{"thread":{"id":"24269","subject":"[PATCH v6 0/3] Merge renormalization, config renamed","startedAt":"2010-07-02T19:20:46Z","lastAt":"2010-08-05T19:02:08Z","messageCount":35,"participants":["Eyvind Bernhardsen","Junio C Hamano","Jonathan Nieder","Ævar Arnfjörð Bjarmason","Bert Wesarg"],"isPatch":true,"patchVersion":6,"patchTotal":3},"messages":[{"id":"144705","messageId":"cover.1278093311.git.eyvind.bernhardsen@gmail.com","threadId":"24269","inReplyTo":null,"subject":"[PATCH v6 0/3] Merge renormalization, config renamed","fromName":"Eyvind Bernhardsen","fromEmail":"eyvind.bernhardsen@gmail.com","sentAt":"2010-07-02T19:20:46Z","receivedAt":"2010-07-02T19:20:46Z","isPatch":true,"sender":{"key":"eyvind.bernhardsen@gmail.com","avatar":"https://avatars.githubusercontent.com/u/106762?v=4"},"body":"Here's a new version of the merge normalization series that renames the\nconfiguration variable. Since \"merge.renormalize\" got the best\nresponse, I went with that.\n\nJunio, I hope you don't mind that I squashed your patch to introduce the\nconfig variable into the first patch. It seemed a bit out-of-place as a\nseparate commit.\n\n- Eyvind\n\nEyvind Bernhardsen (3):\n Avoid conflicts when merging branches with mixed normalization\n Try normalizing files to avoid delete/modify conflicts when merging\n Don't expand CRLFs when normalizing text during merge\n\n Documentation/gitattributes.txt | 34 ++++++++++++++++++++\n Documentation/merge-config.txt | 10 ++++++\n builtin/merge.c | 3 ++\n cache.h | 2 +\n convert.c | 37 +++++++++++++++++++----\n environment.c | 1 +\n ll-merge.c | 15 +++++++++\n merge-recursive.c | 51 +++++++++++++++++++++++++++++-\n t/t6038-merge-text-auto.sh | 64 +++++++++++++++++++++++++++++++++++++++\n 9 files changed, 209 insertions(+), 8 deletions(-)\n create mode 100755 t/t6038-merge-text-auto.sh\n\n-- \n1.7.1.575.g383de\n"},{"id":"144711","messageId":"42dab0b23c001f3981180cc9e0069c4dbb49ae4a.1278093311.git.eyvind.bernhardsen@gmail.com","threadId":"24269","inReplyTo":"cover.1278093311.git.eyvind.bernhardsen@gmail.com","subject":"[PATCH v6 1/3] Avoid conflicts when merging branches with mixed normalization","fromName":"Eyvind Bernhardsen","fromEmail":"eyvind.bernhardsen@gmail.com","sentAt":"2010-07-02T19:20:47Z","receivedAt":"2010-07-02T19:20:47Z","isPatch":true,"sender":{"key":"eyvind.bernhardsen@gmail.com","avatar":"https://avatars.githubusercontent.com/u/106762?v=4"},"body":"Currently, merging across changes in line ending normalization is\npainful since files containing CRLF will conflict with normalized files,\neven if the only difference between the two versions is the line\nendings. Additionally, any \"real\" merge conflicts that exist are\nobscured because every line in the file has a conflict.\n\nAssume you start out with a repo that has a lot of text files with CRLF\nchecked in (A):\n\n o---C\n / \\\n A---B---D\n\nB: Add \"* text=auto\" to .gitattributes and normalize all files to\n LF-only\n\nC: Modify some of the text files\n\nD: Try to merge C\n\nYou will get a ridiculous number of LF/CRLF conflicts when trying to\nmerge C into D, since the repository contents for C are \"wrong\" wrt the\nnew .gitattributes file.\n\nFix ll-merge so that the \"base\", \"theirs\" and \"ours\" stages are passed\nthrough convert_to_worktree() and convert_to_git() before a three-way\nmerge. This ensures that all three stages are normalized in the same\nway, removing from consideration differences that are only due to\nnormalization.\n\nThis feature is optional for now since it changes a low-level mechanism\nand is not necessary for the majority of users. The \"merge.renormalize\"\nconfig variable enables it.\n\nSigned-off-by: Junio C Hamano \nSigned-off-by: Eyvind Bernhardsen \n---\n Documentation/gitattributes.txt | 34 ++++++++++++++++++++\n Documentation/merge-config.txt | 10 ++++++\n builtin/merge.c | 3 ++\n cache.h | 2 +\n convert.c | 16 ++++++++-\n environment.c | 1 +\n ll-merge.c | 15 +++++++++\n t/t6038-merge-text-auto.sh | 64 +++++++++++++++++++++++++++++++++++++++\n 8 files changed, 143 insertions(+), 2 deletions(-)\n create mode 100755 t/t6038-merge-text-auto.sh\n\ndiff --git a/Documentation/gitattributes.txt b/Documentation/gitattributes.txt\nindex 564586b..da553ff 100644\n--- a/Documentation/gitattributes.txt\n+++ b/Documentation/gitattributes.txt\n@@ -317,6 +317,17 @@ command is \"cat\").