| Message ID | 20221013175727.17139-1-palmer@rivosinc.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | shazam: Add the --merge-base argument | expand |
Hey Konstantin, On Thu, Oct 13, 2022 at 10:57:27AM -0700, Palmer Dabbelt wrote: > I was just handling a patch set where the author used English to > describe the dependencies. They hadn't yet been merged at the time the > patch set was posted so I don't think there's really any way to make > sure the computers always understand the base, this just lets me quickly > override the automatic merge base detection when I run into something > non-canonical. > > Link: https://lore.kernel.org/all/20220913061817.22564-1-zong.li@sifive.com/ I guess this just got lost somewhere (missing a CC?), or perhaps we've missed some reason why this is not needed? Doesn't apply cleanly against 0.12.0, but with -3 it does & appears to work - at least in my simple test cases For example, the following gets automagically assigned -rc4 as a base: /stuff/b4/b4.sh shazam -s -t shazam 20230119094447.21939-3-walker.chen@starfivetech.com -H But I don't want to merge -rc4 into my for-next branch, and being able to explicitly pass my preferred base of -rc1 is really useful. Tested-by: Conor Dooley <conor.dooley@microchip.com> Thanks, Conor. > Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com> > --- > b4/command.py | 2 ++ > b4/mbox.py | 13 ++++++++----- > man/b4.5.rst | 2 ++ > 3 files changed, 12 insertions(+), 5 deletions(-) > > diff --git a/b4/command.py b/b4/command.py > index b16043e..5bb674e 100644 > --- a/b4/command.py > +++ b/b4/command.py > @@ -171,6 +171,8 @@ def cmd(): > sp_sh.add_argument('--guess-lookback', dest='guessdays', type=int, default=21, > help=('(use with -H or -M) When guessing base, go back this many days from the patch date ' > '(default: 3 weeks)')) > + sp_sh.add_argument('--merge-base', dest='mergebase', type=str, default=None, > + help=('(use with -H or -M) Force this base when merging')) > sp_sh.set_defaults(func=cmd_shazam) > > # b4 pr > diff --git a/b4/mbox.py b/b4/mbox.py > index ff96a11..c1a0660 100644 > --- a/b4/mbox.py > +++ b/b4/mbox.py > @@ -231,14 +231,17 @@ def make_am(msgs, cmdargs, msgid): > logger.critical(' Link: %s', linkurl) > > base_commit = None > - matches = re.search(r'base-commit: .*?([\da-f]+)', first_body, re.MULTILINE) > - if matches: > - base_commit = matches.groups()[0] > + if cmdargs.mergebase is not None: > + base_commit = cmdargs.mergebase > else: > - # Try a more relaxed search > - matches = re.search(r'based on .*?([\da-f]{40})', first_body, re.MULTILINE) > + matches = re.search(r'base-commit: .*?([\da-f]+)', first_body, re.MULTILINE) > if matches: > base_commit = matches.groups()[0] > + else: > + # Try a more relaxed search > + matches = re.search(r'based on .*?([\da-f]{40})', first_body, re.MULTILINE) > + if matches: > + base_commit = matches.groups()[0] > > if not base_commit and topdir and cmdargs.guessbase: > logger.critical(' Base: attempting to guess base-commit...') > diff --git a/man/b4.5.rst b/man/b4.5.rst > index 073f32a..e889542 100644 > --- a/man/b4.5.rst > +++ b/man/b4.5.rst > @@ -226,6 +226,8 @@ options: > Attempt to merge series as if it were a pull request (execs git-merge) > --guess-lookback GUESSDAYS > (use with -H or -M) When guessing base, go back this many days from the patch date (default: 3 weeks) > + --merge-base COMMIT > + (use with -H or -M) Force this base when merging > > *Example*: b4 shazam -H 20200313231252.64999-1-keescook@chromium.org > > -- > 2.38.0 >
On Fri, Jan 20, 2023 at 10:34:01PM +0000, Conor Dooley wrote: > Hey Konstantin, > > On Thu, Oct 13, 2022 at 10:57:27AM -0700, Palmer Dabbelt wrote: > > I was just handling a patch set where the author used English to > > describe the dependencies. They hadn't yet been merged at the time the > > patch set was posted so I don't think there's really any way to make > > sure the computers always understand the base, this just lets me quickly > > override the automatic merge base detection when I run into something > > non-canonical. > > > > Link: https://lore.kernel.org/all/20220913061817.22564-1-zong.li@sifive.com/ > > I guess this just got lost somewhere (missing a CC?), or perhaps we've > missed some reason why this is not needed? > > Doesn't apply cleanly against 0.12.0, but with -3 it does & appears to > work - at least in my simple test cases > For example, the following gets automagically assigned -rc4 as a base: > > /stuff/b4/b4.sh shazam -s -t shazam 