Обсуждение: Git configuration
Hi, Magnus added a check of the commiter so that only one email address is usable for a commit. It works great. But I have one issue. The only address I can use is my work one (dalibo.com). I would prefer to use my personal one because that's the one already on the other commits. But if it's not possible, I don't care. But if we stay with my work email address, all my commits don't get through the mailing list. I have to moderate them. So, there is at least one thing to do: * accept my personal email address for commits or * add my work email address in a whitelist for the mailing list I prefer the first solution, but I won't object if we choose the second one. -- Guillaume http://www.postgresql.fr http://dalibo.com
Le 24/07/2010 19:41, Guillaume Lelarge a écrit : > Hi, > > Magnus added a check of the commiter so that only one email address is > usable for a commit. It works great. But I have one issue. The only > address I can use is my work one (dalibo.com). I would prefer to use my > personal one because that's the one already on the other commits. But if > it's not possible, I don't care. But if we stay with my work email > address, all my commits don't get through the mailing list. I have to > moderate them. > > So, there is at least one thing to do: > > * accept my personal email address for commits > > or > > * add my work email address in a whitelist for the mailing list > > I prefer the first solution, but I won't object if we choose the second one. > BTW, it forces us to stop using --author command line option. I would prefer if we could keep using it. I know the PostgreSQL folks don't want it already, but should we follow their view on this? -- Guillaume http://www.postgresql.fr http://dalibo.com
On Sun, Jul 25, 2010 at 8:40 AM, Guillaume Lelarge <guillaume@lelarge.info> wrote: > Le 24/07/2010 19:41, Guillaume Lelarge a écrit : >> Hi, >> >> Magnus added a check of the commiter so that only one email address is >> usable for a commit. It works great. But I have one issue. The only >> address I can use is my work one (dalibo.com). I would prefer to use my >> personal one because that's the one already on the other commits. But if >> it's not possible, I don't care. But if we stay with my work email >> address, all my commits don't get through the mailing list. I have to >> moderate them. >> >> So, there is at least one thing to do: >> >> * accept my personal email address for commits >> >> or >> >> * add my work email address in a whitelist for the mailing list >> >> I prefer the first solution, but I won't object if we choose the second one. Don't blame Magnus - it was me. Per our IM conversation I've changed the address. > BTW, it forces us to stop using --author command line option. I would > prefer if we could keep using it. I know the PostgreSQL folks don't want > it already, but should we follow their view on this? No, that doesn't. We have 2 things setup on the repo: Policy enforcement: This ensures that you always commit using the correct email address Commit messages: This sens the commit message to the list. The script is currently using the Author address as the sender, which I'm going to look at fixing to use the Committer address. The way it was setup, was that your policy approved address wasn't subscribed to the mailing list. -- Dave Page EnterpriseDB UK: http://www.enterprisedb.com The Enterprise Postgres Company
Le 25/07/2010 11:50, Dave Page a écrit : > On Sun, Jul 25, 2010 at 8:40 AM, Guillaume Lelarge > <guillaume@lelarge.info> wrote: >> Le 24/07/2010 19:41, Guillaume Lelarge a écrit : >>> Hi, >>> >>> Magnus added a check of the commiter so that only one email address is >>> usable for a commit. It works great. But I have one issue. The only >>> address I can use is my work one (dalibo.com). I would prefer to use my >>> personal one because that's the one already on the other commits. But if >>> it's not possible, I don't care. But if we stay with my work email >>> address, all my commits don't get through the mailing list. I have to >>> moderate them. >>> >>> So, there is at least one thing to do: >>> >>> * accept my personal email address