https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=218023
Unbreak lang/picoc by fixing MASTER_SITES
Pass maintainership to submitter
Fix WWW URL
Differential D10119
Unbreak lang/picoc krion on Mar 23 2017, 11:46 AM. Authored by Tags None Referenced Files
Subscribers None
Details
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=218023 Unbreak lang/picoc by fixing MASTER_SITES
Diff Detail
Event TimelineComment Actions Seeing all the changes in the patches, it is not the same as the previous 2.1 version, so PORTVERSION should be updated, maybe 2.1g<commit-date>, doing that, the commit message should look more like: Update to a later commit and follow upstream change Pass maintainership to submitter Don't mention the WWW change in the commit message, it is implied by the fact that upstream changed. Comment Actions I re-did the patches with "make makepatch" to make portlint happy, I can point it out in commit log. Comment Actions Did you approve or left for changes this revision, I couldn't identify it from your comment :-) Comment Actions You can't have just run "make patch" then "make makepatch" there are too many things changing to be the result of only that. Comment Actions I made them manually and did makepatch, what could be wrong with it? Anyway, I don't care and will bring back old files/* versions. Comment Actions What I mean is that if you need to change the patches manually, then it is not the same upstream version, and then, PORTVERSION must change. Comment Actions Sorry, did I understand it right, if I make patches conform with portlint(1) and run "make makepatch" I've to bump PORTVERSION ?? Where's documented, what policy applies to it and what the reason behind it? Comment Actions No, running make makepatch to get portlint to stop complaining (which is something you should not do) does not require you to bump PORTREVISION. On the other hand, like I was saying, this is not what is happening here, the patches are changing way to much for it to be the he result of just make makepatch. When you run make makepatch, what may change is:
That's it. Here, you have hunks going away, and hunks being changed. Comment Actions Mat, and now look at this picture, I totally understand what you mean, I totally understand this point. What we're doing here is - discussing the need to change the patches to shut up portlint warnings or not. I totally agree - a) it could lead to unneeded problems if you're not careful enough b) there's not much sense to change it because we change just hunk positions and c) we can interfere with spaces and tabs and break things. Now let's calculate the moves I've done before sending it to you: Should I calculate the most valuable asset we had and which we've lost for these reviews and changes - time. Instead of - "no-go, because of a) b) c) - redo" or "go ahead, approved". Mat, the project degraded a lot because of all of that talkie-talkie stuff, I'm noticing it by reading our mailing lists, because "everyone and my dog" would like to comment on everything around what moves and what we see. Sorry for that, maybe, rude points, but that what I see in the project currently and it makes me worry :) Comment Actions Ok, let's backtrack a moment. What is happening here is upstream changed, and so has the distfile, it went from picoc-2.1.tar.bz2 that was ~70kB to zsaleeba-picoc-2.1-4555e84_GH0.tar.gz that is about 610kB. So, the content of the distfile changed, a lot. We're in the "What is the proper procedure for updating the checksum for a port's distfile when the file changes without a version change" case here. As the distfile changed a lot, I'm nitpicking about the patches because it feels like the code changed. There are two cases:
I don't know that you checked that the distfile content is the same or not, you have not spoken about it, this is, again, why I try to make sure the patches changes are normal, which is why I have not said "you need to change this" or "this is ok", I'm still waiting to see if something has to change or not. As for the reviews we do after the fact, yes, the project has changed, it has evolved. Getting what you do reviewed by your peers is a good thing, if someone raises a point, it means that something is wrong/not great/not clear, and that it could maybe be done in a better way. I do not feel at all that it has degraded, some things may take a bit more time to do, but in the end, it means they are done in a better way and we all learn things. Comment Actions My intention was to bump PORTEPOCH for that, as we did all the time after rerolling distfile. In terms of review, pkg, poudriere you're right - project evolved significantly, but in terms of intercommunications and relations between people inside of project it degraded dramatically Comment Actions I'm sure you meant PORTREVISION. But that is only the case if the distfile is still the same version. Here, it is not the case, the distfile has 2.1 in its name. But in that case, any change you mean to do to the port when you commit it must be present here, in this review, because you are still being mentored, and you cannot commit anything that has not been approved. Until some of this are done, here, in this review, it cannot go forward.. |