Nick Hengeveld [off-list ref] writes:
The way I understand this sequence, one and two will not be writing to
the same file. When two unlinks tmpfile, one will still be able to
write to its local, but the body of the file that one is writing will
be removed when it closes local.
True. What I meant by "interesting" is that two is reading from
what one is writing.
quoted
Not that I think the multiple instances should be prevented at
this low level --- if they stomp on each other at this level, it
is very likely the they are doing duplicated work on the network
side as well, and should be prevented from doing so at much
higher level than this, I think. That's why I said I do not
mind BKL upfront in git-fetch.sh.
True, is that something I should include with the partial patch?
No.
On Tue, Sep 27, 2005 at 10:09:49AM -0700, Junio C Hamano wrote:
True. What I meant by "interesting" is that two is reading from
what one is writing.
Excellent point. Given the potential for problems related to this
issue or to compressed representation, would it make sense to enable
partial transfers via a command-line option and leave the feature
disabled by default?
--
For a successful technology, reality must take precedence over public
relations, for nature cannot be fooled.
Nick Hengeveld [off-list ref] writes:
On Tue, Sep 27, 2005 at 10:09:49AM -0700, Junio C Hamano wrote:
quoted
True. What I meant by "interesting" is that two is reading from
what one is writing.
Excellent point. Given the potential for problems related to this
issue or to compressed representation, would it make sense to enable
partial transfers via a command-line option and leave the feature
disabled by default?
Here is what I think. Multiple transfers at the same time, if
it becomes problem in pratcice, should be prevented at the upper
layer anyway, so this "interesting" problem is a moot point. So
I'd say let's not worry about that.
Doing partial transfer support for packs and indices would be
more useful addition to your code than dealing with the above --
packs are much bigger than individual objects and would get much
bigger benefit from restartable download.
Different compressed representation is just a theoretical
curiosity and we do not know if it is a problem in practice. As
long as you can detect the situation and die, it would be OK, I
think.