openoffice-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Matthias Seidel <matthias.sei...@hamburg.de>
Subject Re: get Revision from Git (short Hash)
Date Wed, 14 Aug 2019 22:07:38 GMT
Hi Kay,

Am 15.08.19 um 00:02 schrieb Kay Schenk:
> On Wed, Aug 14, 2019 at 1:24 PM Marcus <marcus.mail@wtnet.de> wrote:
>
>> Am 14.08.19 um 22:02 schrieb Jim Jagielski:
>>>> On Aug 14, 2019, at 10:51 AM, Andrea Pescetti <pescetti@apache.org>
>> wrote:
>>>> Matthias Seidel wrote:
>>>>> We already have the build id, the build
>>>>> date and now the git hash (which is a unique link to the last commit
it
>>>>> was based on)
>>>>> This is how we did it with SVN, why should we change it?
>>>> Because we are dropping information. The SVN revisions are always
>> increasing, and thus (independent on the build date, which can be at any
>> moment) I can compare two builds and retain information on which came first.
>>>> With git of course this doesn't hold, i.e., you cannot say which commit
>> came earlier between abcd1234 and 5678abcd. So I see some added value if we
>> enrich it this way.
>>> Is that needed though? I had thought the basic reason for having the SVN
>> ID is that the end-user knows, for sure, which SVN revision their app was
>> built from.
>>
>> it's unrealistic that the commit was done, e.g., today but the build
>> weeks later. So, Git hash and build date is not done at the exact same
>> date and time. But nearly. And here it think it's sufficiant.
>>
>> But when we decide to prefix the hash with a date value it's OK for me,
>> too.
>>
>> Marcus
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
>> For additional commands, e-mail: dev-help@openoffice.apache.org
>>
>>
> I think the date and hash should be displayed in the "build information"
> screen as the revision information was previously. In Jim's sample display,
> although the date is displayed, there is no indication of actual "revision"
> (hash).

This is simply because the code we are discussing about is still not
committed.

I applied Peters patch and it looks like this:

https://www.dropbox.com/s/tkal1y9b09vrhse/VirtualBox_Windows%2010%20AOO-Build_14_08_2019_16_14_33.png?dl=0

Matthias

>


Mime
View raw message