Daniel wrote:
please do the patch to store the new value in _tr_buffer and attach it to mailing list or bug tracker. I haven't looked at code yet, but sounds like there is indeed an issue. I will review the patch and apply it.
What was the conclusion about this? Was the patch ever submitted and applied? If not, is the bug still there or has is been fixed by some other means? I'm asking because this looked like a quite serious issue.
-- Juha
Hello,
On 5/10/13 9:53 AM, Juha Heinanen wrote:
Daniel wrote:
please do the patch to store the new value in _tr_buffer and attach it to mailing list or bug tracker. I haven't looked at code yet, but sounds like there is indeed an issue. I will review the patch and apply it.
What was the conclusion about this? Was the patch ever submitted and applied? If not, is the bug still there or has is been fixed by some other means? I'm asking because this looked like a quite serious issue.
it was a patch that solved only for one case ( $var(x) = $(var(x){trans}) ), my comment on it what that it seems incomplete. If no patch shows up soon, I will just do it myself, not being something that big.
Thinking more of it, there are some corner cases when the same variabale is set with some transformation of it (not all transformations). If the expression on the right of the assignment has more that the variable and transformation, the it is not affected. Also, not all variables are exposed even when only with right side is the variable and transformation.
For example: $var(x) = $(var(x){s.substr...}); --- is affected $var(x) = $(var(x){s.substr...}) + "abc"; --- is not affected
Cheers, Daniel
Daniel-Constantin Mierla writes:
it was a patch that solved only for one case ( $var(x) = $(var(x){trans}) ), my comment on it what that it seems incomplete. If no patch shows up soon, I will just do it myself, not being something that big.
thanks, a fix would be highly appreciated.
-- juha
On 5/13/13 10:58 AM, Juha Heinanen wrote:
Daniel-Constantin Mierla writes:
it was a patch that solved only for one case ( $var(x) = $(var(x){trans}) ), my comment on it what that it seems incomplete. If no patch shows up soon, I will just do it myself, not being something that big.
thanks, a fix would be highly appreciated.
I committed a patch to fix it last evening. If goes fine on master branch, it can be backported at any time.
Cheers, Daniel
Daniel-Constantin Mierla writes:
I committed a patch to fix it last evening. If goes fine on master branch, it can be backported at any time.
ok, thanks. perhaps those people who reported the issue can test the fix. i have similar statements in my config but don't have a repeatable test case.
-- juha