\n \tsmudge = cat\n ------------------------\n \n+For best results, `clean` should not alter its output further if it is\n+run twice (\"clean->clean\" should be equivalent to \"clean\"), and\n+multiple `smudge` commands should not alter `clean`'s output\n+(\"smudge->smudge->clean\" should be equivalent to \"clean\"). See the\n+section on merging below.\n+\n+The \"indent\" filter is well-behaved in this regard: it will not modify\n+input that is already correctly indented. In this case, the lack of a\n+smudge filter means that the clean filter _must_ accept its own output\n+without modifying it.\n+\n \n Interaction between checkin/checkout attributes\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n@@ -331,6 +342,29 @@ In the check-out codepath, the blob content is first converted\n with `text`, and then `ident` and fed to `filter`.\n \n \n+Merging branches with differing checkin/checkout attributes\n+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n+\n+If you have added attributes to a file that cause the canonical\n+repository format for that file to change, such as adding a\n+clean/smudge filter or text/eol/ident attributes, merging anything\n+where the attribute is not in place would normally cause merge\n+conflicts.\n+\n+To prevent these unnecessary merge conflicts, git can be told to run a\n+virtual check-out and check-in of all three stages of a file when\n+resolving a three-way merge by setting the `merge.renormalize`\n+configuration variable. This prevents changes caused by check-in\n+conversion from causing spurious merge conflicts when a converted file\n+is merged with an unconverted file.\n+\n+As long as a \"smudge->clean\" results in the same output as a \"clean\"\n+even on files that are already smudged, this strategy will\n+automatically resolve all filter-related conflicts. Filters that do\n+not act in this way may cause additional merge conflicts that must be\n+resolved manually.\n+\n+\n Generating diff text\n ~~~~~~~~~~~~~~~~~~~~\n \ndiff --git a/Documentation/merge-config.txt b/Documentation/merge-config.txt\nindex a403155..b72f533 100644\n--- a/Documentation/merge-config.txt\n+++ b/Documentation/merge-config.txt\n@@ -15,6 +15,16 @@ merge.renameLimit::\n \tduring a merge; if not specified, defaults to the value of\n \tdiff.renameLimit.\n \n+merge.renormalize::\n+\tTell git that canonical representation of files in the\n+\trepository has changed over time (e.g. earlier commits record\n+\ttext files with CRLF line endings, but recent ones use LF line\n+\tendings). In such a repository, git can convert the data\n+\trecorded in commits to a canonical form before performing a\n+\tmerge to reduce unnecessary conflicts. For more information,\n+\tsee section \"Merging branches with differing checkin/checkout\n+\tattributes\" in linkgit:gitattributes[5].\n+\n merge.stat::\n \tWhether to print the diffstat between ORIG_HEAD and the merge result\n \tat the end of the merge. True by default.\ndiff --git a/builtin/merge.c b/builtin/merge.c\nindex 37ce4f5..b836e9c 100644\n--- a/builtin/merge.c\n+++ b/builtin/merge.c\n@@ -503,6 +503,9 @@ static int git_merge_config(const char *k, const char *v, void *cb)\n \t\treturn git_config_string(&pull_octopus, k, v);\n \telse if (!strcmp(k, \"merge.log\") || !strcmp(k, \"merge.summary\"))\n \t\toption_log = git_config_bool(k, v);\n+\telse if (!strcmp(k, \"merge.renormalize\")) {\n+\t\tmerge_renormalize = git_config_bool(k, v);\n+\t}\n \treturn git_diff_ui_config(k, v, cb);\n }\n \ndiff --git a/cache.h b/cache.h\nindex c9fa3df..ed73da8 100644\n--- a/cache.h\n+++ b/cache.h\n@@ -551,6 +551,7 @@ extern int read_replace_refs;\n extern int fsync_object_files;\n extern int core_preload_index;\n extern int core_apply_sparse_checkout;\n+extern int merge_renormalize;\n \n enum safe_crlf {\n \tSAFE_CRLF_FALSE = 0,\n@@ -1054,6 +1055,7 @@ extern void trace_argv_printf(const char **argv, const char *format, ...);\n extern int convert_to_git(const char *path, const char *src, size_t len,\n struct strbuf *dst, enum safe_crlf checksafe);\n extern int convert_to_working_tree(const