Re: [pde-dev] Discontinue API Tools ( was Re: What do API tools want her

[Vikas Chandra <Vikas.Chandra@xxxxxxxxxx>]

Hi Christoph,

Regarding the annoying bug, here is my comment - https://github.com/eclipse-pde/eclipse.pde/pull/397#issuecomment-1309040721
Further to this, we can discuss on https://github.com/eclipse-pde/eclipse.pde/issues/232

Regarding the list with label api_tools label, I will have a quick look at this over this weekend and see if we can incorporate
anything for 4.26 RC1. 

Thanks and Regards,
Vikas
Eclipse PDE Lead
-- 

 

On 09/11/22, 11:03 AM, "pde-dev on behalf of Christoph Läubrich" <pde-dev-bounces@xxxxxxxxxxx on behalf of laeubi@xxxxxxxxxxxxxx> wrote:

    Hi Vikas,

    thats great to hear, I have labeled all api-tools issues (that I know) 
    at github:

    https://github.com/eclipse-pde/eclipse.pde/issues?q=is%3Aopen+is%3Aissue+label%3Aapi-tools  

    I think the most prominent/annoying is the "611 failures" problem 
    described here (it has a quite long history now):

    https://github.com/eclipse-pde/eclipse.pde/issues/232  

    so if you can help out there it would be very welcome!