for commits >>> >>> or >>> >>> * add my work email address in a whitelist for the mailing list >>> >>> I prefer the first solution, but I won't object if we choose the second one. > > Don't blame Magnus - it was me. Per our IM conversation I've changed > the address. > >> BTW, it forces us to stop using --author command line option. I would >> prefer if we could keep using it. I know the PostgreSQL folks don't want >> it already, but should we follow their view on this? > > No, that doesn't. We have 2 things setup on the repo: > > Policy enforcement: This ensures that you always commit using the > correct email address > Commit messages: This sens the commit message to the list. The script > is currently using the Author address as the sender, which I'm going > to look at fixing to use the Committer address. > Did you fix the script? do you need help? I can probably take some time to look at it. -- Guillaume http://www.postgresql.fr http://dalibo.com
On Tue, Jul 27, 2010 at 10:30 AM, Guillaume Lelarge <guillaume@lelarge.info> wrote: > Did you fix the script? do you need help? I can probably take some time > to look at it. Not yet. Feel free. It's in commitmsg.py iirc: http://github.com/mhagander/pg_githooks -- Dave Page EnterpriseDB UK: http://www.enterprisedb.com The Enterprise Postgres Company
Le 27/07/2010 11:37, Dave Page a écrit : > On Tue, Jul 27, 2010 at 10:30 AM, Guillaume Lelarge > <guillaume@lelarge.info> wrote: > >> Did you fix the script? do you need help? I can probably take some time >> to look at it. > > Not yet. Feel free. It's in commitmsg.py iirc: > > http://github.com/mhagander/pg_githooks > Well, I could be completely wrong but I think you just need to change one parameter in hook/policyenforce.ini: committerequalsauthor must be equal to 0. If you do this, the hook won't try to match commiter's and author's name, but it will still enforce commiters. Which is what we want, right? -- Guillaume http://www.postgresql.fr http://dalibo.com
On Tue, Jul 27, 2010 at 11:16 AM, Guillaume Lelarge <guillaume@lelarge.info> wrote: > Le 27/07/2010 11:37, Dave Page a écrit : >> On Tue, Jul 27, 2010 at 10:30 AM, Guillaume Lelarge >> <guillaume@lelarge.info> wrote: >> >>> Did you fix the script? do you need help? I can probably take some time >>> to look at it. >> >> Not yet. Feel free. It's in commitmsg.py iirc: >> >> http://github.com/mhagander/pg_githooks >> > > Well, I could be completely wrong but I think you just need to change > one parameter in hook/policyenforce.ini: committerequalsauthor must be > equal to 0. If you do this, the hook won't try to match commiter's and > author's name, but it will still enforce commiters. Which is what we > want, right? That isn't the problem (and committerequalsauthor is already zero). The problem is that the commitmsg script sends the email "from" the author, not the committer. -- Dave Page EnterpriseDB UK: http://www.enterprisedb.com The Enterprise Postgres Company
Le 27/07/2010 13:08, Dave Page a écrit : > On Tue, Jul 27, 2010 at 11:16 AM, Guillaume Lelarge > <guillaume@lelarge.info> wrote: >> Le 27/07/2010 11:37, Dave Page a écrit : >>> On Tue, Jul 27, 2010 at 10:30 AM, Guillaume Lelarge >>> <guillaume@lelarge.info> wrote: >>> >>>> Did you fix the script? do you need help? I can probably take some time >>>> to look at it. >>> >>> Not yet. Feel free. It's in commitmsg.py iirc: >>> >>> http://github.com/mhagander/pg_githooks >>> >> >> Well, I could be completely wrong but I think you just need to change >> one parameter in hook/policyenforce.ini: committerequalsauthor must be >> equal to 0. If you do this, the hook won't try to match commiter's and >> author's name, but it will still enforce commiters. Which is what we >> want, right? > > That isn't the problem (and committerequalsauthor is already zero). > The problem is that the commitmsg script sends the email "from" the > author, not the committer. > Oh OK. I thought I still couldn't use the --author CLI option. Now, I understand the issue. I'll try to work on this today. -- Guillaume http://www.postgresql.fr http://dalibo.com
