Git/log: Difference between revisions

(7 intermediate revisions by the same user not shown)

Line 1:

We use <tt>git log</tt> for creating [[RELEASE NOTES]]

[[File:Git log color.png|thumb]]

We use <code>git log</code> for creating [[RELEASE NOTES]] (Refer to the script there.)

Line 278:

Line 279:

You also get '''<tt>tformat</tt>''' behavior if you do not specify any 'format' word, but use '%-encoded' formatting strings. Thus, the following is equivalent to the previous example:<br>

<code>git log -3 --pretty="* %h (%as) %(decorate:prefix=,suffix=,tag=v,separator= ) - %an: %s %b" 39.5.0...HEAD</code>

== CSV Output for spreadsheets ==

By simply formatting your output into a "Comma Separated Values" file, you can easily work with log data in a spreadsheet with LibreOffice Calc. (Note that LibreOffice Calc does not support reading data directly from STDIN, so you can't use process substitution. You must create a (temporary) file from the log, and then open that file with LibreOffice.

<code>git log nasa/main...origin/REL1_43 --date=iso-strict --pretty=format:'"%h","%ad","%an","%s"' --cherry-pick --right-only > meza.log.csv && libreoffice --calc meza.log.csv</code>

== Keywords ==

If your team writes [[Git/hacks#Good_and_Consistent_commit_messages|good commit messages]] with consistent keywords to describe the type of commits made, you can later easily filter and sort issues for better [[RELEASE NOTES]] In the example below, we filter and sort commits based on subjects that start with common actions or keywords.

If you have no idea what keywords the team is using, have a look:

<code>git log 39.5.0..HEAD --pretty=format:"%s" --no-merges | awk '{print tolower($1)}' | sort | uniq -c</code>

or for bigger repos, limit the lookback, and sort things numerically in reverse order, showing the top 20

<code>git log --since "2 years ago" --pretty=format:"%s" --no-merges | awk '{print tolower($1)}' | sort | uniq -c | sort -nr | head -20</code>

== Comparing branches and foreign (upstream) repos ==

As you may know, the triple dot notation means to find the common ancestor. Using that with <code>shortlog</code> allows to quickly find the differences between branches or even 'upstream' repos - especially in conjunction with <code>--cherry-pick --right-only</code>.

This is the command you want if you've ever wondered how GitHub shows you that your fork is '''X commits ahead, Y commits behind''<nowiki/>' repo ACME and you want to get a quick overview of the situation.

<code>git shortlog nasa/main...origin/REL1_43 --cherry-pick --right-only</code>

== Cherry on top ==

By using the --cherry-pick option with --right-only, we can omit equivalent commits that appear on both sides.

git log --graph --full-history nasa/main...origin/REL1_43 --color --date=short --pretty=format:"%x1b[31m%h%x09%x1b[32m%d%x1b[0m%x20%ad %s" --cherry-pick --right-only

</syntaxhighlight>

@@ Line 1: / Line 1: @@
-We use <tt>git log</tt> for creating [[RELEASE NOTES]]
+[[File:Git log color.png|thumb]]
+We use <code>git log</code> for creating [[RELEASE NOTES]] (Refer to the script there.)
@@ Line 278: / Line 279: @@
 You also get '''<tt>tformat</tt>''' behavior if you do not specify any 'format' word, but use '%-encoded' formatting strings. Thus, the following is equivalent to the previous example:<br>
 <code>git log -3 --pretty="* %h (%as) %(decorate:prefix=,suffix=,tag=v,separator= ) - %an: %s %b" 39.5.0...HEAD</code>
+== CSV Output for spreadsheets ==
+By simply formatting your output into a "Comma Separated Values" file, you can easily work with log data in a spreadsheet with LibreOffice Calc. (Note that LibreOffice Calc does not support reading data directly from STDIN, so you can't use process substitution. You must create a (temporary) file from the log, and then open that file with LibreOffice.
+<code>git log nasa/main...origin/REL1_43 --date=iso-strict --pretty=format:'"%h","%ad","%an","%s"' --cherry-pick --right-only > meza.log.csv && libreoffice --calc meza.log.csv</code>
+== Keywords ==
+If your team writes [[Git/hacks#Good_and_Consistent_commit_messages|good commit messages]] with consistent keywords to describe the type of commits made, you can later easily filter and sort issues for better [[RELEASE NOTES]] In the example below, we filter and sort commits based on subjects that start with common actions or keywords.
+<code>git log 39.5.0..HEAD --pretty=format:"%s" --no-merges | grep -Ei "^(add|bug|chore|delete|doc|feat|fix|refactor|remove|style|test|update)" | sort</code>
+If you have no idea what keywords the team is using, have a look:
+<code>git log 39.5.0..HEAD --pretty=format:"%s" --no-merges | awk '{print tolower($1)}' | sort | uniq -c</code>
+or for bigger repos, limit the lookback, and sort things numerically in reverse order, showing the top 20
+<code>git log --since "2 years ago" --pretty=format:"%s" --no-merges | awk '{print tolower($1)}' | sort | uniq -c | sort -nr | head -20</code>
+== Comparing branches and foreign (upstream) repos ==
+As you may know, the triple dot notation means to find the common ancestor. Using that with <code>shortlog</code> allows to quickly find the differences between branches or even 'upstream' repos - especially in conjunction with <code>--cherry-pick --right-only</code>.
+This is the command you want if you've ever wondered how GitHub shows you that your fork is '''X commits ahead, Y commits behind''<nowiki/>' repo ACME and you want to get a quick overview of the situation.
+<code>git shortlog  nasa/main...origin/REL1_43 --cherry-pick --right-only</code>
+== Cherry on top ==
+By using the --cherry-pick option with --right-only, we can omit equivalent commits that appear on both sides.
+<syntaxhighlight lang="bash">
+git log --graph --full-history nasa/main...origin/REL1_43 --color --date=short --pretty=format:"%x1b[31m%h%x09%x1b[32m%d%x1b[0m%x20%ad %s" --cherry-pick --right-only
+</syntaxhighlight>