chadnetzer / hardlinkable Goto Github PK

A tool to scan directories and report on the space that could be saved by hardlinking identical files. It can also perform the linking. Written in Go.

License: MIT License

Go 99.84% AMPL 0.16%

go hardlink compression unix posix tools cmd filesystem file files

hardlinkable's People

Contributors

Stargazers

Watchers

Forkers

jamescassell

hardlinkable's Issues

RFE: "oldest" behavior, opposite of "newest" behavior

Would be helpful to my use case to have an option to reverse the default "newest" behavior to have the program link instead against the oldest candidate.

Is this something you'd consider adding?

Seem like it'd be a customization or duplication of your code here:

hardlinkable/dolink.go

Lines 55 to 81 in 29ad3a2

 if fs.Options.UseNewestLink { 

 // Use destination file times if it's most recently modified 

 dstTime := dst.Mtim 

 if dstTime.After(src.Mtim) { 

 err := os.Chtimes(src.Pathsplit.Join(), dstTime, dstTime) 

 if err != nil { 

 fs.Results.FailedLinkChtimesCount++ 

 // Ignore this error, and just return early, as we 

 // don't want to abort the Run(). 

 return nil 

 } 

 // Keep cached inode.StatInfo time updated 

 si := fs.inoStatInfo[src.Ino] 

 si.Mtim = dst.Mtim 

 // Change uid/gid if possible 

 err = os.Lchown(src.Pathsplit.Join(), int(src.Uid), int(src.Gid)) 

 if err != nil { 

 fs.Results.FailedLinkChownCount++ 

 return nil 

 } 

 // Chown succeeded, so update the cached stat structures 

 si.Uid = dst.Uid 

 si.Gid = dst.Gid 

 } 

 }

Edit: note to add that this would be an additional option; keep the current behavior as your preferred default

Update file.Read() to process read bytes before checking error results

In the following golang/go commit by Rob Pike:

golang/go@7c2718b
golang/go#27818

it is demonstrated that for Go Read() errors with a non-zero byte count, it is
important to first process the returned bytes. In particular an EOF can be
returned with a non-zero read. This is documented for the io.Reader
interface, but not mentioned for the os.File type (and in fact implies
otherwise, imo).

In any case, our readerContentsEqual() func should probably be made to only
accept *os.File, instead of reader interfaces, because we specifically only
intend to support file comparisons. Further, adjusting the processing to
support the wording of the io.Reader() docs and the above example is probably
preferred; perform the comparison of any bytes returned before handling the
Read() error conditions.

chadnetzer / hardlinkable Goto Github PK

hardlinkable's People

Contributors

Stargazers

Watchers

Forkers

hardlinkable's Issues

RFE: "oldest" behavior, opposite of "newest" behavior

Update file.Read() to process read bytes before checking error results

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent

	if fs.Options.UseNewestLink {
	// Use destination file times if it's most recently modified
	dstTime := dst.Mtim
	if dstTime.After(src.Mtim) {
	err := os.Chtimes(src.Pathsplit.Join(), dstTime, dstTime)
	if err != nil {
	fs.Results.FailedLinkChtimesCount++
	// Ignore this error, and just return early, as we
	// don't want to abort the Run().
	return nil
	}

	// Keep cached inode.StatInfo time updated
	si := fs.inoStatInfo[src.Ino]
	si.Mtim = dst.Mtim

	// Change uid/gid if possible
	err = os.Lchown(src.Pathsplit.Join(), int(src.Uid), int(src.Gid))
	if err != nil {
	fs.Results.FailedLinkChownCount++
	return nil
	}
	// Chown succeeded, so update the cached stat structures
	si.Uid = dst.Uid
	si.Gid = dst.Gid
	}
	}