I agree, moving these components out of the main repo should result in a major version change. The idea of dividing the whole process in two stages (v3 and v4) seems reasonable too.
On Monday, July 27, 2020 at 5:15:00 AM UTC+3 bl...@furius.ca wrote: > What about this: In version 2.4 (we're now at 2.3.1), a minor version, I > moved the reporting code, the web app, the price fetching script, and > possibly the ingest library out of the main repo. Is that reasonable? I > don't think so. > > What could make sense, is to make that a major revision (v3) - just v2 > with the tools removed - and the new work in C++ would eventuawlly become > v4. I feel like waiting until the new version has at least some meat before > bothering people currently using v2. > > > > > > On Sun, Jul 26, 2020 at 8:16 AM Kirill Goncharov <kdgon...@gmail.com> > wrote: > >> >I assume many users here run directly from source code and it's probably >> a minority who pip install. >> >> I think that the opposite is true. >> According to this site, beancount package has ~50 downloads on average >> per day: https://pypistats.org/packages/beancount , many of these >> downloads are probably coming from Fava installations. >> I also use only packaged stable version. >> >> On Sunday, July 26, 2020 at 5:24:39 AM UTC+3 bl...@furius.ca wrote: >> >>> On Sat, Jul 25, 2020 at 4:18 PM Daniele Nicolodi <dan...@grinta.net> >>> wrote: >>> >>>> On 25/07/2020 13:14, Martin Blais wrote: >>>> > On Sat, Jul 25, 2020 at 2:23 PM Daniele Nicolodi <dan...@grinta.net >>>> > Would it make sense to have v2 depend on the new project instead >>>> that >>>> > having the code in two places? beancount v2 setup.py could >>>> restrict the >>>> > version to (for example) 'beanprice < 2.0' so that incompatible >>>> changes >>>> > can be pushed to another branch. >>>> > >>>> > This only requires publishing beanprice to PyPI and declaring the >>>> > dependency in setup.py (and merge the patch that removes >>>> beanprice from >>>> > master to the v2 branch). I can do that if there is agreement >>>> about this >>>> > way forward. >>>> > >>>> > >>>> > I thought about removing it from v2, and doing the same for >>>> report/web, >>>> > but I kind-of feel that until v3 has at least a "beta" level base to >>>> > work from, it's probably best not to mess too much with v2, current >>>> > users should be able to just keep working without having to make >>>> changes. >>>> >>>> Unless I am missing something, there would not be any user visible >>>> change other than one more package downloaded and installed when 'pip >>>> install beancount' is run. >>>> >>> >>> That's assuming people aren't running from source. >>> I assume many users here run directly from source code and it's probably >>> a minority who pip install. >>> I don't really know TBH. >>> >>> I think of v2 as "frozen" at this point, more or less, new development >>> will occur on v3. As long as the changes are trivially mergeable to v3 and >>> introduce only very minimal changes (e.g. bug fixes), I probably will do >>> them on the v2 branch, but anything interesting will happen on v3. There's >>> no point in breaking things on v2. Let's break things in v3 (and make >>> progress). >>> >>> >>> -- >> You received this message because you are subscribed to the Google Groups >> "Beancount" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to beancount+...@googlegroups.com. >> > To view this discussion on the web visit >> https://groups.google.com/d/msgid/beancount/7e4f68fc-af1d-4cd7-9a45-2e29bbafea71n%40googlegroups.com >> >> <https://groups.google.com/d/msgid/beancount/7e4f68fc-af1d-4cd7-9a45-2e29bbafea71n%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> > -- You received this message because you are subscribed to the Google Groups "Beancount" group. To unsubscribe from this group and stop receiving emails from it, send an email to beancount+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/0d3401f9-296f-459c-8051-9bd4ee563d2cn%40googlegroups.com.