char *path, const char *src, size_t len, struct strbuf *dst);\n+extern int renormalize_buffer(const char *path, const char *src, size_t len, struct strbuf *dst);\n \n /* add */\n /*\ndiff --git a/convert.c b/convert.c\nindex e41a31e..0203be8 100644\n--- a/convert.c\n+++ b/convert.c\n@@ -93,7 +93,8 @@ static int is_binary(unsigned long size, struct text_stat *stats)\n \treturn 0;\n }\n \n-static enum eol determine_output_conversion(enum action action) {\n+static enum eol determine_output_conversion(enum action action)\n+{\n \tswitch (action) {\n \tcase CRLF_BINARY:\n \t\treturn EOL_UNSET;\n@@ -693,7 +694,8 @@ static int git_path_check_ident(const char *path, struct git_attr_check *check)\n \treturn !!ATTR_TRUE(value);\n }\n \n-enum action determine_action(enum action text_attr, enum eol eol_attr) {\n+static enum action determine_action(enum action text_attr, enum eol eol_attr)\n+{\n \tif (text_attr == CRLF_BINARY)\n \t\treturn CRLF_BINARY;\n \tif (eol_attr == EOL_LF)\n@@ -773,3 +775,13 @@ int convert_to_working_tree(const char *path, const char *src, size_t len, struc\n \t}\n \treturn ret | apply_filter(path, src, len, dst, filter);\n }\n+\n+int renormalize_buffer(const char *path, const char *src, size_t len, struct strbuf *dst)\n+{\n+\tint ret = convert_to_working_tree(path, src, len, dst);\n+\tif (ret) {\n+\t\tsrc = dst->buf;\n+\t\tlen = dst->len;\n+\t}\n+\treturn ret | convert_to_git(path, src, len, dst, 0);\n+}\ndiff --git a/environment.c b/environment.c\nindex 83d38d3..81a3682 100644\n--- a/environment.c\n+++ b/environment.c\n@@ -53,6 +53,7 @@ enum object_creation_mode object_creation_mode = OBJECT_CREATION_MODE;\n char *notes_ref_name;\n int grafts_replace_parents = 1;\n int core_apply_sparse_checkout;\n+int merge_renormalize;\n \n /* Parallel index stat data preload? */\n int core_preload_index = 0;\ndiff --git a/ll-merge.c b/ll-merge.c\nindex 3764a1a..5068fe0 100644\n--- a/ll-merge.c\n+++ b/ll-merge.c\n@@ -321,6 +321,16 @@ static int git_path_check_merge(const char *path, struct git_attr_check check[2]\n \treturn git_checkattr(path, 2, check);\n }\n \n+static void normalize_file(mmfile_t *mm, const char *path)\n+{\n+\tstruct strbuf strbuf = STRBUF_INIT;\n+\tif (renormalize_buffer(path, mm->ptr, mm->size, &strbuf)) {\n+\t\tfree(mm->ptr);\n+\t\tmm->size = strbuf.len;\n+\t\tmm->ptr = strbuf_detach(&strbuf, NULL);\n+\t}\n+}\n+\n int ll_merge(mmbuffer_t *result_buf,\n \t const char *path,\n \t mmfile_t *ancestor, const char *ancestor_label,\n@@ -334,6 +344,11 @@ int ll_merge(mmbuffer_t *result_buf,\n \tconst struct ll_merge_driver *driver;\n \tint virtual_ancestor = flag & 01;\n \n+\tif (merge_renormalize) {\n+\t\tnormalize_file(ancestor, path);\n+\t\tnormalize_file(ours, path);\n+\t\tnormalize_file(theirs, path);\n+\t}\n \tif (!git_path_check_merge(path, check)) {\n \t\tll_driver_name = check[0].value;\n \t\tif (check[1].value) {\ndiff --git a/t/t6038-merge-text-auto.sh b/t/t6038-merge-text-auto.sh\nnew file mode 100755\nindex 0000000..127baf8\n--- /dev/null\n+++ b/t/t6038-merge-text-auto.sh\n@@ -0,0 +1,64 @@\n+#!/bin/sh\n+\n+test_description='CRLF merge conflict across text=auto change'\n+\n+. ./test-lib.sh\n+\n+test_expect_success setup '\n+\tgit config merge.renormalize true &&\n+\tgit config core.autocrlf false &&\n+\techo first line | append_cr >file &&\n+\techo first line >control_file &&\n+\techo only line >inert_file &&\n+\tgit add file control_file inert_file &&\n+\tgit commit -m \"Initial\" &&\n+\tgit tag initial &&\n+\tgit branch side &&\n+\techo \"* text=auto\" >.gitattributes &&\n+\ttouch file &&\n+\tgit add .gitattributes file &&\n+\tgit commit -m \"normalize file\" &&\n+\techo same line | append_cr >>file &&\n+\techo same line >>control_file &&\n+\tgit add file control_file &&\n+\tgit commit -m \"add line from a\" &&\n+\tgit tag a &&\n+\tgit rm .gitattributes &&\n+\trm file &&\n+\tgit checkout file &&\n+\tgit commit -m \"remove .gitattributes\" &&\n+\tgit tag c &&\n+\tgit checkout side &&\n+\techo same line | append_cr >>file &&\n+\techo same line >>control_file &&\n+\tgit add file control_file &&\n+\tgit commit -m \"add line from b\" &&\n+\tgit tag b &&\n+\tgit checkout master\n+'\n+\n+test_expect_success 'Check merging after