On Tue, Jul 27, 2010 at 1:10 PM, Guillaume Lelarge <guillaume@lelarge.info> wrote: > Le 27/07/2010 13:08, Dave Page a écrit : >> On Tue, Jul 27, 2010 at 11:16 AM, Guillaume Lelarge >> <guillaume@lelarge.info> wrote: >>> Le 27/07/2010 11:37, Dave Page a écrit : >>>> On Tue, Jul 27, 2010 at 10:30 AM, Guillaume Lelarge >>>> <guillaume@lelarge.info> wrote: >>>> >>>>> Did you fix the script? do you need help? I can probably take some time >>>>> to look at it. >>>> >>>> Not yet. Feel free. It's in commitmsg.py iirc: >>>> >>>> http://github.com/mhagander/pg_githooks >>>> >>> >>> Well, I could be completely wrong but I think you just need to change >>> one parameter in hook/policyenforce.ini: committerequalsauthor must be >>> equal to 0. If you do this, the hook won't try to match commiter's and >>> author's name, but it will still enforce commiters. Which is what we >>> want, right? >> >> That isn't the problem (and committerequalsauthor is already zero). >> The problem is that the commitmsg script sends the email "from" the >> author, not the committer. >> > > Oh OK. I thought I still couldn't use the --author CLI option. Now, I > understand the issue. I'll try to work on this today. Well you shouldn't, as it will send email in someone else's name until this is fixed. -- Dave Page EnterpriseDB UK: http://www.enterprisedb.com The Enterprise Postgres Company
Le 27/07/2010 14:12, Dave Page a écrit : > On Tue, Jul 27, 2010 at 1:10 PM, Guillaume Lelarge > <guillaume@lelarge.info> wrote: >> Le 27/07/2010 13:08, Dave Page a écrit : >>> On Tue, Jul 27, 2010 at 11:16 AM, Guillaume Lelarge >>> <guillaume@lelarge.info> wrote: >>>> Le 27/07/2010 11:37, Dave Page a écrit : >>>>> On Tue, Jul 27, 2010 at 10:30 AM, Guillaume Lelarge >>>>> <guillaume@lelarge.info> wrote: >>>>> >>>>>> Did you fix the script? do you need help? I can probably take some time >>>>>> to look at it. >>>>> >>>>> Not yet. Feel free. It's in commitmsg.py iirc: >>>>> >>>>> http://github.com/mhagander/pg_githooks >>>>> >>>> >>>> Well, I could be completely wrong but I think you just need to change >>>> one parameter in hook/policyenforce.ini: committerequalsauthor must be >>>> equal to 0. If you do this, the hook won't try to match commiter's and >>>> author's name, but it will still enforce commiters. Which is what we >>>> want, right? >>> >>> That isn't the problem (and committerequalsauthor is already zero). >>> The problem is that the commitmsg script sends the email "from" the >>> author, not the committer. >>> >> >> Oh OK. I thought I still couldn't use the --author CLI option. Now, I >> understand the issue. I'll try to work on this today. > > Well you shouldn't, as it will send email in someone else's name until > this is fixed. > Here is a patch to fix this. Can you try it? Thanks. -- Guillaume http://www.postgresql.fr http://dalibo.com
On Tue, Jul 27, 2010 at 5:36 PM, Guillaume Lelarge <guillaume@lelarge.info> wrote: > Here is a patch to fix this. Can you try it? Thanks. Looks OK (except for the spelling of committerinfo). But it didn't work when I tried it (no email was produced), and I can't seem to get it to show me any debug info. So it got reverted again for now. -- Dave Page EnterpriseDB UK: http://www.enterprisedb.com The Enterprise Postgres Company
Le 28/07/2010 13:26, Dave Page a écrit : > On Tue, Jul 27, 2010 at 5:36 PM, Guillaume Lelarge > <guillaume@lelarge.info> wrote: > >> Here is a patch to fix this. Can you try it? Thanks. > > Looks OK (except for the spelling of committerinfo). But it didn't > work when I tried it (no email was produced), and I can't seem to get > it to show me any debug info. So it got reverted again for now. > What kind of error do you get? -- Guillaume http://www.postgresql.fr http://dalibo.com
Le 28/07/2010 14:50, Guillaume Lelarge a écrit : > Le 28/07/2010 13:26, Dave Page a écrit : >> On Tue, Jul 27, 2010 at 5:36 PM, Guillaume Lelarge >> <guillaume@lelarge.info> wrote: >> >>> Here is a patch to fix this. Can you try it? Thanks. >> >> Looks OK (except for the spelling of committerinfo). But it didn't >> work when I tried it (no email was produced), and I can't seem to get >> it to show me any debug info. So it got reverted again for now. >> > > What kind of error do you get? > Worked a bit on it, and here is a new patch (checked with the debug config). The issue with the previous patch was that full format doesn't include the date of the commit. So, we need to use the format "fuller". Anyway, hope this one will have better results. -- Guillaume http://www.postgresql.fr http://dalibo.com