20230119094447.21939-3-walker.chen@starfivetech.com -H > > But I don't want to merge -rc4 into my for-next branch, and being able > to explicitly pass my preferred base of -rc1 is really useful. > > Tested-by: Conor Dooley <conor.dooley@microchip.com> Heh, the functionality itself is tested - but I get an error doing: b4 am -3 20230116074259.22874-4-walker.chen@starfivetech.com Traceback (most recent call last): File "/stuff/b4/b4/command.py", line 379, in <module> cmd() File "/stuff/b4/b4/command.py", line 362, in cmd cmdargs.func(cmdargs) File "/stuff/b4/b4/command.py", line 91, in cmd_am b4.mbox.main(cmdargs) File "/stuff/b4/b4/mbox.py", line 710, in main make_am(msgs, cmdargs, msgid) File "/stuff/b4/b4/mbox.py", line 214, in make_am if cmdargs.mergebase is not None: AttributeError: 'Namespace' object has no attribute 'mergebase' I may go fix that up, since the feature itself is useful to me. > > Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com> > > --- > > b4/command.py | 2 ++ > > b4/mbox.py | 13 ++++++++----- > > man/b4.5.rst | 2 ++ > > 3 files changed, 12 insertions(+), 5 deletions(-) > > > > diff --git a/b4/command.py b/b4/command.py > > index b16043e..5bb674e 100644 > > --- a/b4/command.py > > +++ b/b4/command.py > > @@ -171,6 +171,8 @@ def cmd(): > > sp_sh.add_argument('--guess-lookback', dest='guessdays', type=int, default=21, > > help=('(use with -H or -M) When guessing base, go back this many days from the patch date ' > > '(default: 3 weeks)')) > > + sp_sh.add_argument('--merge-base', dest='mergebase', type=str, default=None, > > + help=('(use with -H or -M) Force this base when merging')) > > sp_sh.set_defaults(func=cmd_shazam) > > > > # b4 pr > > diff --git a/b4/mbox.py b/b4/mbox.py > > index ff96a11..c1a0660 100644 > > --- a/b4/mbox.py > > +++ b/b4/mbox.py > > @@ -231,14 +231,17 @@ def make_am(msgs, cmdargs, msgid): > > logger.critical(' Link: %s', linkurl) > > > > base_commit = None > > - matches = re.search(r'base-commit: .*?([\da-f]+)', first_body, re.MULTILINE) > > - if matches: > > - base_commit = matches.groups()[0] > > + if cmdargs.mergebase is not None: > > + base_commit = cmdargs.mergebase > > else: > > - # Try a more relaxed search > > - matches = re.search(r'based on .*?([\da-f]{40})', first_body, re.MULTILINE) > > + matches = re.search(r'base-commit: .*?([\da-f]+)', first_body, re.MULTILINE) > > if matches: > > base_commit = matches.groups()[0] > > + else: > > + # Try a more relaxed search > > + matches = re.search(r'based on .*?([\da-f]{40})', first_body, re.MULTILINE) > > + if matches: > > + base_commit = matches.groups()[0] > > > > if not base_commit and topdir and cmdargs.guessbase: > > logger.critical(' Base: attempting to guess base-commit...') > > diff --git a/man/b4.5.rst b/man/b4.5.rst > > index 073f32a..e889542 100644 > > --- a/man/b4.5.rst > > +++ b/man/b4.5.rst > > @@ -226,6 +226,8 @@ options: > > Attempt to merge series as if it were a pull request (execs git-merge) > > --guess-lookback GUESSDAYS > > (use with -H or -M) When guessing base, go back this many days from the patch date (default: 3 weeks) > > + --merge-base COMMIT > > + (use with -H or -M) Force this base when merging > > > > *Example*: b4 shazam -H 20200313231252.64999-1-keescook@chromium.org > > > > -- > > 2.38.0 > >
On Sat, Jan 21, 2023 at 05:19:02PM +0000, Conor Dooley wrote: > On Fri, Jan 20, 2023 at 10:34:01PM +0000, Conor Dooley wrote: > > Hey Konstantin, > > > > On Thu, Oct 13, 2022 at 10:57:27AM -0700, Palmer Dabbelt wrote: > > > I was just handling a patch set where the author used English to > > > describe the dependencies. They hadn't yet been merged at the time the > > > patch set was posted so I don't think there's really any way to make > > > sure the computers always understand the base, this just lets me quickly > > > override the automatic merge base detection when I run into something > > > non-canonical. > > > > > > Link: https://lore.kernel.org/all/20220913061817.22564-1-zong.li@sifive.com/ > > > > I guess this just got lost somewhere (missing a CC?), or perhaps we've > > missed some reason why this is not needed? > > > > Doesn't apply cleanly against 0.12.0, but with -3 it does & appears to > > work - at least in my simple test cases > > For example, the following gets automagically assigned -rc4 as a base: > > > > /stuff/b4/b4.sh shazam -s -t shazam 20230119094447.21939-3-walker.chen@starfivetech.com -H > > > > But I don't want to merge -rc4 into my for-next branch, and being able > > to