setting text=auto' '\n+\tgit reset --hard a &&\n+\tgit merge b &&\n+\tcat file | remove_cr >file.temp &&\n+\ttest_cmp file file.temp\n+'\n+\n+test_expect_success 'Check merging addition of text=auto' '\n+\tgit reset --hard b &&\n+\tgit merge a &&\n+\tcat file | remove_cr >file.temp &&\n+\ttest_cmp file file.temp\n+'\n+\n+test_expect_failure 'Test delete/normalize conflict' '\n+\tgit checkout side &&\n+\tgit reset --hard initial &&\n+\tgit rm file &&\n+\tgit commit -m \"remove file\" &&\n+\tgit checkout master &&\n+\tgit reset --hard a^ &&\n+\tgit merge side\n+'\n+\n+test_done\n-- \n1.7.1.575.g383de\n"},{"id":"144707","messageId":"0652be563a15e811d053d59532eff62ca5684203.1278093311.git.eyvind.bernhardsen@gmail.com","threadId":"24269","inReplyTo":"cover.1278093311.git.eyvind.bernhardsen@gmail.com","subject":"[PATCH v6 2/3] Try normalizing files to avoid delete/modify conflicts when merging","fromName":"Eyvind Bernhardsen","fromEmail":"eyvind.bernhardsen@gmail.com","sentAt":"2010-07-02T19:20:48Z","receivedAt":"2010-07-02T19:20:48Z","isPatch":true,"sender":{"key":"eyvind.bernhardsen@gmail.com","avatar":"https://avatars.githubusercontent.com/u/106762?v=4"},"body":"If a file is modified due to normalization on one branch, and deleted on\nanother, a merge of the two branches will result in a delete/modify\nconflict for that file even if it is otherwise unchanged.\n\nTry to avoid the conflict by normalizing and comparing the \"base\" file\nand the modified file when their sha1s differ. If they compare equal,\nthe file is considered unmodified and is deleted.\n\nSigned-off-by: Eyvind Bernhardsen \nSigned-off-by: Junio C Hamano \n---\n merge-recursive.c | 51 ++++++++++++++++++++++++++++++++++++++++++-\n t/t6038-merge-text-auto.sh | 2 +-\n 2 files changed, 50 insertions(+), 3 deletions(-)\n\ndiff --git a/merge-recursive.c b/merge-recursive.c\nindex 856e98c..593c0b1 100644\n--- a/merge-recursive.c\n+++ b/merge-recursive.c\n@@ -1056,6 +1056,53 @@ static unsigned char *stage_sha(const unsigned char *sha, unsigned mode)\n \treturn (is_null_sha1(sha) || mode == 0) ? NULL: (unsigned char *)sha;\n }\n \n+static int read_sha1_strbuf(const unsigned char *sha1, struct strbuf *dst)\n+{\n+\tvoid *buf;\n+\tenum object_type type;\n+\tunsigned long size;\n+\tbuf = read_sha1_file(sha1, &type, &size);\n+\tif (!buf)\n+\t\treturn error(\"cannot read object %s\", sha1_to_hex(sha1));\n+\tif (type != OBJ_BLOB) {\n+\t\tfree(buf);\n+\t\treturn error(\"object %s is not a blob\", sha1_to_hex(sha1));\n+\t}\n+\tstrbuf_attach(dst, buf, size, size + 1);\n+\treturn 0;\n+}\n+\n+static int blob_unchanged(const unsigned char *o_sha,\n+\t\t\t const unsigned char *a_sha,\n+\t\t\t const char *path)\n+{\n+\tstruct strbuf o = STRBUF_INIT;\n+\tstruct strbuf a = STRBUF_INIT;\n+\tint ret = 0; /* assume changed for safety */\n+\n+\tif (sha_eq(o_sha, a_sha))\n+\t\treturn 1;\n+\tif (!merge_renormalize)\n+\t\treturn 0;\n+\n+\tassert(o_sha && a_sha);\n+\tif (read_sha1_strbuf(o_sha, &o) || read_sha1_strbuf(a_sha, &a))\n+\t\tgoto error_return;\n+\t/*\n+\t * Note: binary | is used so that both renormalizations are\n+\t * performed. Comparison can be skipped if both files are\n+\t * unchanged since their sha1s have already been compared.\n+\t */\n+\tif (renormalize_buffer(path, o.buf, o.len, &o) |\n+\t renormalize_buffer(path, a.buf, o.len, &a))\n+\t\tret = (o.len == a.len && !memcmp(o.buf, a.buf, o.len));\n+\n+error_return:\n+\tstrbuf_release(&o);\n+\tstrbuf_release(&a);\n+\treturn ret;\n+}\n+\n /* Per entry merge function */\n static int process_entry(struct merge_options *o,\n \t\t\t const char *path, struct stage_data *entry)\n@@ -1075,8 +1122,8 @@ static int process_entry(struct merge_options *o,\n \tif (o_sha && (!a_sha || !b_sha)) {\n \t\t/* Case A: Deleted in one */\n \t\tif ((!a_sha && !b_sha) ||\n-\t\t (sha_eq(a_sha, o_sha) && !b_sha) ||\n-\t\t (!a_sha && sha_eq(b_sha, o_sha))) {\n+\t\t (!b_sha && blob_unchanged(o_sha, a_sha, path)) ||\n+\t\t (!a_sha && blob_unchanged(o_sha, b_sha, path))) {\n \t\t\t/* Deleted in both or deleted in one and\n \t\t\t * unchanged in the other */\n \t\t\tif (a_sha)\ndiff --git a/t/t6038-merge-text-auto.sh b/t/t6038-merge-text-auto.sh\nindex 127baf8..d1ab86e 