decent manifests. > > The Commons Pom includes a lot of common setup that would otherwise > have to be added to each component's parent pom. > That is not an efficient way of proceeding. > > There are various other additional optional features in the Commons > pom which are useful for testing. > > I think the CP pom works fine for single module projects. > I'm not sure about multimodule projects, but is the ASF pom any better? > > Rather than raise unsubstantiated criticisms of the CP pom, file bugs > for any problems and/or try fixing them. > > > LieGrue, > > strub > > > > - > > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > > For additional commands, e-mail: dev-h...@commons.apache.org > > > > - > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > > >>> >>> - >>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>> For additional commands, e-mail: dev-h...@commons.apache.org >>> >> >> - >> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >> For additional commands, e-mail: dev-h...@commons.apache.org >> > > - > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > - To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
Re: commons-parent maven pom highly broken?
Just file an JIRA if something doesn't work and you are done. > Releasing is tedious because Maven does not really support non-Maven > release directories (i.e. www.apache.org/dist/commons) This is totally non-ASF style. Just use the standard stuff like every other ASF project and you are again done. > Until very recently the ASF parent pom was broken, and with no sign of > any fixes.To me'not fixing anything' would mean that all the apache-parent > releases would be useless? Or you just did not check it? > There are stil issues with LICENSE and NOTICE, because the > automatically created ones are not always correct. The rule is straight forward. If you provide your own one then it will be used. If not, then a standard one will be packaged. > The Commons Pom includes a lot of common setup that would otherwise > have to be added to each component's parent pom. And that would be? Most of the stuff is taken over from the old ant builds and nowadays do not reflect reality anymore. > There are various other additional optional features in the Commons > pom which are useful for testing. what more than testng OR junit (depending on the project) + maybe mock tools do you need? > I think the CP pom works fine for single module projects. > I'm not sure about multimodule projects, but is the ASF pom any better? Well, most of todays projects are multi-module ones. And this works perfectly fine with the ASF parent pom. LieGrue, strub > On Saturday, 13 December 2014, 18:41, sebb wrote: > > On 13 December 2014 at 12:26, Mark Struberg wrote: >> Hi! >> >> I've never seen any other ASF project where it is such a torture to > release. >> This is partly because the quality level is really high, but a big part of > it is that we don't have a mature parent pom. > > Sorry, but I don't think that is at all relevant. > Releasing is tedious because Maven does not really support non-Maven > release directories (i.e. www.apache.org/dist/commons) so work-rounds > are needed. > This is true whatever the parent pom is used. > > It's partly also that Maven does some things well, but when > adjustments are needed, it can be all but impossible to work out how > to coax it to do what's needed. > >> >> I have no clue why we don't just use the common apache parent pom. > I've NEVER experienced such issues like missing NOTICE and LICENSE with it. > It's really much more solid than our own one. > > Until very recently the ASF parent pom was broken, and with no sign of > any fixes. > For example, it did not allow override of the compiler plugin version, > and there were a few other issues with it. > > There are stil issues with LICENSE and NOTICE, because the > automatically created ones are not always correct. > I'm not sure it creates the appropriate source and javadoc jars either. > > And it does not create decent manifests. > > The Commons Pom includes a lot of common setup that would otherwise > have to be added to each component's parent pom. > That is not an efficient way of proceeding. > > There are various other additional optional features in the Commons > pom which are useful for testing. > > I think the CP pom works fine for single module projects. > I'm not sure about multimodule projects, but is the ASF pom any better? > > Rather than raise unsubstantiated criticisms of the CP pom, file bugs > for any problems and/or try fixing them. > > >> LieGrue, >> strub >> >> - >> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >> For additional commands, e-mail: dev-h...@commons.apache.org >> > - To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
[math] preparing release 3.4 ?
Hi all, I would like to push version 3.4 soon. There are 3 remaining issues in JIRA. I propose to postpone 1066 to next version as there are no news of the original poster about refactoring the patch. I don't know the status about 1105, could it be done in the next few days or if not could it be postponed to next version? Concerning 1132, I would say resolve it as WONT'FIX for now, and perhaps reopen it later if needed. There have been lots of discussions on the list a few months ago, and the current status seem to be sufficient for everybody (at least we did not get further complaints). I am willing to be the release manager and am already doing some dry runs on it. My goal is really to do RC1 as soon as possible, and no later than middle of next week. best regards Luc - To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
Re: [math] preparing release 3.4 ?
On Sat, 13 Dec 2014 21:51:30 +0100, Luc Maisonobe wrote: Hi all, I would like to push version 3.4 soon. There are 3 remaining issues in JIRA. I propose to postpone 1066 to next version as there are no news of the original poster about refactoring the patch. I don't know the status about 1105, could it be done in the next few days or if not could it be postponed to next version? Concerning 1132, I would say resolve it as WONT'FIX for now, and perhaps reopen it later if needed. There have been lots of discussions on the list a few months ago, and the current status seem to be sufficient for everybody (at least we did not get further complaints). I am willing to be the release manager and am already doing some dry runs on it. My goal is really to do RC1 as soon as possible, and no later than middle of next week. There are two issues for which I have code to be committed: * MATH-1172 There was no reaction to http://markmail.org/message/cfrtdxm3y5o5adgj * MATH-1173 I have coded a replacement to "TricubicSplineInterpolator" to go with "TricubicInterpolatingFunction" (i.e. using finite difference approximation to the derivatives), but I don't have unit tests yet... Suggestions welcome. Best, Gilles - To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org