Add a useful return value to git-check-attr
PATCH 1/2 Add a useful return value to git-check-attr
git-check-attr previously always returned success. With this patch it returns success when the requested attribute is found for all supplied paths.
This lets you check in a script whether a file has an attribute:
$ git-check-attr attribute -- file && echo "file has attribute"
---
This is handy in itself, but I added it to support the tests in the next patch.
builtin-check-attr.c | 5 ++++- 1 files changed, 4 insertions(+), 1 deletions(-)
builtin-check-attr.c +4 -1
@@ -9,6 +9,7 @@ int cmd_check_attr(int argc, const char **argv, const char *prefix) {struct git_attr_check *check;int cnt, i, doubledash; +int allset =1 ; doubledash =-1 ;for (i =1 ; doubledash <0 && i < argc; i++) {
@@ -43,7 +44,9 @@ int cmd_check_attr(int argc, const char **argv, const char *prefix) (check[j].isset <0 ) ?"unspecified" : (check[j].isset ==0 ) ?"unset" : "set"); +if ( check[j].isset <=0 ) + allset =0 ; } } -return 0 ; +return !allset; }
Signature · 1 line
1.5.1.1.821.g88bdb
Quoted 7 lines · Andy Parkins
git-check-attr previously always returned success. With this patch it returns success when the requested attribute is found for all supplied paths.
This lets you check in a script whether a file has an attribute:
$ git-check-attr attribute -- file && echo "file has attribute"
I wrote check-attr primarily to have a quick way to make sure the attribute stack is working properly, but I agree making gitattributes available to scripts is important to make it useful in general.
And your patch was a step in the right direction based on my previous patch.
But sorry, you were shooting at a target that is still moving too fast, and you missed. It's not your fault, as even I did not know where it was heading to, but my plan has been to have the scripts parse the text output, as we would need more than true/false values, as I did in the four patch series I sent out tonight.
Quoted 9 lines · Junio C Hamano
And your patch was a step in the right direction based on my previous patch.
But sorry, you were shooting at a target that is still moving too fast, and you missed. It's not your fault, as even I did not know where it was heading to, but my plan has been to have the scripts parse the text output, as we would need more than true/false values, as I did in the four patch series I sent out tonight.
Not a problem. It was a couple of seconds work. I added it so that I could test my keyword expansion stuff.
Andy
Signature · 2 lines
Dr Andy Parkins, M Eng (hons), MIET [email protected]