GAMS writer fails for power nested in a log

[qtothec]

The functionality to selectively replace ** with power() fails in an edge case when ** is nested within a log function, e.g. log(x ** 2). In replace_power(), the assert that arg is bound by ( and ) fails because the arg is bound by log( and ). This is possibly also true with other similar functions, e.g. exp().

@gseastream

[whart222]

The changes included in PR #272 address this issue on the expr_dev branch.

[qtothec]

I'm not sure it's good practice to close issues before their associated PR is merged. Github supports automatic closing of linked issues: https://github.com/blog/1506-closing-issues-via-pull-requests

[whart222]

The issue here is that code change on another branch resolved the issue. John argued that we should re-open and accept this PR, since it's a minor change that will impact users soon. But in general there isn't a standard way to address the situation that we have, where conflicting changes are happening on separate branches. I think it's OK to close a PR like this to avoid duplicating work.

…

On Mon, Dec 4, 2017 at 7:24 PM, Qi Chen ***@***.***> wrote: I'm not sure it's good practice to close issues before their associated PR is merged. Github supports automatic closing of linked issues: https://github.com/blog/1506-closing-issues-via-pull-requests — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#244 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAsb-Gzy3XG5RVHE3aL5kVZb_Mm7yAZzks5s9KlUgaJpZM4QJ8ue> .

[qtothec]

Huh? I think that I'm a bit confused here because I agree with what you say about conflicting PR's, but this is an issue, not a PR.

Rendered moot by the fact that #263 was merged, but just explaining the reason for my earlier post.