← Back to team overview

anvil-dev team mailing list archive

Re: [Devstackpy] execute_template() shell argument

 

I got it working this morning. See my latest pull request.

On Tue, Mar 20, 2012 at 10:27 PM, Doug Hellmann <doug.hellmann@xxxxxxxxxxxxx> wrote:
Ok, I'll experiment tomorrow and see what happens. :)

Doug

On Mar 20, 2012, at 9:23 PM, Joshua Harlow <harlowja@xxxxxxxxxxxxx> wrote:

Sure,

It seemed liked it stopped python from escaping some of the mysql commands, like the stuff inside quotes.

If your change doesn’t need that then might as well remove it.

>From what I remember it was needed, but might not be if investigated more.

On 3/20/12 6:22 PM, "Doug Hellmann" <doug.hellmann@xxxxxxxxxxxxx> wrote:

I guess the question really is, why is it a shell command sometimes and not other times?

Doug

On Mar 20, 2012, at 6:50 PM, Joshua Harlow <harlowja@xxxxxxxxxxxxx> wrote:

Re: [Devstackpy] execute_template() shell argument I think the subprocess says for a shell command it should be a string rather than a list.

Or that’s what I thought.

That’s why that is there. If we can get rid of the flag then I’m all for it.

On 3/20/12 3:07 PM, "Doug Hellmann" <doug.hellmann@xxxxxxxxxxxxx> wrote:

Joshua,

The utils.execute_template() function and the execute() function in the shell module use a "shell" flag and treat the incoming command differently. That's causing some confusion with the db commands like grant_all, since the commands have some funky quote escaping. Is there any reason to parse that command line, substitute values in the individual parts, then re-assemble it to a single big string before executing it? I could either skip the parsing, or skip the reassembly, but I wasn't sure why it ended up the way it is now so I wasn't sure which way to go.

I'll be on IRC tomorrow if my question is too vague to make sense. :-)

Doug





References