| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2021-10-08 | exp: repo 4k, better summary format, fix diff sizes | n-peugnet | |
| 2021-10-08 | add chunkSize parameter to cli | n-peugnet | |
| 2021-10-07 | fix storeDelta by keeping previous raw bytes | n-peugnet | |
| this ensures that the last patch is created on top of the previous one. This way we know that it will be correctly applied on loadDelta Also directly write to wrapper instead of using an intermediate buffer. | |||
| 2021-10-07 | better exp logs and unlimit git rename detection | n-peugnet | |
| This should make for better git diff size perfs | |||
| 2021-10-06 | use repo deltacodec for recipe and files | n-peugnet | |
| this should not only be more efficient in space but also in time drop old Commit and Restore tests that do not use compression | |||
| 2021-10-06 | rename buildclean into mostlyclean target in root Makefil | n-peugnet | |
| 2021-10-06 | refactor: move repo and delta in their own package | n-peugnet | |
| 2021-10-06 | refactor main and logger | n-peugnet | |
| 2021-10-04 | add checks for consistency | n-peugnet | |
| It works ok, but man, git is good | |||
| 2021-10-04 | export env vars instead of passing params to exp.sh | n-peugnet | |
| 2021-10-01 | add some comments, summary, TODOs and infos | n-peugnet | |
| 2021-10-01 | compare against git diff gzipped | n-peugnet | |
| this is not looking good... | |||
| 2021-10-01 | call clean and all targets on submakes | n-peugnet | |
| 2021-09-30 | simplify exp by not using mktemp | n-peugnet | |
| 2021-09-30 | move script in its own file for comfort | n-peugnet | |
| and start making some stats | |||
| 2021-09-30 | run dna-backup on different commits | n-peugnet | |
| 2021-09-30 | first add of exp and Makefiles | n-peugnet | |
| 2021-09-30 | fix symlink tests for windows | n-peugnet | |
| ye olde filepath.Separator | |||
| 2021-09-30 | properly manage relative symlinks will CI pass ?? | n-peugnet | |
| 2021-09-30 | add some comments | n-peugnet | |
| 2021-09-29 | fix typos in README | n-peugnet | |
| 2021-09-29 | okay EvalSymlinks() should be what we want instead of Clean() | n-peugnet | |
| 2021-09-29 | ok this should do the trick for macos symlink test | n-peugnet | |
| fingerscrossed | |||
| 2021-09-29 | this time symlink test should be fixed for macos | n-peugnet | |
| 2021-09-29 | echo actual logger output in symlink test | n-peugnet | |
| 2021-09-29 | catch all symlink test errors | n-peugnet | |
| 2021-09-29 | add debug print for symlink test | n-peugnet | |
| 2021-09-29 | try to fix symlink tests for MacOS | n-peugnet | |
| 2021-09-29 | support symlinks if it is internal to source | n-peugnet | |
| adapt tests | |||
| 2021-09-29 | move old funcs to repo_test and add roundtrip test | n-peugnet | |
| 2021-09-29 | merge tryDeltaEncodeChunk and encodeTempChunk | n-peugnet | |
| to reuse the sketch calculation | |||
| 2021-09-29 | fix some typos | n-peugnet | |
| 2021-09-29 | add details and some todos | n-peugnet | |
| 2021-09-27 | add CLI with subcommands | n-peugnet | |
| 2021-09-24 | fix symlink tests on windows | n-peugnet | |
| split into symlink and NotReadable tests and do not build nowindows_test on windows. | |||
| 2021-09-23 | Testing with Fdelta as the differ | n-peugnet | |
| 2021-09-23 | better error handling and start checking symlinks | n-peugnet | |
| 2021-09-23 | use repo wrappers for files and recipe | n-peugnet | |
| 2021-09-23 | add String to Delta struct and use it in logs | n-peugnet | |
| Also switch to external package for slice_test as it was already a case of "black box testing". | |||
| 2021-09-23 | add SetLevel function to logger package | n-peugnet | |
| 2021-09-23 | better tests (util/io subtests + check log output) | n-peugnet | |
| 2021-09-22 | fix mystical bug by making multiple passes | n-peugnet | |
| We make ne passes while new blocks have been added. This way we are assured that the result will be the same on following runs. result : 161 /tmp/test-1/00001/recipe 161 /tmp/test-1/00001/files 0 /tmp/test-1/00001/hashes 40 /tmp/test-1/00001/chunks 482 /tmp/test-1/00001 9904052 /tmp/test-1/00000/recipe 5377304 /tmp/test-1/00000/files 2061331 /tmp/test-1/00000/hashes 207958371 /tmp/test-1/00000/chunks 225301178 /tmp/test-1/00000 225301740 /tmp/test-1 | |||
| 2021-09-22 | add tests for io utils | n-peugnet | |
| 2021-09-22 | rename and remove unused utils | n-peugnet | |
| 2021-09-22 | fix calling with source "." | n-peugnet | |
| 2021-09-22 | export unprefix in utils and fix error msg | n-peugnet | |
| 2021-09-22 | fix for stream smaller than chunkSize | n-peugnet | |
| 2021-09-22 | add debug log level | n-peugnet | |
| 2021-09-22 | trying to fix the mystical bug | n-peugnet | |
| If prev is not null and no match if found, always encode both remaining chunks. Previously some chunks of chunkSize could have been stored as TempChunks in the recipe instead of as StoredChunks with hashes and an Id. This did not fix the mystical bug. But it helped finding where it came from. | |||
| 2021-09-22 | misc log messages and refactoring | n-peugnet | |
 |
index : dna-backup | |
| deduplicated versionned backups for dna | Nicolas Peugnet |
| aboutsummaryrefslogtreecommitdiff |