Re: [Scons-dev] SCons and Python 3
Dirk, I'll take a look at windows. It does seem a bit odd.. -Bill On Sat, Dec 19, 2015 at 4:12 AM, Dirk Bächle wrote: > Russel, > > On 19.12.2015 10:39, Russel Winder wrote: > >> On Wed, 2015-12-16 at 09:22 +0100, Dirk Bächle wrote: >> >>> >>> […] >> >>> @Russel: If all Buildbots turn out to be "green", feel free to merge >>> onto the python3 branch... >>> >> >> It appears as though Windows is still a bit red, but I cannot in all >> honest profess any sadness. >> > > feel free to latch on. The Windows bot seems to have had a "hiccup"...it > didn't find any updates on its last "hg pull", but the last proper run was > green. @Bill: Can you check the Windows setup? It looks to me as if it has > a problem with merges over more than one commit. The slave seems to think > that it has to process several commits, but pulls all of them in the first > run anyway? That's how it looks to me at least... > > As all the Linux variants are green, I > >> shall take this as a "go" to do a merge to the Python 3 branch. This >> will undoubtedly lead to a right royal mess, but then that is the whole >> point – fix the mess. >> >> My assumptions will be: >> >> Python 3.4 and 3.5 >> Python 2.7.10 >> Debian Sid >> Fedora Rawhide >> No packages not in the standard distribution >> A lack of CI to cover Solaris, OSX or Windows >> >> > Sounds good. > > I may be gone a while… >> >> Progress will be reflected in https://bitbucket.org/russel/scons__pyth >> on3 >> >> > I'll definitely track that one. If you need help, just let us know...I'm > sure that all the other devs (me included) are ready to jump in and help > where they can. > > Meanwhile, we others should try to come up with a plan of how to test > Python 2.7 and 3.x in parallel in the future. Some people have nominated > "tox" for this task already, but is this still the way to go...or can > anybody recommend a better tool? > > Best regards, > > > Dirk > > ___ > Scons-dev mailing list > Scons-dev@scons.org > https://pairlist2.pair.net/mailman/listinfo/scons-dev > ___ Scons-dev mailing list Scons-dev@scons.org https://pairlist2.pair.net/mailman/listinfo/scons-dev
Re: [Scons-dev] Huge merges in Mercurial
With GIT you can merge at a commit level (IE. you don't have to merge a whole branch at once). Can you do this with Mercurial? That way you can do it in steps rather than all at once. V/R, William On Sat, Dec 19, 2015 at 1:51 PM, Dirk Bächle wrote: > Russel, > > On 19.12.2015 14:31, Russel Winder wrote: > >> Hi, >> >> I tried an experimental merge and much of the stuff looks like it will >> go straight through. However there is a lot of 3-way meld resolution to >> be done, hours of it, possibly days. Clearly there is no problem the >> human taking breaks as long as the computer doesn't. But what if… >> >> Has anyone any experience of checkpointing and restarting big Mercurial >> merge resolutions? >> >> > not with Mercurial so far, but here's what I tend to do at work > (Clearcase) when things get complicated: > > For every conflicting file I'll accept the "new" changes completely, > overwriting old stuff. This is just to get the basic "merge" operation > through. > Afterwards, I look at the single "diffs" and do the actual merge work by > trying to amend those places where some of the "old" stuff is still needed. > Note how you can list the "to be resolved" files with > > hg resolve -la > > Hope this helps you further, you may want to combine the "amend" work with > using the MQ extension... > > Regards, > > Dirk > > ___ > Scons-dev mailing list > Scons-dev@scons.org > https://pairlist2.pair.net/mailman/listinfo/scons-dev > ___ Scons-dev mailing list Scons-dev@scons.org https://pairlist2.pair.net/mailman/listinfo/scons-dev
Re: [Scons-dev] SCons and Python 3
Hi Russel; last time I did this it wasn't all _that_ painful. Took a few hours if I remember rightly. (Of course a lot of time has passed, but still not that much new code compared to the entire code base.) Look at the diff of the last merge before you start, to get a sense of the top two or three things that have to get tweaked. On Sat, Dec 19, 2015 at 4:39 AM, Russel Winder wrote: > On Wed, 2015-12-16 at 09:22 +0100, Dirk Bächle wrote: > > > […] > > @Russel: If all Buildbots turn out to be "green", feel free to merge > > onto the python3 branch... > > It appears as though Windows is still a bit red, but I cannot in all > honest profess any sadness. As all the Linux variants are green, I > shall take this as a "go" to do a merge to the Python 3 branch. This > will undoubtedly lead to a right royal mess, but then that is the whole > point – fix the mess. > > My assumptions will be: > > Python 3.4 and 3.5 > Python 2.7.10 > Debian Sid > Fedora Rawhide > No packages not in the standard distribution > A lack of CI to cover Solaris, OSX or Windows > > I may be gone a while… > > Progress will be reflected in https://bitbucket.org/russel/scons__pyth > on3 > > Should anyone track progress and want to chip in feel free. If you see > any lack of progress, it will either be me having to work on organizing > ACCU 2016 (accu.org) or choosing to work for a while on GPars or Me TV > – or eating and especially drinking, it is Christmas after all. Do not > drink and code. > > -- > Russel. > > = > Dr Russel Winder t: +44 20 7585 2200 voip: > sip:russel.win...@ekiga.net > 41 Buckmaster Roadm: +44 7770 465 077 xmpp: rus...@winder.org.uk > London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder > > > ___ > Scons-dev mailing list > Scons-dev@scons.org > https://pairlist2.pair.net/mailman/listinfo/scons-dev > > -- Gary ___ Scons-dev mailing list Scons-dev@scons.org https://pairlist2.pair.net/mailman/listinfo/scons-dev
