summaryrefslogtreecommitdiffstats
path: root/test
Commit message (Collapse)AuthorAgeFilesLines
* - Some more QUILT_PATCHES_PREFIX fixes.Andreas Gruenbacher2004-07-125-38/+39
|
* - Switch from echo to printf for all translations: This is moreAndreas Gruenbacher2004-07-0917-148/+188
| | | | | | | | | | | | | | | stable than having expansions in messages that might change independent of the message. - Add print_patch function and always use it when printing patch names. Remove -n options from various scripts and use a global switch QUILT_PATCHES_PREFIX in .quiltrc to decide between patch names with and without directory prefix. Depending on user experience this switch may eventually go away. - Add --diffstat option to refresh command: If given, this option inserts diffstat statistics at the end of the patch header, or refreshes the existing diffstat output. No special tags in the path file (%diffstat or the like) are needed. - Minor fix inserting changelog into RPM specfile.
* - Add common GNU Diff format options to diff and refresH commands:Andreas Gruenbacher2004-07-042-3/+99
| | | | | | | | | | | -u, -U num, -c, -C num. REname old -c option (combine patches) of diff command to --combine. - Extend the function that splits comments froM patches and syntax coloring to handLe combined diff format. - Honor the LAnG environment variable again; this goT broken at some point. - Set TEXTDOMAINDIR to where the translations are put so that Messages are found even when not installing into /usr.
* - Several Fixes from Joe Green <jgreen@mvista.com>:Andreas Gruenbacher2004-07-022-2/+3
| | | | | | | | | | | | | | | | + Another sed \t\r problem, in Makefile.in + Bad error message on "quilt delete" of non-existent patch + "import" doesn't remove all leading components on patch file + Misspelling in "import" help text - With using the GNU diff --label option, file timestamps disappeared. Specify the timestamps in --label by hand: This gives us added flexibility when generating patches; we may some day want to support diff styles other than unified. Also reported by Joe Green. - Add --backup option to quilt refresh. - Update documentation to QUILT_${COMMAND}_ARGS in .quiltrc. (The old settings still work but they are no longer documented.) - Update German translation.
* - Oops, forgot to add files.v0.34Andreas Gruenbacher2004-06-101-0/+44
|
* - Add meta-data version check/upgrade code from Martin Quinson andAndreas Gruenbacher2004-06-106-11/+21
| | | | | | | | | | | | | | me; new `quilt upgrade' command. - Handle new/removed files correctly in generated patches: If the old or new file is missing, the file name in the patch should be /dev/null. Fix test suite accordingly, and add test/create-delete.diff. GNU patch recognizes this, and does additional file existance tests when a patch creates/deletes a file. Also use the --label option of GNU diff instead of hand-editing file name headers. - Add test/Makefile for running the test suite. Note: The tests run against the installed version of quilt! - Bump version to 0.34.
* - Preserve the order of files in patches in the diff and refreshAndreas Gruenbacher2004-06-061-3/+8
| | | | | | | commands. Files added to a patch appear at the end of a patch. The files command also lists the files in the order in which they appear in patches. - Update to version 0.33.
* - backup-file.c: Add code to recursively search .pc directories.Andreas Gruenbacher2004-06-062-0/+55
| | | | | | | | | | Add a "no-op" mode as default, and allow to just unlink files. - apatch/rpatch: let backup-files search .pc directories instead of generating temporary file lists. - rpatch/pop: unlink files of topmost patch after popping instead of unlinking at each step. - If QUILTRC is set in the environment, use this as the configuration file.
* - Revert `quilt fork' to how it originally worked: fork theAndreas Gruenbacher2004-03-133-15/+138
| | | | | | | | | | | | | | *topmost* patch, not the next unapplied patch. Adapt the test suite accordingly. - Add flag to backup-files.c to make sure that files in the working tree have a link count of at most one. Use this where appropriate. - Fix a bug in `quilt snapshot': Taking snapshots caused files to get link counts bigger than one, so modifying them could cause multiple files to get modified. Speed up the command, too. - Update the documentation. Add example1.test from the documentation.