100755\n--- a/t/t6038-merge-text-auto.sh\n+++ b/t/t6038-merge-text-auto.sh\n@@ -51,7 +51,7 @@ test_expect_success 'Check merging addition of text=auto' '\n \ttest_cmp file file.temp\n '\n \n-test_expect_failure 'Test delete/normalize conflict' '\n+test_expect_success 'Test delete/normalize conflict' '\n \tgit checkout side &&\n \tgit reset --hard initial &&\n \tgit rm file &&\n-- \n1.7.1.575.g383de\n"},{"id":"144706","messageId":"6a8530fa3cba7cbc14882a9f5979d13452ab52b5.1278093311.git.eyvind.bernhardsen@gmail.com","threadId":"24269","inReplyTo":"cover.1278093311.git.eyvind.bernhardsen@gmail.com","subject":"[PATCH v6 3/3] Don't expand CRLFs when normalizing text during merge","fromName":"Eyvind Bernhardsen","fromEmail":"eyvind.bernhardsen@gmail.com","sentAt":"2010-07-02T19:20:49Z","receivedAt":"2010-07-02T19:20:49Z","isPatch":true,"sender":{"key":"eyvind.bernhardsen@gmail.com","avatar":"https://avatars.githubusercontent.com/u/106762?v=4"},"body":"Disable CRLF expansion when convert_to_working_tree() is called from\nnormalize_buffer(). This improves performance when merging branches\nwith conflicting line endings when core.eol=crlf or core.autocrlf=true\nby making the normalization act as if core.eol=lf.\n\nSigned-off-by: Eyvind Bernhardsen \nSigned-off-by: Junio C Hamano \n---\n convert.c | 27 ++++++++++++++++++++-------\n 1 files changed, 20 insertions(+), 7 deletions(-)\n\ndiff --git a/convert.c b/convert.c\nindex 0203be8..01de9a8 100644\n--- a/convert.c\n+++ b/convert.c\n@@ -741,7 +741,9 @@ int convert_to_git(const char *path, const char *src, size_t len,\n \treturn ret | ident_to_git(path, src, len, dst, ident);\n }\n \n-int convert_to_working_tree(const char *path, const char *src, size_t len, struct strbuf *dst)\n+static int convert_to_working_tree_internal(const char *path, const char *src,\n+\t\t\t\t\t size_t len, struct strbuf *dst,\n+\t\t\t\t\t int normalizing)\n {\n \tstruct git_attr_check check[5];\n \tenum action action = CRLF_GUESS;\n@@ -767,18 +769,29 @@ int convert_to_working_tree(const char *path, const char *src, size_t len, struc\n \t\tsrc = dst->buf;\n \t\tlen = dst->len;\n \t}\n-\taction = determine_action(action, eol_attr);\n-\tret |= crlf_to_worktree(path, src, len, dst, action);\n-\tif (ret) {\n-\t\tsrc = dst->buf;\n-\t\tlen = dst->len;\n+\t/*\n+\t * CRLF conversion can be skipped if normalizing, unless there\n+\t * is a smudge filter. The filter might expect CRLFs.\n+\t */\n+\tif (filter || !normalizing) {\n+\t\taction = determine_action(action, eol_attr);\n+\t\tret |= crlf_to_worktree(path, src, len, dst, action);\n+\t\tif (ret) {\n+\t\t\tsrc = dst->buf;\n+\t\t\tlen = dst->len;\n+\t\t}\n \t}\n \treturn ret | apply_filter(path, src, len, dst, filter);\n }\n \n+int convert_to_working_tree(const char *path, const char *src, size_t len, struct strbuf *dst)\n+{\n+\treturn convert_to_working_tree_internal(path, src, len, dst, 0);\n+}\n+\n int renormalize_buffer(const char *path, const char *src, size_t len, struct strbuf *dst)\n {\n-\tint ret = convert_to_working_tree(path, src, len, dst);\n+\tint ret = convert_to_working_tree_internal(path, src, len, dst, 1);\n \tif (ret) {\n \t\tsrc = dst->buf;\n \t\tlen = dst->len;\n-- \n1.7.1.575.g383de\n"},{"id":"144727","messageId":"7vk4pdzema.fsf@alter.siamese.dyndns.org","threadId":"24269","inReplyTo":"cover.1278093311.git.eyvind.bernhardsen@gmail.com","subject":"Re: [PATCH v6 0/3] Merge renormalization, config renamed","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2010-07-02T22:46:21Z","receivedAt":"2010-07-02T22:46:21Z","isPatch":true,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Eyvind Bernhardsen writes:\n\n> Here's a new version of the merge normalization series that renames the\n> configuration variable. Since \"merge.renormalize\" got the best\n> response, I went with that.\n>\n> Junio, I hope you don't mind that I squashed your patch to introduce the\n> config variable into the first patch. It seemed a bit out-of-place as a\n> separate commit.\n\nIt was meant to be squashed into the first one. Thanks.