This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author pitrou
Recipients Ramchandra Apte, barry, jcea, kxroberto, pitrou, r.david.murray
Date 2011-12-12.12:14:30
SpamBayes Score 5.551115e-17
Marked as misclassified No
Message-id <1323692057.21067.19.camel@localhost.localdomain>
In-reply-to <1323688736.04.0.379367816874.issue13580@psf.upfronthosting.co.za>
Content
> Overall I wonder why you post here on the main topic "resource usage",
> when you don't care about issues of magnitude 2x memory usage.

Because you are talking about a fixed overhead of a mere 2MB (IIUC),
which is moreover shared between all processes using libssl/libcrypto
(and chances are these libraries are already loaded by something else on
your system - crypto is useful after all - so Python would actually not
add anything substantial in that regard). It is IMHO not interesting at
all, but still, you are welcome to post an issue about it if you are
concerned.

On the other hand, we do care about "resource usage" issues when they
are about dynamic memory consumption, e.g. reducing the size of objects.
Or, of course, memory leaks.

[...]
> You have this sort of black-white arguments which are green and
> somehow really I think, you are perhaps misplaced in this category
> "resource usage".
[...]
> I indeed wonder about that careless style here meanwhile. Its not as
> it was years ago.

Nobody here is interested in exchanging personal attacks, I'm afraid.

> You are free to divide the issue if you really think its worth multiple.
> But why close it swift-handed before that is sorted out / set up?

It *is* sorted out. You complained about libssl being linked into the
executable and it turned out to be Debian-specific. Just because you
think there is some kind of "big picture" problem involved doesn't make
it ok to turn this issue into a catch-all for all related issues.

> So are you definitely saying/knowing, there is really no such
> mentioned optimized module selection  (~50% of so modules since
> Python2.5 on Debian) somewhere in the Python build files?
> (I ask first here to not create unnecessary lots of issues)

Let's say that I don't know about such a selection, and neither probably
do other core developers on this issue, otherwise they would already
have chimed in. There is certainly no such documented or tested thing,
anyway, so even if there were I wonder why the Debian people would
decide to use it.
History
Date User Action Args
2011-12-12 12:14:31pitrousetrecipients: + pitrou, barry, jcea, kxroberto, r.david.murray, Ramchandra Apte
2011-12-12 12:14:30pitroulinkissue13580 messages
2011-12-12 12:14:30pitroucreate