On Sat, 23 Feb 2019 at 15:48, Vladimir Sitnikov <[email protected]> wrote: > > > > 3) Tags (e.g. version tags) are located at separate branches which > > > complicates history browsing > > Can you give an example of this? I don't get what you mean. > > Please find an attached gitk view of the current Git mirror. > > In case you are not familiar with output, commits grow from the bottom > (just like regular trees do). > > It bothers me in multiple ways: > 1) V5_1_RC1, v5_1_RC2, and v5_1 tags are placed OUTSIDE of the trunk > history. In other words, if I browse "trunk" history alone, I don't > see tags at all > 2) It is quite obscure which repository state is represented by RC. > 3) V4_0 tags seem to be a complete mess (see attached > jmeter_v4_history.png). For instance, can you tell which commits are > there between V4_0_RC2 and V4_0_RC4?
$ svn log https://svn.apache.org/repos/asf/jmeter/tags/v4_0_RC2 -v -l 1 ------------------------------------------------------------------------ r1822462 | milamber | 2018-01-28 14:39:59 +0000 (Sun, 28 Jan 2018) | 1 line Changed paths: A /jmeter/tags/v4_0_RC2 (from /jmeter/trunk:1822461) Tag for pre-release ------------------------------------------------------------------------ $ svn log https://svn.apache.org/repos/asf/jmeter/tags/v4_0_RC4 -v -l 1 ------------------------------------------------------------------------ r1822613 | milamber | 2018-01-30 07:50:10 +0000 (Tue, 30 Jan 2018) | 1 line Changed paths: A /jmeter/tags/v4_0_RC4 (from /jmeter/trunk:1822612) M /jmeter/tags/v4_0_RC4/build.xml Tag for pre-release ------------------------------------------------------------------------ i.e. v4_0_RC2 was created from jmeter/trunk:1822461 v4_0_RC4 was created from jmeter/trunk:1822612 >From which you can work out what trunk changes were made between the two tags. > Can you tell why V4_0_RC5 is not between V4_0_RC4 and V4_0_RC6? The derivations as above are: A /jmeter/tags/v4_0_RC5 (from /jmeter/trunk:1822958) A /jmeter/tags/v4_0_RC6 (from /jmeter/trunk:1822964) Seems to me that the tags *are* in the correct order in SVN. If they are incorrect in git, then it seems to me that the conversion to Git has failed to preserve history. > We don't have lots of tags, so we could even "migrate" that manually, > and it could be faster than trying to make that automatic. > What do you think? > A bit of a side question: what is the purpose of having docs-x.y branches? > What is the purpose of placing generated javadocs to docs-x.y branch? > > Vladimir
