[Lldb-commits] [PATCH] D28305: [Host] Handle short reads and writes, take 3

2018-11-26 Thread Pavel Labath via Phabricator via lldb-commits
labath added a comment. This is a stale patch which I am abandoning. I am just cleaning up my queue. CHANGES SINCE LAST ACTION https://reviews.llvm.org/D28305/new/ https://reviews.llvm.org/D28305 ___ lldb-commits mailing list lldb-commits@lists.l

[Lldb-commits] [PATCH] D28305: [Host] Handle short reads and writes, take 3

2018-11-26 Thread Greg Clayton via Phabricator via lldb-commits
clayborg added a comment. Any reason we are removing File::SeekFromCurrent and File::SeekFromEnd? CHANGES SINCE LAST ACTION https://reviews.llvm.org/D28305/new/ https://reviews.llvm.org/D28305 ___ lldb-commits mailing list lldb-commits@lists.llvm

[Lldb-commits] [PATCH] D28305: [Host] Handle short reads and writes, take 3

2017-01-06 Thread Pavel Labath via Phabricator via lldb-commits
labath added a comment. Ok, this is going to be more complicated than I expected. libedit takes a `FILE*` in its interface, so it's going to be very hard to get rid of it. It would be possible to still use file descriptors everywhere, and create a fake FILE* using functions like `funopen` for t

[Lldb-commits] [PATCH] D28305: [Host] Handle short reads and writes, take 3

2017-01-05 Thread Pavel Labath via Phabricator via lldb-commits
labath added a comment. I've done a bit of investigation, and there is still a lot of `FILE*` usage in random places. I believe most of them would be pretty easy to get rid of (https://reviews.llvm.org/D28356 is one of the more complicated cases). However, the main problem I see is the SB API,

Re: [Lldb-commits] [PATCH] D28305: [Host] Handle short reads and writes, take 3

2017-01-04 Thread Zachary Turner via lldb-commits
Honestly I would just assume we blanket kill FILE* across the entire codebase. It's non-portable, doesn't interoperate with Python 3, doesn't interoperate with C++ library facilities, and is clunky to deal with on Windows since most of the useful APIs won't accept one. On Wed, Jan 4, 2017 at 2:04

[Lldb-commits] [PATCH] D28305: [Host] Handle short reads and writes, take 3

2017-01-04 Thread Greg Clayton via Phabricator via lldb-commits
clayborg added a comment. In https://reviews.llvm.org/D28305#635877, @labath wrote: > The way this deals with it is by using `GetDescriptor()` which extracts the > raw fd from FILE *, and then always deals with those. I guess this has the > potential of bypassing any libc cache that FILE * may

[Lldb-commits] [PATCH] D28305: [Host] Handle short reads and writes, take 3

2017-01-04 Thread Pavel Labath via Phabricator via lldb-commits
labath added a comment. The way this deals with it is by using `GetDescriptor()` which extracts the raw fd from FILE *, and then always deals with those. I guess this has the potential of bypassing any libc cache that FILE * may be using. I don't know whether we care about that, but if we do th

[Lldb-commits] [PATCH] D28305: [Host] Handle short reads and writes, take 3

2017-01-04 Thread Greg Clayton via Phabricator via lldb-commits
clayborg requested changes to this revision. clayborg added a comment. This revision now requires changes to proceed. This change removes support for using a FILE* instead of a file descriptor. This needs to be fixed. The old Read function used to do this: ssize_t bytes_read = -1; if (Descri

[Lldb-commits] [PATCH] D28305: [Host] Handle short reads and writes, take 3

2017-01-04 Thread Pavel Labath via Phabricator via lldb-commits
labath created this revision. labath added reviewers: zturner, clayborg. labath added a subscriber: lldb-commits. Herald added a subscriber: mgorny. This is a rework of https://reviews.llvm.org/D25783, which was blocked by my objections to using pwrite wholesale. This version keeps the benefits o