* - Remove stray ``../'' in test script.Andreas Gruenbacher2004-03-101-1/+1
|
* - Sub-directory support:Andreas Gruenbacher2004-02-204-14/+38
| | | | | | | | | | + Fix another bug that triggers when /patches exists. + Working in sub-directories: Force `quilt new' to always create the new patch in the current working directory even if there is a patches/ sub-directory further up the directory tree. This should restore the behavior from before Jan 28 very well. - quilt files, diff, refresh: Sort the files in the patch alphabetically.
* - No longer remove common extensions (.dif, .diff, .patch, .gz,Andreas Gruenbacher2004-01-3111-109/+105
| | | | | | | .bz2) from patch names. This allows patches in the same series that only differ by extension, and cleans up the scripts a little. Also don't remove extensions when printing patch names. - Remove unused diffstat code.
* - Add a few more test cases for subdirectory support.Andreas Gruenbacher2004-01-284-5/+234
| | | | - Fix a few bugs introduced while merging Chris's patch.
* - Add support for working in subdirectories of the base directoryAndreas Gruenbacher2004-01-281-0/+73
| | | | | | | that contains patches/ and .pc/. In the unlikely case that quilt shall operate in the sub-directory itself, it is sufficient to create a patches/ directory there. - Add a small testcase for subdirectory support.
* - `quilt fold' did not create directories for additional files,Andreas Gruenbacher2003-11-271-9/+9
| | | | and the code for recovering from failures was wrong.
* - Add new command `quilt fold' to fold one patch into another.Andreas Gruenbacher2003-11-172-4/+57
| | | | | | | | - Fix bug in `quilt files': Files that are neither present in the working directory and have an empty backup copy sometimes were still listed. - Some minor cleanups in apatch. - Add test/fold.test, and fix a now-useless check in one.test.
* - Add failpop.test that checks if modified patches are recognized.Andreas Gruenbacher2003-11-041-0/+27
| | | | | - Fix the check for modified patches: That broke with the .pc file removal.
* - No longer verify if patches remove cleanly by reverse applyingAndreas Gruenbacher2003-10-312-3/+3
| | | | | | | | | them: This doesn't always succeed, and the the equivalent of `quilt diff -z' should suffice, anyway. - Add `touch .pc/$patch' after successful refresh: Speeds up `quilt pop' after a refresh. - Update test suite. - apatch was printing a spurious newline.
* - Remove .pc/*/.pc files, and adjust various scripts accordingly.Andreas Gruenbacher2003-10-281-16/+19
| | | | - Run test/one.test inside sub-directory.
* Add test case for patch merging with `quilt diff'.Andreas Gruenbacher2003-10-271-0/+87
|
* Add a new test case for conflicts while pushing.Andreas Gruenbacher2003-10-271-0/+139
|
* - Rename snapshot directory from ".snap0" to ".snap".Andreas Gruenbacher2003-10-261-0/+61
| | | | | | | | | - `quilt diff: Speed up file list generation in `quilt diff'. Don't use modified_files anymore. Instead, add first_modified_by function. I want to get rid of the patches_per_file cache; it sometimes was not brought up to date. Some cleanups. - Add test/snapshot.test. - Fix two errors in the date format in quilt.changes.
* - Never reorder files in patches. (Previously the file listAndreas Gruenbacher2003-04-094-172/+345
| | | | | | was run through (sort | uniq) to remove duplicate entries.) - Update test script to new format, and add regression test cases.
* Update test script to reflect recent patch format changes.Andreas Gruenbacher2003-03-231-2/+7
|
* Make test more robust by popping all patches with -R. Before no modeAndreas Gruenbacher2003-03-221-4/+4
| | | | | was specified and quilt was using a mix of reverse-applying and fast restoring.
* Merge version of James's temp file patch (some further improvements,Andreas Gruenbacher2003-02-161-4/+8
| | | | | hopefully). Also, fix some messages that seem to have been reset to a previous version during I18N, and update the test script.
* Start with a very minimal test suiteAndreas Gruenbacher2003-02-062-0/+338