Re: [PATCH] kernel32: fix command line argument escaping in build_command_line

[Vincent Povirk]

Is it really necessary to rewrite this loop? It'd be better to fix the
existing code, if possible.

[PATCH] kernel32: fix command line argument escaping in build_command_line

[Jacob Lifshay]

The old loop seemed confusing to me and it overestimated the size of the
memory to allocate by a large factor. The old loop would need to have about
half of it rewritten to fix the size overestimation and a little more for
fixing the original bug. If you think it's very important to keep as much
of the old code as possible, I can rewrite the patch that way, but I don't
think that's necessary.
Jacob Lifshay

On Feb 7, 2017 7:43 PM, "Vincent Povirk"  wrote:

Is it really necessary to rewrite this loop? It'd be better to fix the
existing code, if possible.

[PATCH] kernel32: fix command line argument escaping in build_command_line

[Alexandre Julliard]

Jacob Lifshay  writes:
> The old loop seemed confusing to me and it overestimated the size of
> the memory to allocate by a large factor. The old loop would need to
> have about half of it rewritten to fix the size overestimation and a
> little more for fixing the original bug. If you think it's very
> important to keep as much of the old code as possible, I can rewrite
> the patch that way, but I don't think that's necessary.

Please try to fix separate issues as separate patches instead of
rewriting everything. The overestimation is very unlikely to be a
problem; in fact I'd say that something simple like doubling the
incoming size would be better than a complicated calculation.