EOL filter only applied to files when first checked out
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Bazaar |
Fix Released
|
High
|
Martin Pool | ||
2.0 |
Fix Released
|
High
|
Ian Clatworthy |
Bug Description
The initial checkout applies the content filtering (and you run into https:/
Subsequent updates to the branch do not apply the content filter.
Reproduce (possibly not the most streamlined way of doing this)
# On Linux
bzr init one --1.14
cd one
echo -e "one\ntwo" > test.txt
bzr add
bzr commit -m "One"
# On Windows
Edit %appdata%
[name *.txt]
eol = native
bzr co bzr+ssh:
cd one
bzr st
# "text.txt" is listed as changed ; #362030 - already known and fixed
# open file in notepad2 or other modern editor and you see it has CRLF endings
# Linux
echo -e "three\nfour" > test2.txt
bzr add
bzr commit -m "Two"
# Windows
bzr up # test2.txt is added
bzr st
# Still only test.txt is changed
# open test2.txt and it has LF endings
Related branches
- Robert Collins (community): Needs Fixing
- Diff: 325 lines
- Robert Collins (community): Approve
- Diff: None lines
Changed in bzr: | |
assignee: | nobody → Ian Clatworthy (ian-clatworthy) |
Changed in bzr: | |
importance: | Undecided → High |
status: | New → Confirmed |
Changed in bzr: | |
status: | Confirmed → In Progress |
tags: | added: content-filtering |
Changed in bzr: | |
milestone: | 2.0 → 2.0rc2 |
Changed in bzr: | |
assignee: | Ian Clatworthy (ian-clatworthy) → Martin Pool (mbp) |
Changed in bzr: | |
milestone: | 2.0 → none |
status: | Fix Committed → In Progress |
Changed in bzr: | |
status: | In Progress → Fix Released |
Changed in bzr: | |
milestone: | none → 2.1.0b4 |
Adrian,
If you get a chance, please give the linked branch a *quick* test and let me know if it improves things. (I've done next to no testing on the branch but I think/hope it's in the right direction.)
If it proves successful, I'd be curious as to how many edge cases it solve. As well as update, it might also fix similar problems with merge and pull? And for good measure, hopefully it does the "right thing" w.r.t. generation of conflict-related files (x.THIS, x.OTHER, x.BASE)?