\n"},{"id":"147087","messageId":"20100804031935.GA19699@burratino","threadId":"24269","inReplyTo":"cover.1278093311.git.eyvind.bernhardsen@gmail.com","subject":"[PATCH/RFC eb/double-convert-before-merge 0/6] merge -Xrenormalize","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2010-08-04T03:19:35Z","receivedAt":"2010-08-04T03:19:35Z","isPatch":true,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"Hi Eyvind,\n\nEyvind Bernhardsen wrote:\n\n> Here's a new version of the merge normalization series that renames the\n> configuration variable. Since \"merge.renormalize\" got the best\n> response, I went with that.\n\nI have no idea how the following branch happened in my local git tree,\nbut it’s here now so maybe it’s worth something. Somehow I got the\nidea that the merge_renormalize variable was a nice short-term\nprotection but long-term scary, and so a few patches emerged to\nbanish it.\n\nPlease let me emphasize that in their current form I do not think\nthese patches should be very usable.\n\nBut for later, they tell a nice story. The idea is that the\n\"merge.renormalize\" is not really about the behavior of the \"git\nmerge\" command but about the merge-recursive driver; and so it\nshould be usable everywhere else that driver is used, too.\n\nSo in an ideal world, you would be able to, e.g.:\n\n\tgit checkout -m -Xrenormalize otherbranch\n\nor\n\n\tgit revert -Xrenormalize otherpatch\n\nor\n\n\tgit pull --rebase -Xrenormalize\n\nOf course, the plumbing is not all there yet; many commands that\nuse merge drivers cannot pass through arbitrary options through\nyet.\n\nWell, enough talking. Here are the patches; what do you think?\n\nPatch 1 teaches blob_unchanged() to stop paying attention to the\nmerge_recursive variable. It uses a parameter instead. So from then\non, it can watch, bemused, a neutral party.\n\nPatch 2 adds and respects a renormalize option in the merge_options\nstruct. It defaults to the value of the merge_recursive global to\nsave callers (in particular, \"git merge\") the pain of adjusting.\n\nPatch 3 lets ll_merge merge callers decide whether to renormalize,\ntoo. Most callers never renormalize, since they are not \"git merge\"\n(and this is noted in new comments).\n\nPatch 4 is a sort of a digression. It makes \"git rerere\" produce nice\nhelp output with the \"git rerere -h\" option.\n\nPatch 5 lets rerere callers decide whether to renormalize, too. Most\ncallers never renormalize. But looking at that patch reveals a\nserious problem: there are all kinds of other merge options (e.g.,\n-Xsubtree) that rerere callers are unable to control.\n\nFollowing the principle that scripted users should have about as\nmuch power as internal ones, patch 5 reluctantly exposes the\nrenormalize option to rerere as a new --renormalize option. It\nshould be -Xrenormalize instead.\n\nPatch 6 eliminates the merge_recursive global once and for all, by\nteaching merge_recursive callers to set the renormalize merge option\nappropriately (and removing the backward-compatibility default).\nCallers that are not \"git merge\" still never use renormalize.\n\nFollowing the principle that scripted users should have as much\npower as internal ones, it also exposes the nice interface of an\n-Xrenormalize option. Some redundancy between builtin/merge and\nbuiltin/merge-recursive is noticed but left for another topic.\n\nJonathan Nieder (6):\n merge-trees: push choice to renormalize away from low level\n merge-trees: let caller decide whether to renormalize\n ll-merge: let caller decide whether to renormalize\n rerere: migrate to parse-options API\n rerere: let caller decide whether to renormalize\n merge-recursive: add -Xrenormalize option\n\n Documentation/merge-strategies.txt | 8 +++++\n builtin/checkout.c | 13 ++++++++-\n builtin/commit.c | 4 ++\n builtin/merge-recursive.c | 2 +\n builtin/merge.c | 24 +++++++++++----\n builtin/rerere.c | 56 ++++++++++++++++++++---------------\n builtin/revert.c | 18 ++++++++++-\n cache.h | 1 -\n environment.c | 1 -\n ll-merge.c | 4 +-\n ll-merge.h | 2 +-\n merge-file.c | 2 +-\n merge-recursive.c | 11 ++++--\n merge-recursive.h | 1 +\n rerere.c | 20 ++++++++----\n rerere.h | 1 +\n 16 files changed, 117 insertions(+), 51 deletions(-)\n\n-- \n1.7.2.1.544.ga752d.dirty\n"},{"id":"147088","messageId":"20100804032043.GB19699@burratino","threadId":"24269","inReplyTo":"20100804031935.GA19699@burratino","subject":"[PATCH 1/6] merge-trees: push choice to renormalize away from low level","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2010-08-04T03:20:43Z","receivedAt":"2010-08-04T03:20:43Z","isPatch":true,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"The merge machinery decides whether to resmudge and clean relevant\nentries based on the global merge_renormalize setting, which is set by\n\"git merge\" based on its configuration (and left alone by other\ncommands).