    Am 09.11.22 um 05:01 schrieb Vikas Chandra:
    >>>the guys that had deep knowledge of API tools codebase, design and etc
    > 
    > Just saw this. I have fixed close to 200 bugs in API tools in last 5-7 
    > years ( *including many*
    > 
    > *enhancements*) .  I believe I have decent knowledge of API code base 
    > and design
    > 
    > I recommend putting blocker/regression tag on any issue that needs more 
    > attention.
    > 
    > My priorities in API tools are to not have a regression or a general 
    > blocker issue. Also I
    > 
    > try to have at least 1-2 enhancements per year ( in API tools).
    > 
    > If some other bug needs to be addressed, a quality patch should be 
    > provided or a “/constructive/”
    > 
    > discussion should be initiated.
    > 
    > Thanks and Regards,
    > 
    > Vikas
    > 
    > Eclipse PDE Lead
    > 
    > -- 
    > 
    > *From: *pde-dev <pde-dev-bounces@xxxxxxxxxxx> on behalf of Aleksandar 
    > Kurtakov <akurtako@xxxxxxxxxx>
    > *Reply to: *"Eclipse PDE general developers list." <pde-dev@xxxxxxxxxxx>
    > *Date: *Monday, 7 November 2022 at 1:29 PM
    > *To: *"Eclipse PDE general developers list." <pde-dev@xxxxxxxxxxx>
    > *Subject: *[EXTERNAL] Re: [pde-dev] Discontinue API Tools ( was Re: What 
    > do API tools want here?)
    > 
    > On Mon, Nov 7, 2022 at 9: 52 AM Ed Merks <ed. merks@ gmail. com> wrote: 
    > It seems to me that we often tell people who complain about a problem 
    > that a PR would be most welcome.    In this case where we have a problem 
    > with a shoe and "we"
    > 
    > ZjQcmQRYFpfptBannerStart
    > 
    > *This Message Is From an External Sender *
    > 
    > This message came from outside your organization.
    > 
    > ZjQcmQRYFpfptBannerEnd
    > 
    > On Mon, Nov 7, 2022 at 9:52 AM Ed Merks <ed.merks@xxxxxxxxx 
    > <mailto:ed.merks@xxxxxxxxx>> wrote:
    > 
    >     It seems to me that we often tell people who complain about a
    >     problem that a PR would be most welcome.   In this case where we
    >     have a problem with a shoe and "we" are the shoemaker of that shoe,
    >     that suggestion seems even more appropriate.  We are PDE...
    > 
    >     In any case, you opened
    >     https://github.com/eclipse-pde/eclipse.pde/issues/378  
    >     <https://github.com/eclipse-pde/eclipse.pde/issues/378  > and the
    >     issue was determined to be a surprising and nasty corner case of our
    >     own manufacture.  So I think we can now safely call off the
    >     demolition crew.
    > 
    >     -----------------
    > 
    >     It is frustrating that there is little in the way of responsiveness
    >     when asking questions.  For example I *know *there is a bug in
    >     org.eclipse.pde.api.tools.internal.builder.Reference.getParameterList(), and I have a fix for it (I think).  But I don't know what exactly that method is supposed to do, so I asked in the bug where that method was introduced:
    > 
    >     https://bugs.eclipse.org/bugs/show_bug.cgi?id=332767
    >     <https://bugs.eclipse.org/bugs/show_bug.cgi?id=332767>
    > 
    >     That was almost a month ago...
    > 
    > Unfortunately, the guys that had deep knowledge of API tools codebase, 
    > design and etc. haven't been working with us for years. As we are the 
    > people with commit rights, we are the "experts" so acting according to 
    > your best knowledge is everything we can do.
    > 
    >     On 06.11.2022 07:09, Christoph Läubrich wrote:
    > 
    >          > I understand and agree that the current situation is
    >         unpleasant or
    >          > better annoying.
    > 
    >         611 failing test anyone? ;-)
    > 
    >         That's just another example, we have random failures, none seem
    >         to know why or what can be done even though we have tried some
    >         things but it still occurs.
    > 
    >          > In case it would require more time/effort to adjust the
    >         existing code
    >          > base to the new tooling (e.g. remove all required-bundles) it
    >         could be
    >          > simpler to fix the PDE API-tools instead.
    > 
    >         For sure it depends on the tooling, and I won't recommend to
    >         just migrate everything, but it seems code debt for API tools is
    >         increasing and actually no one "simply fix it" even simple(?)
    >         ones like [1] just hanging around.
    > 
    >         So it is better to decide if there is a future for API tools or
    >         if we better look *now* for alternatives and let user know there
    >         is most likely no more support for them (similar for pde-build
    >         that seems unmaintained for a long time as well!).
    > 
    > 
    > 
    >         [1] https://github.com/eclipse-pde/eclipse.pde/issues/124  
    >         <https://github.com/eclipse-pde/eclipse.pde/issues/124  >
    > 
    > 
    > 
    >         Am 05.11.22 um 22:19 schrieb Hannes Wellmann:
    > 
    >             I understand and agree that the current situation is
    >             unpleasant or better annoying.
    >             Nevertheless we need some form of API tooling and I don't
    >             know if other techniques and tools can deal better with the
    >             PDE specialties like re-exported Required-Bundles (which is
    >             likely causing your issues)?
    >             What would be the alternatives in this case?
    >             In case it would require more time/effort to adjust the
    >             existing code base to the new tooling (e.g. remove all
    >             required-bundles) it could be simpler to fix the PDE
    >             API-tools instead.
    >             Using Re-export in general is bad and I fully agree/support
    >             improving PDE's Import-Package support in general but since
    >             probably many rely on re-exported require-bundle that it
    >             will take some time and discussion to get rid of that and
    >             will probably not work until there is a equivalent
    >             convenient solution available (like automatically calculated
    >             Import-Packages)
    >             But my blind guess is that other tools (probably from BND?)
    >             would also have their difficulties with that.
