$ git version
git version 2.9.3.windows.1
I seem to be having trouble with the + (one or more) pattern matcher in .gitignore. Here is a test with + and *used as a pattern qualifier for digits embedded in the file name. I want to keep Test.txt in source control and ignore the Test[0-9]+.txt files.
Directory:
$ ll | grep Test
-rw-r--r-- 1 asdf 1049089 0 Apr 5 15:55 Test.txt
-rw-r--r-- 1 asdf 1049089 0 Apr 5 15:55 Test123.txt
-rw-r--r-- 1 asdf 1049089 0 Apr 5 15:55 Test1235.txt
-rw-r--r-- 1 asdf 1049089 0 Apr 5 15:55 Test124.txt
Git ignore file contents Test 1:
Test[0-9]+.txt
Test 1:
$ git status | grep Test
Test.txt
Test123.txt
Test1235.txt
Test124.txt
This should have filtered out the 2nd through 4th files.
Git ignore file contents Test 2:
Test[0-9]*.txt
Test 2:
$ git status | grep Test
Test.txt
This should have filtered out ALL of the files.
Why is this happening?
The .gitignore file is using globbing, not regular expressions. So using + is just not valid.
So your second attempt is actually closer to what you want.
Test*.txt
This would filter out files starting with Test and ending with .txt with anything in between. But it would also catch Test.txt itself because a * in glob can be zero-width.
What you have may be the simplest option.
Test[0-9]*.txt
That will match Test followed by any digit, followed by anything (including nothing), followed by .txt.
So it is imperfect, in that it will also match on
Test1foo.txt
Which has non-numeric text past the digit. But how specific your requirement is will dictate this. You can also just add a few patterns, like
Test[0-9].txt
Test[0-9][0-9].txt
Test[0-9][0-9][0-9].txt
Test[0-9][0-9][0-9][0-9].txt
Test[0-9][0-9][0-9][0-9][0-9].txt
If you really have to be sure it's all digits, then just pick how many digits, and use multiple entries like that to make it a very specific match.