On Wed, Jul 28, 2010 at 2:08 PM, Guillaume Lelarge <guillaume@lelarge.info> wrote: > Le 28/07/2010 14:50, Guillaume Lelarge a écrit : >> Le 28/07/2010 13:26, Dave Page a écrit : >>> On Tue, Jul 27, 2010 at 5:36 PM, Guillaume Lelarge >>> <guillaume@lelarge.info> wrote: >>> >>>> Here is a patch to fix this. Can you try it? Thanks. >>> >>> Looks OK (except for the spelling of committerinfo). But it didn't >>> work when I tried it (no email was produced), and I can't seem to get >>> it to show me any debug info. So it got reverted again for now. >>> >> >> What kind of error do you get? >> > > Worked a bit on it, and here is a new patch (checked with the debug > config). The issue with the previous patch was that full format doesn't > include the date of the commit. So, we need to use the format "fuller". > Anyway, hope this one will have better results. Nope still doesn't work - and I'm going to stop testing it on the pgAdmin repo now, via GIT. Magnus won't be pleased with me messing up his GIT history, even if the whole thing is still in test. -- Dave Page EnterpriseDB UK: http://www.enterprisedb.com The Enterprise Postgres Company
Le 28/07/2010 17:47, Dave Page a écrit : > On Wed, Jul 28, 2010 at 2:08 PM, Guillaume Lelarge > <guillaume@lelarge.info> wrote: >> Le 28/07/2010 14:50, Guillaume Lelarge a écrit : >>> Le 28/07/2010 13:26, Dave Page a écrit : >>>> On Tue, Jul 27, 2010 at 5:36 PM, Guillaume Lelarge >>>> <guillaume@lelarge.info> wrote: >>>> >>>>> Here is a patch to fix this. Can you try it? Thanks. >>>> >>>> Looks OK (except for the spelling of committerinfo). But it didn't >>>> work when I tried it (no email was produced), and I can't seem to get >>>> it to show me any debug info. So it got reverted again for now. >>>> >>> >>> What kind of error do you get? >>> >> >> Worked a bit on it, and here is a new patch (checked with the debug >> config). The issue with the previous patch was that full format doesn't >> include the date of the commit. So, we need to use the format "fuller". >> Anyway, hope this one will have better results. > > Nope still doesn't work - and I'm going to stop testing it on the > pgAdmin repo now, via GIT. Magnus won't be pleased with me messing up > his GIT history, even if the whole thing is still in test. > I don't see why it still doesn't work. Anyway, you're right. I won't spend more time on this. My problem is: I have at least one commit to do with another author. Can I? -- Guillaume http://www.postgresql.fr http://dalibo.com
On Wed, Jul 28, 2010 at 4:59 PM, Guillaume Lelarge <guillaume@lelarge.info> wrote: > I don't see why it still doesn't work. Anyway, you're right. I won't > spend more time on this. My problem is: I have at least one commit to do > with another author. Can I? Yes, but just do it the old fashioned way - with credit in the commit message. We'll just have to hold off on this new feature until Magnus can help. -- Dave Page EnterpriseDB UK: http://www.enterprisedb.com The Enterprise Postgres Company
On Wed, Jul 28, 2010 at 18:02, Dave Page <dpage@pgadmin.org> wrote: > On Wed, Jul 28, 2010 at 4:59 PM, Guillaume Lelarge > <guillaume@lelarge.info> wrote: >> I don't see why it still doesn't work. Anyway, you're right. I won't >> spend more time on this. My problem is: I have at least one commit to do >> with another author. Can I? > > Yes, but just do it the old fashioned way - with credit in the commit message. > > We'll just have to hold off on this new feature until Magnus can help. Fixed. You were *really* close to getting it right ;) I've installed the new version on git.postgresql.org, and Guillaume will likely make a proper test commit soon :-) -- Magnus Hagander Me: http://www.hagander.net/ Work: http://www.redpill-linpro.com/