explicitly pass my preferred base of -rc1 is really useful. > > > > Tested-by: Conor Dooley <conor.dooley@microchip.com> > > Heh, the functionality itself is tested - but I get an error doing: > b4 am -3 20230116074259.22874-4-walker.chen@starfivetech.com > > Traceback (most recent call last): > File "/stuff/b4/b4/command.py", line 379, in <module> > cmd() > File "/stuff/b4/b4/command.py", line 362, in cmd > cmdargs.func(cmdargs) > File "/stuff/b4/b4/command.py", line 91, in cmd_am > b4.mbox.main(cmdargs) > File "/stuff/b4/b4/mbox.py", line 710, in main > make_am(msgs, cmdargs, msgid) > File "/stuff/b4/b4/mbox.py", line 214, in make_am > if cmdargs.mergebase is not None: > AttributeError: 'Namespace' object has no attribute 'mergebase' > > I may go fix that up, since the feature itself is useful to me. Perhaps it is fixed up by simply doing what is done to "protect" shazam from am specific options. I'm not a pythonist, but I did try each of am, and shazam with & without the flag. They all seem to work with this diff applied: diff --git a/b4/mbox.py b/b4/mbox.py index ce04062..0b0fc40 100644 --- a/b4/mbox.py +++ b/b4/mbox.py @@ -211,7 +211,7 @@ def make_am(msgs: List[email.message.Message], cmdargs: argparse.Namespace, msgi logger.critical(' Link: %s', linkurl) base_commit = None - if cmdargs.mergebase is not None: + if cmdargs.mergebase: base_commit = cmdargs.mergebase else: matches = re.search(r'base-commit: .*?([\da-f]+)', first_body, re.MULTILINE) @@ -673,8 +673,8 @@ def refetch(dest: str) -> None: def main(cmdargs: argparse.Namespace) -> None: + # We force some settings if cmdargs.subcmd == 'shazam': - # We force some settings cmdargs.checknewer = True cmdargs.threeway = False cmdargs.nopartialreroll = False @@ -686,6 +686,8 @@ def main(cmdargs: argparse.Namespace) -> None: cmdargs.guessbase = True else: cmdargs.guessbase = False + else: + cmdargs.mergebase = False if cmdargs.checknewer: # Force nocache mode
diff --git a/b4/command.py b/b4/command.py index b16043e..5bb674e 100644 --- a/b4/command.py +++ b/b4/command.py @@ -171,6 +171,8 @@ def cmd(): sp_sh.add_argument('--guess-lookback', dest='guessdays', type=int, default=21, help=('(use with -H or -M) When guessing base, go back this many days from the patch date ' '(default: 3 weeks)')) + sp_sh.add_argument('--merge-base', dest='mergebase', type=str, default=None, + help=('(use with -H or -M) Force this base when merging')) sp_sh.set_defaults(func=cmd_shazam) # b4 pr diff --git a/b4/mbox.py b/b4/mbox.py index ff96a11..c1a0660 100644 --- a/b4/mbox.py +++ b/b4/mbox.py @@ -231,14 +231,17 @@ def make_am(msgs, cmdargs, msgid): logger.critical(' Link: %s', linkurl) base_commit = None - matches = re.search(r'base-commit: .*?([\da-f]+)', first_body, re.MULTILINE) - if matches: - base_commit = matches.groups()[0] + if cmdargs.mergebase is not None: + base_commit = cmdargs.mergebase else: - # Try a more relaxed search - matches = re.search(r'based on .*?([\da-f]{40})', first_body, re.MULTILINE) + matches = re.search(r'base-commit: .*?([\da-f]+)', first_body, re.MULTILINE) if matches: base_commit = matches.groups()[0] + else: + # Try a more relaxed search + matches = re.search(r'based on .*?([\da-f]{40})', first_body, re.MULTILINE) + if matches: + base_commit = matches.groups()[0] if not base_commit and topdir and cmdargs.guessbase: logger.critical(' Base: attempting to guess base-commit...') diff --git a/man/b4.5.rst b/man/b4.5.rst index 073f32a..e889542 100644 --- a/man/b4.5.rst +++ b/man/b4.5.rst @@ -226,6 +226,8 @@ options: Attempt to merge series as if it were a pull request (execs git-merge) --guess-lookback GUESSDAYS (use with -H or -M) When guessing base, go back this many days from the patch date (default: 3 weeks) + --merge-base COMMIT + (use with -H or -M) Force this base when merging *Example*: b4 shazam -H 20200313231252.64999-1-keescook@chromium.org
I was just handling a patch set where the author used English to describe the dependencies. They hadn't yet been merged at the time the patch set was posted so I don't think there's really any way to make sure the computers always understand the base, this just lets me quickly override the automatic merge base detection when I run into something non-canonical. Link: https://lore.kernel.org/all/20220913061817.22564-1-zong.li@sifive.com/ Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com> --- b4/command.py | 2 ++ b4/mbox.py | 13 ++++++++----- man/b4.5.rst | 2 ++ 3 files changed, 12 insertions(+), 5 deletions(-)