Page MenuHomeFreeBSD

Improve and extend sed_checked.sh
Needs ReviewPublic

Authored by mandree on Mar 31 2020, 3:53 PM.
Tags
None
Referenced Files
Unknown Object (File)
Sun, Mar 31, 4:31 AM
Unknown Object (File)
Feb 8 2024, 7:08 AM
Unknown Object (File)
Jan 26 2024, 6:57 PM
Unknown Object (File)
Dec 28 2023, 11:11 AM
Unknown Object (File)
Dec 22 2023, 11:17 PM
Unknown Object (File)
Dec 14 2023, 6:51 PM
Unknown Object (File)
Nov 15 2023, 9:26 AM
Unknown Object (File)
Nov 14 2023, 7:58 AM
Subscribers

Details

Summary

The current sed_checked.sh validates REINPLACE_CMD expressions in port Makefiles.
It has a few shortcomings that this diff strives to address:

  • the test is blind to whether a REINPLACE_CMD line is completely useless, or partially ineffective - report that because for some machine edits, a broader glob has been common practice
  • REINPLACE_CMD may suppress creation of backup files (-i ''), which invalidates this check, so warn about that
  • report found data immediately, and again through REWARNFILE.

Diff Detail

Repository
rP FreeBSD ports repository
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

mandree created this object with edit policy "committers (Project)".

same code, rediffed with "full" context

@portmgr @manu @mat @swills

Guys, this has been sitting for five weeks since the full context was provided, and almost three months since initial review request.
Can we get this moving? Can I get approval for commit? I've been using this for 10+ weeks.

No more complaints received in a month, can I commit?

This revision is now accepted and ready to land.Sep 23 2020, 4:56 PM

No more complaints received in a month, can I commit?

We do not have more complains, but we still do not agree with the change, so the answer is still no.

OK, then this will escalate to core@ for abuse because there has been no feedback as to why it's unacceptable in spite of solving practical problems.

In D24240#592496, @mat wrote:

No more complaints received in a month, can I commit?

We do not have more complains, but we still do not agree with the change, so the answer is still no.

Can you articulate a reason? A blanket "no" without an actionable explanation is frowned upon. The code looks fine. The concept seems on its face OK. What's the holdup?

This revision was automatically updated to reflect the committed changes.

since swills@ has reverted this as of r552736, reopen.

In the light of findings mentioned in D27954 we might want to add removal of .bak files, but other than that my proposal already looks in-depth at the REINPLACE_ARGS, to complain if something uses a nonstandard -i argument (backup suffix).