GTNPORTAL-3560 - Changed the way the cache-control is sanitized.

[jpkrohling]

No description provided.

[jpkrohling]

@ppalaga or @lucasponce , could you review it? I was a bit concerned at first with the possible performance impact of replaceAll, but as the outcome will always be the same for a long time, it should be optimized by the VM.

[lucasponce]

Hi Juca,

Compiler should optimize that. But for example, looking in other project
(Undertow) with performance constraints, they use something like a chain:

return msg.replace("<", "<").replace(">", ">").replace("&", "&");

Just in case if you prefer this other approach.

On Wed, Nov 12, 2014 at 1:22 PM, Juraci Paixão Kröhling <
notifications@github.com> wrote:

@ppalaga https://github.com/ppalaga or @lucasponce
https://github.com/lucasponce , could you review it? I was a bit
concerned at first with the possible performance impact of replaceAll, but
as the outcome will always be the same for a long time, it should be
optimized by the VM.

—
Reply to this email directly or view it on GitHub
#917 (comment).

[jpkrohling]

Lucas, the problem is that there might be more than one new line on the string. So, I'd need to iterate to replace all of them. So, in comparison, I think replaceAll is a better solution.

[ppalaga]

@jpkrohling +1

[lucasponce]

Hey, the patch is perfect :-).

[ppalaga]

Squashed the two commits and merged to master 979cbd2 and 3.8.x 69c37c8