Renato Golin
2015-Mar-15 15:48 UTC
[LLVMdev] [cfe-dev] Bikeshedding commit message policy - Round 3 - Fight!
On 15 March 2015 at 15:06, Hal Finkel <hfinkel at anl.gov> wrote:> I used to use CSE:, but have now switched to using [CSE] because that seems to be the prevailing convention (and is somewhat more visually distinctive). I think it makes sense to codify that convention, but not to require them. Sometimes, there is nothing appropriate to use. Sometimes, the first or second word of the commit message is naturally the same as what the title tag would be, and so including the title tag seems redundant.I agree with you that [CSE] is the most visually striking, but I wonder how much do we want to code when to use them. There will always be arguments to all sides, and I think this is not a topic important enough for us to make it official. See how the exclamation mark became self important on the attribution and you'll see what I mean. I think common sense will always prevail if we don't try to push too many standards. cheers, --renato
Hal Finkel
2015-Mar-15 16:31 UTC
[LLVMdev] [cfe-dev] Bikeshedding commit message policy - Round 3 - Fight!
----- Original Message -----> From: "Renato Golin" <renato.golin at linaro.org> > To: "Hal Finkel" <hfinkel at anl.gov> > Cc: "LLVM Dev" <llvmdev at cs.uiuc.edu>, "Clang Dev" <cfe-dev at cs.uiuc.edu> > Sent: Sunday, March 15, 2015 10:48:37 AM > Subject: Re: [cfe-dev] Bikeshedding commit message policy - Round 3 - Fight! > > On 15 March 2015 at 15:06, Hal Finkel <hfinkel at anl.gov> wrote: > > I used to use CSE:, but have now switched to using [CSE] because > > that seems to be the prevailing convention (and is somewhat more > > visually distinctive). I think it makes sense to codify that > > convention, but not to require them. Sometimes, there is nothing > > appropriate to use. Sometimes, the first or second word of the > > commit message is naturally the same as what the title tag would > > be, and so including the title tag seems redundant. > > I agree with you that [CSE] is the most visually striking, but I > wonder how much do we want to code when to use them.I don't want to code when to use them. But it makes sense to say, "If you want to include a title tag, do it like this...".> There will > always > be arguments to all sides, and I think this is not a topic important > enough for us to make it official. > > See how the exclamation mark became self important on the attribution > and you'll see what I mean. I think common sense will always prevail > if we don't try to push too many standards.I think this is different for a number of reasons. -Hal> > cheers, > --renato >-- Hal Finkel Assistant Computational Scientist Leadership Computing Facility Argonne National Laboratory
Renato Golin
2015-Mar-15 18:49 UTC
[LLVMdev] [cfe-dev] Bikeshedding commit message policy - Round 3 - Fight!
On 15 March 2015 at 16:31, Hal Finkel <hfinkel at anl.gov> wrote:> I don't want to code when to use them. But it makes sense to say, "If you want to include a title tag, do it like this...".I'm ok with that. So, do we have consensus? 1. Don't require, but recommend using [] for tags. 2. Don't specify attribution more than just "patch by Foo." and expect people to be sensible. Current scripts already work with a good variation anyway. cheers, --renato