Re: SGE-Pygame mysteriously downgraded on PyPI
l...@gnu.org (Ludovic Courtès) writes: > Hello, > > Kei Kebreau skribis: > >> l...@gnu.org (Ludovic Courtès) writes: >> >>> Hi! >>> >>> Kei Kebreau skribis: >>> I don't know if anyone else has noticed, but it seems that our sge-pygame package is failing to build because the package version on PyPI has downgraded from 1.5.1 to 1.5. >>> >>> How can that downgrade lead to a build failure of our package though? >>> >> >> Our sge-pygame definition uses the source code from PyPI, not from GNU >> Savannah and its mirrors where the project seems to be hosted. Because >> version >> 1.5.1 is no longer available from PyPI servers, our package fails to >> build. >> >> I can change the definition to get the tarball from Savannah and its >> mirrors because we already do that for python-xsge, a similar package by >> the same developer. > > Ah, got it. > > Well, I suggest you should do whichever looks most sensible to you. If > the downgrade on PyPI was motivated, but we should probably follow > suite. Otherwise, we could simply use the 1.5.1 tarball from Savannah. > > Thoughts? > > Ludo’. For any who may be following this thread, I've elected to use the 1.5.1 tarball from Savannah. The code change is in commit 7f652029908f23d85629e2932cc70585ba3e8a1b. signature.asc Description: PGP signature
Re: SGE-Pygame mysteriously downgraded on PyPI
Hello, Kei Kebreau skribis: > l...@gnu.org (Ludovic Courtès) writes: > >> Hi! >> >> Kei Kebreau skribis: >> >>> I don't know if anyone else has noticed, but it seems that our >>> sge-pygame package is failing to build because the package version on >>> PyPI has downgraded from 1.5.1 to 1.5. >> >> How can that downgrade lead to a build failure of our package though? >> > > Our sge-pygame definition uses the source code from PyPI, not from GNU > Savannah and its mirrors where the project seems to be hosted. Because version > 1.5.1 is no longer available from PyPI servers, our package fails to > build. > > I can change the definition to get the tarball from Savannah and its > mirrors because we already do that for python-xsge, a similar package by > the same developer. Ah, got it. Well, I suggest you should do whichever looks most sensible to you. If the downgrade on PyPI was motivated, but we should probably follow suite. Otherwise, we could simply use the 1.5.1 tarball from Savannah. Thoughts? Ludo’.
Re: SGE-Pygame mysteriously downgraded on PyPI
l...@gnu.org (Ludovic Courtès) writes: > Hi! > > Kei Kebreau skribis: > >> I don't know if anyone else has noticed, but it seems that our >> sge-pygame package is failing to build because the package version on >> PyPI has downgraded from 1.5.1 to 1.5. > > How can that downgrade lead to a build failure of our package though? > Our sge-pygame definition uses the source code from PyPI, not from GNU Savannah and its mirrors where the project seems to be hosted. Because version 1.5.1 is no longer available from PyPI servers, our package fails to build. I can change the definition to get the tarball from Savannah and its mirrors because we already do that for python-xsge, a similar package by the same developer. signature.asc Description: PGP signature
Re: SGE-Pygame mysteriously downgraded on PyPI
Hi! Kei Kebreau skribis: > I don't know if anyone else has noticed, but it seems that our > sge-pygame package is failing to build because the package version on > PyPI has downgraded from 1.5.1 to 1.5. How can that downgrade lead to a build failure of our package though? > I'm not sure why this has happened, as the recommended version on the > SGE-Pygame website is still 1.5.1. This gives me two questions: Indeed: --8<---cut here---start->8--- $ guix refresh python-sge-pygame gnu/packages/game-development.scm:279:13: warning: 1.5.1 is greater than the latest known version of python-sge-pygame (1.5) --8<---cut here---end--->8--- > 1. Do we have a protocol for dealing with this sort of issue? No. > 2. As I've had no luck searching on my own, where can one keep track of > the reasons behind changes like this on PyPI? No idea. That’s a weird situation! Ludo’.
SGE-Pygame mysteriously downgraded on PyPI
Hello Guix, I don't know if anyone else has noticed, but it seems that our sge-pygame package is failing to build because the package version on PyPI has downgraded from 1.5.1 to 1.5. I'm not sure why this has happened, as the recommended version on the SGE-Pygame website is still 1.5.1. This gives me two questions: 1. Do we have a protocol for dealing with this sort of issue? 2. As I've had no luck searching on my own, where can one keep track of the reasons behind changes like this on PyPI? signature.asc Description: PGP signature