\n\nA nicer interface would make that decision a parameter to merge_trees\nso callers would pass in a choice made on a call-by-call basis.\nStart by making blob_unchanged stop examining the merge_renormalize\nglobal.\n\nIn other words, this change is a trivial no-op, but it brings us\ncloser to something good.\n\nCc: Eyvind Bernhardsen \nSigned-off-by: Jonathan Nieder \n---\n merge-recursive.c | 9 +++++----\n 1 files changed, 5 insertions(+), 4 deletions(-)\n\ndiff --git a/merge-recursive.c b/merge-recursive.c\nindex 5ad8fc9..2b55fc2 100644\n--- a/merge-recursive.c\n+++ b/merge-recursive.c\n@@ -1074,7 +1074,7 @@ static int read_sha1_strbuf(const unsigned char *sha1, struct strbuf *dst)\n \n static int blob_unchanged(const unsigned char *o_sha,\n \t\t\t const unsigned char *a_sha,\n-\t\t\t const char *path)\n+\t\t\t int renormalize, const char *path)\n {\n \tstruct strbuf o = STRBUF_INIT;\n \tstruct strbuf a = STRBUF_INIT;\n@@ -1082,7 +1082,7 @@ static int blob_unchanged(const unsigned char *o_sha,\n \n \tif (sha_eq(o_sha, a_sha))\n \t\treturn 1;\n-\tif (!merge_renormalize)\n+\tif (!renormalize)\n \t\treturn 0;\n \n \tassert(o_sha && a_sha);\n@@ -1112,6 +1112,7 @@ static int process_entry(struct merge_options *o,\n \tprint_index_entry(\"\\tpath: \", entry);\n \t*/\n \tint clean_merge = 1;\n+\tint normalize = merge_renormalize;\n \tunsigned o_mode = entry->stages[1].mode;\n \tunsigned a_mode = entry->stages[2].mode;\n \tunsigned b_mode = entry->stages[3].mode;\n@@ -1122,8 +1123,8 @@ static int process_entry(struct merge_options *o,\n \tif (o_sha && (!a_sha || !b_sha)) {\n \t\t/* Case A: Deleted in one */\n \t\tif ((!a_sha && !b_sha) ||\n-\t\t (!b_sha && blob_unchanged(o_sha, a_sha, path)) ||\n-\t\t (!a_sha && blob_unchanged(o_sha, b_sha, path))) {\n+\t\t (!b_sha && blob_unchanged(o_sha, a_sha, normalize, path)) ||\n+\t\t (!a_sha && blob_unchanged(o_sha, b_sha, normalize, path))) {\n \t\t\t/* Deleted in both or deleted in one and\n \t\t\t * unchanged in the other */\n \t\t\tif (a_sha)\n-- \n1.7.2.1.544.ga752d.dirty\n"},{"id":"147089","messageId":"20100804032110.GC19699@burratino","threadId":"24269","inReplyTo":"20100804031935.GA19699@burratino","subject":"[PATCH 2/6] merge-trees: let caller decide whether to renormalize","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2010-08-04T03:21:10Z","receivedAt":"2010-08-04T03:21:10Z","isPatch":true,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"Add a \"renormalize\" option to struct merge_options so callers can\ndecide on a case-by-case basis whether the merge is likely to have\noverlapped with a change in smudge/clean rules. The option defaults\nto the global merge_renormalize setting for now.\n\nNo change in behavior intended.\n\nCc: Eyvind Bernhardsen \nSigned-off-by: Jonathan Nieder \n---\n merge-recursive.c | 3 ++-\n merge-recursive.h | 1 +\n 2 files changed, 3 insertions(+), 1 deletions(-)\n\ndiff --git a/merge-recursive.c b/merge-recursive.c\nindex 2b55fc2..8a49844 100644\n--- a/merge-recursive.c\n+++ b/merge-recursive.c\n@@ -1112,7 +1112,7 @@ static int process_entry(struct merge_options *o,\n \tprint_index_entry(\"\\tpath: \", entry);\n \t*/\n \tint clean_merge = 1;\n-\tint normalize = merge_renormalize;\n+\tint normalize = o->renormalize;\n \tunsigned o_mode = entry->stages[1].mode;\n \tunsigned a_mode = entry->stages[2].mode;\n \tunsigned b_mode = entry->stages[3].mode;\n@@ -1484,6 +1484,7 @@ void init_merge_options(struct merge_options *o)\n \to->buffer_output = 1;\n \to->diff_rename_limit = -1;\n \to->merge_rename_limit = -1;\n+\to->renormalize = merge_renormalize;\n \tgit_config(merge_recursive_config, o);\n \tif (getenv(\"GIT_MERGE_VERBOSITY\"))\n \t\to->verbosity =\ndiff --git a/merge-recursive.h b/merge-recursive.h\nindex b831293..cdb97e9 