[PATCH] Revert "Revert "Render: Use built-in SHA1 library""

[Carl Worth]

On Mon, 12 Apr 2010 19:22:44 +0300, Tiago Vignatti <tiago.vignatti at nokia.com> wrote:
> My simple argument against Keith's decision is simple: we can save 316 kB of
> RSS in a standalone Xorg call. Therefore, I'm in favor to keep our own very
> simple and shiny SHA1 implementation.

It does seem clear to me that several people/projects want a tiny SHA1
implementation. In the notmuch project, I needed SHA1 and ended up not
linking against openssl, but instead included a copy of the libsha1 code
From Brian Gladman and Mikhail Gusarov. (Code which I originally found
in my email inbox in response to using openssl for SHA1 in X).

Of course, people complained about me embedding a copy of code, and that
I should instead be using some library function to ease the ability of
distributions tracking security bugs, etc.

It does seem to me that it would be useful to have a tiny library that
just did SHA1. Then it could be packaged in the distributions, both the
X server and notmuch could use it, and everyone could be happy.

Are the linker bugs that make it expensive to load many small libraries
instead of one large library still prevalent, or have those been fixed?

> Ideally I'd try to analyse some eventual performance regressions on
> render. Do you guys think we should care with this? Carl?

"Eventual" performance regressions? Are you trying to set things up so
that we can test future problems when they arise? Or are there current
regressions you've seen that you'd like to investigate?

-Carl
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://lists.x.org/archives/xorg-devel/attachments/20100413/dd5482b5/attachment.pgp>