larsxschneider@xxxxxxxxx writes: > From: Lars Schneider <larsxschneider@xxxxxxxxx> > > The goal of this series is to avoid launching a new clean/smudge filter > process for each file that is filtered. I tentatively queued this in 'pu' because I wanted to see these changes in context and also wanted to know if there are overlaps and conflicts with other topics in flight. As your earlier steps renamed packet_write() to packet_write_fmt() but didn't add a new packet_write() that has different semantics, it was pleasantly easy to make sure there is no new caller added in the meantime (it did conflict with Duy's shallow-deepen topic and the resolution had to touch outside the <<< conflicted === regions >>>, but it was otherwise trivial). Some details of the protocol (I think Torsten has already pointed out that not all paths can be representable with the current incarnation; there may be other minor nits like that) may still need to be worked out, but I think the series at this point gets the basic code structure right and such additional fixes would not change things too drastically (IOW, I think we are very close to being 'next'-ready). Thanks.