Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Trunk codesize numbers

0 views
Skip to first unread message

Boris Zbarsky

unread,
Sep 15, 2006, 8:47:33 PM9/15/06
to
So I decided to actually take a look at our codesize graphs today.... I really
wish we could annotate them with known "it's OK" issues, because they sure don't
look good. :(

Firefox codesize is up about 600KB on Linux since 6 months ago. Sadly, this is
Z, not mZ, but due to the static build the mZ number for Firefox is useless:

http://build-graphs.mozilla.org/graph/query.cgi?tbox=argo.mozilla.org&testname=codesize&autoscale=1&size=&units=bytes&ltype=&points=&showpoint=2006%3A09%3A15%3A16%3A54%3A03%2C15120636&avg=1&days=180

Seamonkey mZ on Linux shows about a 350KB size increase:

http://build-graphs.mozilla.org/graph/query.cgi?tbox=luna.mozilla.org&testname=codesize_embed&autoscale=1&size=&units=bytes&ltype=&points=&showpoint=2006%3A09%3A15%3A17%3A05%3A05%2C12782066&avg=0&days=180

Interestingly, the really big spikes are in somewhat different places on the two
graphs. That Seamonkey build doesn't build with cairo, so something else was
up, fwiw.

Seamonkey Z on Linux shows a 1.4MB size increase(!) with the huge spike matching
the big spike for mZ well.

In any case, are the big jumps known? Do we even care? I haven't seen any
codesize regression bugs filed or any reaction to codesize jumps on tinderbox....

-Boris

Boris Zbarsky

unread,
Sep 15, 2006, 8:59:16 PM9/15/06
to
Boris Zbarsky wrote:
> Seamonkey Z on Linux shows a 1.4MB size increase(!) with the huge spike
> matching the big spike for mZ well.

Actually, it's closer to 1MB.

It looks like the huge jump in mid-May for seamonkey (order of 200-300KB for Z
and about 100KB for mZ) corresponds to --enable-storage going on by default...
And over here, with an opt build, libstoragecomps.so is 289680 bytes. Stripped.

Is there anything we can do to make that a little smaller?

-Boris

Ludovic Hirlimann

unread,
Sep 19, 2006, 11:11:52 AM9/19/06
to dev-pl...@lists.mozilla.org, bzba...@mit.edu
On 9/16/06, Boris Zbarsky <bzba...@mit.edu> wrote:
> So I decided to actually take a look at our codesize graphs today.... I really
> wish we could annotate them with known "it's OK" issues, because they sure don't
> look good. :(
>
> Firefox codesize is up about 600KB on Linux since 6 months ago. Sadly, this is
> Z, not mZ, but due to the static build the mZ number for Firefox is useless:

care to explain What Z and mZ are ?

[snip]

> In any case, are the big jumps known? Do we even care? I haven't seen any

I do care - but doesn't count -

Ludovic
--
http://perso.hirlimann.net/~ludo/blog/

Boris Zbarsky

unread,
Sep 19, 2006, 1:20:11 PM9/19/06
to
Ludovic Hirlimann wrote:
> care to explain What Z and mZ are ?

Z is total codesize. mZ is "embedding" (core library) codesize. So for
example, layout is in both Z and mZ, while bookmarks are in Z but not in mZ (for
seamonkey; for Firefox it's all weird due to the static thing, as I said).

-Boris

0 new messages