Re: [Scons-dev] Huge merges in Mercurial
Russel, On 19.12.2015 14:31, Russel Winder wrote: Hi, I tried an experimental merge and much of the stuff looks like it will go straight through. However there is a lot of 3-way meld resolution to be done, hours of it, possibly days. Clearly there is no problem the human taking breaks as long as the computer doesn't. But what if… Has anyone any experience of checkpointing and restarting big Mercurial merge resolutions? not with Mercurial so far, but here's what I tend to do at work (Clearcase) when things get complicated: For every conflicting file I'll accept the "new" changes completely, overwriting old stuff. This is just to get the basic "merge" operation through. Afterwards, I look at the single "diffs" and do the actual merge work by trying to amend those places where some of the "old" stuff is still needed. Note how you can list the "to be resolved" files with hg resolve -la Hope this helps you further, you may want to combine the "amend" work with using the MQ extension... Regards, Dirk ___ Scons-dev mailing list Scons-dev@scons.org https://pairlist2.pair.net/mailman/listinfo/scons-dev
[Scons-dev] Huge merges in Mercurial
Hi, I tried an experimental merge and much of the stuff looks like it will go straight through. However there is a lot of 3-way meld resolution to be done, hours of it, possibly days. Clearly there is no problem the human taking breaks as long as the computer doesn't. But what if… Has anyone any experience of checkpointing and restarting big Mercurial merge resolutions? -- Russel. = Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.win...@ekiga.net 41 Buckmaster Roadm: +44 7770 465 077 xmpp: rus...@winder.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder signature.asc Description: This is a digitally signed message part ___ Scons-dev mailing list Scons-dev@scons.org https://pairlist2.pair.net/mailman/listinfo/scons-dev
Re: [Scons-dev] SCons and Python 3
Russel, On 19.12.2015 10:39, Russel Winder wrote: On Wed, 2015-12-16 at 09:22 +0100, Dirk Bächle wrote: […] @Russel: If all Buildbots turn out to be "green", feel free to merge onto the python3 branch... It appears as though Windows is still a bit red, but I cannot in all honest profess any sadness. feel free to latch on. The Windows bot seems to have had a "hiccup"...it didn't find any updates on its last "hg pull", but the last proper run was green. @Bill: Can you check the Windows setup? It looks to me as if it has a problem with merges over more than one commit. The slave seems to think that it has to process several commits, but pulls all of them in the first run anyway? That's how it looks to me at least... As all the Linux variants are green, I shall take this as a "go" to do a merge to the Python 3 branch. This will undoubtedly lead to a right royal mess, but then that is the whole point – fix the mess. My assumptions will be: Python 3.4 and 3.5 Python 2.7.10 Debian Sid Fedora Rawhide No packages not in the standard distribution A lack of CI to cover Solaris, OSX or Windows Sounds good. I may be gone a while… Progress will be reflected in https://bitbucket.org/russel/scons__pyth on3 I'll definitely track that one. If you need help, just let us know...I'm sure that all the other devs (me included) are ready to jump in and help where they can. Meanwhile, we others should try to come up with a plan of how to test Python 2.7 and 3.x in parallel in the future. Some people have nominated "tox" for this task already, but is this still the way to go...or can anybody recommend a better tool? Best regards, Dirk ___ Scons-dev mailing list Scons-dev@scons.org https://pairlist2.pair.net/mailman/listinfo/scons-dev
Re: [Scons-dev] SCons and Python 3
On Wed, 2015-12-16 at 09:22 +0100, Dirk Bächle wrote: > […] > @Russel: If all Buildbots turn out to be "green", feel free to merge > onto the python3 branch... It appears as though Windows is still a bit red, but I cannot in all honest profess any sadness. As all the Linux variants are green, I shall take this as a "go" to do a merge to the Python 3 branch. This will undoubtedly lead to a right royal mess, but then that is the whole point – fix the mess. My assumptions will be: Python 3.4 and 3.5 Python 2.7.10 Debian Sid Fedora Rawhide No packages not in the standard distribution A lack of CI to cover Solaris, OSX or Windows I may be gone a while… Progress will be reflected in https://bitbucket.org/russel/scons__pyth on3 Should anyone track progress and want to chip in feel free. If you see any lack of progress, it will either be me having to work on organizing ACCU 2016 (accu.org) or choosing to work for a while on GPars or Me TV – or eating and especially drinking, it is Christmas after all. Do not drink and code. -- Russel. = Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.win...@ekiga.net 41 Buckmaster Roadm: +44 7770 465 077 xmpp: rus...@winder.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder signature.asc Description: This is a digitally signed message part ___ Scons-dev mailing list Scons-dev@scons.org https://pairlist2.pair.net/mailman/listinfo/scons-dev