Le 19/08/2010 16:00, Magnus Hagander a écrit : > On Wed, Jul 28, 2010 at 18:02, Dave Page <dpage@pgadmin.org> wrote: >> On Wed, Jul 28, 2010 at 4:59 PM, Guillaume Lelarge >> <guillaume@lelarge.info> wrote: >>> I don't see why it still doesn't work. Anyway, you're right. I won't >>> spend more time on this. My problem is: I have at least one commit to do >>> with another author. Can I? >> >> Yes, but just do it the old fashioned way - with credit in the commit message. >> >> We'll just have to hold off on this new feature until Magnus can help. > > Fixed. > > You were *really* close to getting it right ;) > > I've installed the new version on git.postgresql.org, and Guillaume > will likely make a proper test commit soon :-) > Done. But there is an issue: guillaume@laptop:~/freeprojects/git.pgadmin3(master) $ git push Counting objects: 15, done. Delta compression using up to 4 threads. Compressing objects: 100% (8/8), done. Writing objects: 100% (8/8), 761 bytes, done. Total 8 (delta 7), reused 0 (delta 0) fatal: unrecognized argument: --format=full To ssh://git@git.postgresql.org/pgadmin3.git f627445..70a4f28 master -> master See the "fatal: unrecognized argument: --format=full" line? -- Guillaume http://www.postgresql.fr http://dalibo.com
On Thu, Aug 19, 2010 at 16:17, Guillaume Lelarge <guillaume@lelarge.info> wrote: > Le 19/08/2010 16:00, Magnus Hagander a écrit : >> On Wed, Jul 28, 2010 at 18:02, Dave Page <dpage@pgadmin.org> wrote: >>> On Wed, Jul 28, 2010 at 4:59 PM, Guillaume Lelarge >>> <guillaume@lelarge.info> wrote: >>>> I don't see why it still doesn't work. Anyway, you're right. I won't >>>> spend more time on this. My problem is: I have at least one commit to do >>>> with another author. Can I? >>> >>> Yes, but just do it the old fashioned way - with credit in the commit message. >>> >>> We'll just have to hold off on this new feature until Magnus can help. >> >> Fixed. >> >> You were *really* close to getting it right ;) >> >> I've installed the new version on git.postgresql.org, and Guillaume >> will likely make a proper test commit soon :-) >> > > Done. > > But there is an issue: > > guillaume@laptop:~/freeprojects/git.pgadmin3(master) $ git push > Counting objects: 15, done. > Delta compression using up to 4 threads. > Compressing objects: 100% (8/8), done. > Writing objects: 100% (8/8), 761 bytes, done. > Total 8 (delta 7), reused 0 (delta 0) > fatal: unrecognized argument: --format=full > To ssh://git@git.postgresql.org/pgadmin3.git > f627445..70a4f28 master -> master > > > See the "fatal: unrecognized argument: --format=full" line? Meh. Old version of git. Fix committed. -- Magnus Hagander Me: http://www.hagander.net/ Work: http://www.redpill-linpro.com/
Le 19/08/2010 16:35, Magnus Hagander a écrit : > On Thu, Aug 19, 2010 at 16:17, Guillaume Lelarge <guillaume@lelarge.info> wrote: >> Le 19/08/2010 16:00, Magnus Hagander a écrit : >>> On Wed, Jul 28, 2010 at 18:02, Dave Page <dpage@pgadmin.org> wrote: >>>> On Wed, Jul 28, 2010 at 4:59 PM, Guillaume Lelarge >>>> <guillaume@lelarge.info> wrote: >>>>> I don't see why it still doesn't work. Anyway, you're right. I won't >>>>> spend more time on this. My problem is: I have at least one commit to do >>>>> with another author. Can I? >>>> >>>> Yes, but just do it the old fashioned way - with credit in the commit message. >>>> >>>> We'll just have to hold off on this new feature until Magnus can help. >>> >>> Fixed. >>> >>> You were *really* close to getting it right ;) >>> >>> I've installed the new version on git.postgresql.org, and Guillaume >>> will likely make a proper test commit soon :-) >>> >> >> Done. >> >> But there is an issue: >> >> guillaume@laptop:~/freeprojects/git.pgadmin3(master) $ git push >> Counting objects: 15, done. >> Delta compression using up to 4 threads. >> Compressing objects: 100% (8/8), done. >> Writing objects: 100% (8/8), 761 bytes, done. >> Total 8 (delta 7), reused 0 (delta 0) >> fatal: unrecognized argument: --format=full >> To ssh://git@git.postgresql.org/pgadmin3.git >> f627445..70a4f28 master -> master >> >> >> See the "fatal: unrecognized argument: --format=full" line? > > Meh. > > Old version of git. Fix committed. > Works great now. -- Guillaume http://www.postgresql.fr http://dalibo.com