On Jun 10 20:51, Corinna Vinschen wrote: > On Jun 10 16:59, Aaron Schneider wrote: > > > Date: Fri, 10 Jun 2016 12:00:27 +0200 > > > From: corinna-cyg...@cygwin.com > > > > > > Cygwin's /dev/sd* emulation doesn't support more than 15 partitions yet. > > > This would need some rework of the device naming code. > > > > > > > > > Corinna > > > > I see. I though that it was a Windows problem that didn't provide > > the partitions properly to cygwin if they weren't formatted with a > > known filesystem or something related. > > > > Isn't any other way to access those partitions even from Windows > > Powershell? > > There may be a way even using Cygwin. If you know which disk X it is, > you could try accessing it via /proc/sys/Device/HarddiskX/PartitionY. > No guarantee but for raw access this might be sufficient. > > > Is the feature to support more than 15 partition planned > > to be implemented? > > I looked into that and theoretically the changes are small. > > However, there's an undesired side-effect. The device parser code uses > an auto-generated minimally pruned 0-trie, created by a tool called > shilka. The problem is that the resulting object file is already very > big, compared to the rest of the DLL.
> Amazing but true, about 20% of > the object file is just this trie code and its data, whihc, Uh, this sentence makes no sense. Let's try again: Amazing but true, about 20% of the Cygwin DLL size is just this device parser code and its data, which... > IMHO, is > already quite bad. By adding support for partitions 16-63, this object > code gets 2.5 times bigger! > > This is way too much, and I think what we have to do first is to > replace this autogenerated 0-trie code by some simpler, hand-written > parser before we can consider adding support for more partitions. Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Maintainer cygwin AT cygwin DOT com Red Hat
signature.asc
Description: PGP signature