    >               > as even the build-setup is horrible complex and hard to
    >             perform including calling helper mojos and ant scripts.
    >             Since we are working on a API-tools plugin at Tycho, that
    >             problem should no longer exists in the not to distance
    >             future (I hope to have time to continue my work on that soon).
    >             Furthermore I had the intention to improve PDE's API tools
    >             in the future with regard to detecting necessary version
    >             bumps in Features and Plugins already in the IDE.
    >             That being said, if there is a 'drop-in' replacement for
    >             PDE's API tools I would not object to use that instead.
    >             But I would not assume that such immediate drop-in exists,
    >             but please correct me if I'm wrong.
    >             Nevertheless, if a suitable library exists we could consider
    >             to use that in the core of PDE's API tools and add
    >             reasonable PDE specialties on top of that.
    >             Greetings,
    >             Hannes
    >             *Gesendet:* Samstag, 05. November 2022 um 15:31 Uhr
    >             *Von:* "Christoph Läubrich" <laeubi@xxxxxxxxxxxxxx>
    >             <mailto:laeubi@xxxxxxxxxxxxxx>
    >             *An:* pde-dev@xxxxxxxxxxx <mailto:pde-dev@xxxxxxxxxxx>
    >             *Betreff:* [pde-dev] Discontinue API Tools ( was Re: What do
    >             API tools want here?)
    >             So now three month later, still no one know why API tools
    >             complains, or
    >             even what the error message means. Ed has done some analysis
    >             and even
    >             found that all quick fixes are broken here and even adding
    >             code in mini
    >             steps do not help any further.
    > 
    >             For me the question is if there is no support, no
    >             maintenance or
    >             development available if it is time to discontinue API
    >             tools, as even
    >             the build-setup is horrible complex and hard to perform
    >             including
    >             calling helper mojos and ant scripts.
    > 
    >             SO if PDE itself can't master its own API tools I think its
    >             time to face
    >             the truth and recommend to stop using it and migrate to
    >             other techniques.
    > 
    >             Am 06.08.22 um 08:16 schrieb Christoph Läubrich:
    >               > Hi PDE-Devs,
    >               >
    >               > I have a strange error from the API tools and no clue
    >             what it want here,
    >               > the error is:
    >               >
    >               > The type
    >             org.eclipse.pde.ui.launcher.AbstractPDELaunchConfiguration has
    >               > been removed from org.eclipse.pde.ui_3.14.0
    >               >
    >               > see build [1]
    >               > and PR [2]
    >               >
    >               > Any hint would be welcome ... I think it would be the
    >             best to directly
    >               > add any comments to the PR.
    >               >
    >               >
    >               > [1]
    >             https://ci.eclipse.org/pde/job/eclipse.pde.ui/job/PR-214/6/consoleFull <https://ci.eclipse.org/pde/job/eclipse.pde.ui/job/PR-214/6/consoleFull> <https://ci.eclipse.org/pde/job/eclipse.pde.ui/job/PR-214/6/consoleFull> <https://ci.eclipse.org/pde/job/eclipse.pde.ui/job/PR-214/6/consoleFull>
    >               > [2] https://github.com/eclipse-pde/eclipse.pde/pull/214  
    >             <https://github.com/eclipse-pde/eclipse.pde/pull/214  >
    >             <https://github.com/eclipse-pde/eclipse.pde/pull/214  >
    >             <https://github.com/eclipse-pde/eclipse.pde/pull/214  >
    >               > _______________________________________________
    >               > pde-dev mailing list
    >               > pde-dev@xxxxxxxxxxx <mailto:pde-dev@xxxxxxxxxxx>
    >               > To unsubscribe from this list, visit
    >               > https://www.eclipse.org/mailman/listinfo/pde-dev
    >             <https://www.eclipse.org/mailman/listinfo/pde-dev>
    >             <https://www.eclipse.org/mailman/listinfo/pde-dev>
    >             <https://www.eclipse.org/mailman/listinfo/pde-dev>
    >             _______________________________________________
    >             pde-dev mailing list
    >             pde-dev@xxxxxxxxxxx <mailto:pde-dev@xxxxxxxxxxx>
    >             To unsubscribe from this list, visit
    >             https://www.eclipse.org/mailman/listinfo/pde-dev
    >             <https://www.eclipse.org/mailman/listinfo/pde-dev>
    >             <https://www.eclipse.org/mailman/listinfo/pde-dev>
    >             <https://www.eclipse.org/mailman/listinfo/pde-dev>
    > 
    >             _______________________________________________
    >             pde-dev mailing list
    >             pde-dev@xxxxxxxxxxx <mailto:pde-dev@xxxxxxxxxxx>
    >             To unsubscribe from this list, visit
    >             https://www.eclipse.org/mailman/listinfo/pde-dev
    >             <https://www.eclipse.org/mailman/listinfo/pde-dev>
    > 
    >         _______________________________________________
    >         pde-dev mailing list
    >         pde-dev@xxxxxxxxxxx <mailto:pde-dev@xxxxxxxxxxx>
    >         To unsubscribe from this list, visit
    >         https://www.eclipse.org/mailman/listinfo/pde-dev
    >         <https://www.eclipse.org/mailman/listinfo/pde-dev>
    > 
    >     _______________________________________________
    >     pde-dev mailing list
    >     pde-dev@xxxxxxxxxxx <mailto:pde-dev@xxxxxxxxxxx>
    >     To unsubscribe from this list, visit
    >     https://www.eclipse.org/mailman/listinfo/pde-dev
    >     <https://www.eclipse.org/mailman/listinfo/pde-dev>
    > 
    > 
    > 
    > -- 
    > 
    > Aleksandar Kurtakov
    > 
    > Red Hat Eclipse Team
    > 
    > 
    > _______________________________________________
    > pde-dev mailing list
    > pde-dev@xxxxxxxxxxx
    > To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/pde-dev
    _______________________________________________
    pde-dev mailing list
    pde-dev@xxxxxxxxxxx
    To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/pde-dev