#146 Koji doc: discourage use of direct chained builds
Merged by gotmax23. Opened by adamwill.
fedora-docs/ adamwill/package-maintainer-docs koji-discourage-chain-build  into  main

Download 146.patch

Direct chained builds really aren't a good idea any more, now we
create updates for all Rawhide builds and apply gating tests to
some. If a build in the chain fails gating tests it will not be
pushed stable and the chain will fail.

This explains that, but also explains that you can use
chain-build --target to do a chain build to a side tag, and
this might be a nice workflow.

rebased onto 8453bc3cebf9b99f508ab109ac203f3489eadf51

rebased onto 0004fdc8db8116ae12df45c0617d0c67facc907e

Generally, this is good idea. But wouldn't it be better to put it the opposite way using the side tags by default and have there some remark that this could also be used without them, but leave it without too much details? IOW we should try to remove such warnings.

This could be the start of the chapter:

_
Sometimes you want to make sure that one build has succeeded before launching the next one, for example when you want to rebuild a package against a dependency that has just been rebuilt. This is called chain build and it works best with a side-tag._

...

_
Note: You can also use chain build without side-tags, but the usage is limited to Rawhide and has other downsides ....
_

yeah, that's a good point. I initially wrote this just to say DON'T USE CHAIN BUILDS! (more or less) then belatedly realized you can actually do them on side tags, so I did a late rewrite to cover that. Thinking it over from the start it'd probably be better to do it the way you suggest, indeed. I'll redo it. Thanks.

rebased onto e6deb22bdb199fd8fb793f2920cd94013c85bcf5

There, how's that?

`+fedpkg+` provides a `+chain-build+` command to help with this.
`+side-tag-name+`
that package will never reach _stable_,

Hmm, does asciidoc also support plain backticks for inline monspace/code blocks like markdown? I thought it was only the backtick with plus-sign syntax.

Hmm, does asciidoc also support plain backticks for inline monspace/code blocks like markdown? I thought it was only the backtick with plus-sign syntax.

Ah, I guess the plain backticks syntax allows formatting within the monospace text and may require extra escaping; the one with the plus is for literal monospace text. The rest of the document uses the plain backticks, so I guess keep them as is. Sorry for the noise.

that package will never reach _stable_,

I think whether you have a comma there is a subjective call in this construction. I prefer not having one. :)

that package will never reach _stable_,

I think whether you have a comma there is a subjective call in this construction. I prefer not having one. :)

Isn't it two independent clauses separated by a conjunction?

Anyways, this looks good to me, so I'll go ahead and merge it. Thanks for the improvement!

Pull-Request has been merged by gotmax23

Metadata