Hi Paul, On 31/01/20 2:01 am, Paul Wise wrote: > On Thu, Jan 30, 2020 at 8:22 PM Utkarsh Gupta wrote: > >> Ideally, we shouldn't go for a mass commit because we don't know if >> everything would be working fine with it, or do we? > > I think all Perl module packages are unlikely to produce different > packages with R³: no, but Perl app packages might. Yep, that's (partly) what I wanted to convey. >> I've been doing it manually so far via cme. And it works fine, of course. >> Yes, it's a little extra work but it ensures that everything works fine. > > How did you test the changes? Please note that when I said "I've been doing it manually via cme ..", I meant not just for Perl packages, but also for Ruby, Golang, et al. And for testing, exactly how you described below. Building a package without that line and then building with it and then using diffoscope to compare the binaries. (of course, there are some (mostly Ruby) packages that I didn't really diffoscope'd since I was confident that it won't break stuff). > I would recommend building twice, once without R³: no and then once > with it, and use diffoscope to compare the resulting source/binary > packages. If there are any differences to the binary packages then > those differences need to be fixed before using R³: no. Best, Utkarsh
Attachment:
signature.asc
Description: OpenPGP digital signature