100644\n--- a/merge-recursive.h\n+++ b/merge-recursive.h\n@@ -17,6 +17,7 @@ struct merge_options {\n \tint verbosity;\n \tint diff_rename_limit;\n \tint merge_rename_limit;\n+\tint renormalize;\n \tint call_depth;\n \tstruct strbuf obuf;\n \tstruct string_list current_file_set;\n-- \n1.7.2.1.544.ga752d.dirty\n"},{"id":"147090","messageId":"20100804032144.GD19699@burratino","threadId":"24269","inReplyTo":"20100804031935.GA19699@burratino","subject":"[PATCH 3/6] ll-merge: let caller decide whether to renormalize","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2010-08-04T03:21:44Z","receivedAt":"2010-08-04T03:21:44Z","isPatch":true,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"Add a \"renormalize\" parameter to ll_merge so callers can decide on a\ncase-by-case basis whether the merge is likely to have overlapped with\na change in smudge/clean rules. This reveals a few commands that have\nnot been taking that situation into account, though it does not fix\nthem.\n\nNo change in behavior intended.\n\nCc: Eyvind Bernhardsen \nSigned-off-by: Jonathan Nieder \n---\n builtin/checkout.c | 6 +++++-\n ll-merge.c | 4 ++--\n ll-merge.h | 2 +-\n merge-file.c | 2 +-\n merge-recursive.c | 1 +\n rerere.c | 17 +++++++++++------\n 6 files changed, 21 insertions(+), 11 deletions(-)\n\ndiff --git a/builtin/checkout.c b/builtin/checkout.c\nindex 1994be9..5a7ae03 100644\n--- a/builtin/checkout.c\n+++ b/builtin/checkout.c\n@@ -150,8 +150,12 @@ static int checkout_merged(int pos, struct checkout *state)\n \tread_mmblob(&ours, active_cache[pos+1]->sha1);\n \tread_mmblob(&theirs, active_cache[pos+2]->sha1);\n \n+\t/*\n+\t * NEEDSWORK: re-create conflicts from merges with\n+\t * merge.renormalize set, too\n+\t */\n \tstatus = ll_merge(&result_buf, path, &ancestor, \"base\",\n-\t\t\t &ours, \"ours\", &theirs, \"theirs\", 0);\n+\t\t\t &ours, \"ours\", &theirs, \"theirs\", 0, 0);\n \tfree(ancestor.ptr);\n \tfree(ours.ptr);\n \tfree(theirs.ptr);\ndiff --git a/ll-merge.c b/ll-merge.c\nindex 5068fe0..0a0e04d 100644\n--- a/ll-merge.c\n+++ b/ll-merge.c\n@@ -336,7 +336,7 @@ int ll_merge(mmbuffer_t *result_buf,\n \t mmfile_t *ancestor, const char *ancestor_label,\n \t mmfile_t *ours, const char *our_label,\n \t mmfile_t *theirs, const char *their_label,\n-\t int flag)\n+\t int renormalize, int flag)\n {\n \tstatic struct git_attr_check check[2];\n \tconst char *ll_driver_name = NULL;\n@@ -344,7 +344,7 @@ int ll_merge(mmbuffer_t *result_buf,\n \tconst struct ll_merge_driver *driver;\n \tint virtual_ancestor = flag & 01;\n \n-\tif (merge_renormalize) {\n+\tif (renormalize) {\n \t\tnormalize_file(ancestor, path);\n \t\tnormalize_file(ours, path);\n \t\tnormalize_file(theirs, path);\ndiff --git a/ll-merge.h b/ll-merge.h\nindex 57754cc..dd81a1e 100644\n--- a/ll-merge.h\n+++ b/ll-merge.h\n@@ -10,7 +10,7 @@ int ll_merge(mmbuffer_t *result_buf,\n \t mmfile_t *ancestor, const char *ancestor_label,\n \t mmfile_t *ours, const char *our_label,\n \t mmfile_t *theirs, const char *their_label,\n-\t int flag);\n+\t int renormalize, int flag);\n \n int ll_merge_marker_size(const char *path);\n \ndiff --git a/merge-file.c b/merge-file.c\nindex db4d0d5..5afdc01 100644\n--- a/merge-file.c\n+++ b/merge-file.c\n@@ -37,7 +37,7 @@ static void *three_way_filemerge(const char *path, mmfile_t *base, mmfile_t *our\n \t * common ancestor.\n \t */\n \tmerge_status = ll_merge(&res, path, base, NULL,\n-\t\t\t\tour, \".our\", their, \".their\", 0);\n+\t\t\t\tour, \".our\", their, \".their\", 0, 0);\n \tif (merge_status < 0)\n \t\treturn NULL;\n \ndiff --git a/merge-recursive.c b/merge-recursive.c\nindex 8a49844..4838939 100644\n--- a/merge-recursive.c\n+++ b/merge-recursive.c\n@@ -647,6 +647,7 @@ static int merge_3way(struct merge_options *o,\n \n \tmerge_status = ll_merge(result_buf, a->path, &orig, base_name,\n \t\t\t\t&src1, name1, &src2, name2,\n+\t\t\t\to->renormalize,\n \t\t\t\t(!!o->call_depth) | (favor << 1));\n \n \tfree(name1);\ndiff --git a/rerere.c b/rerere.c\nindex 2197890..17dcc3c 100644\n--- a/rerere.c\n+++ b/rerere.c\n@@ -319,9 +319,13 @@ static int handle_cache(const char *path, unsigned char *sha1, const char *outpu\n \t\tif (!mmfile[i].ptr && !mmfile[i].size)\n \t\t\tmmfile[i].ptr = xstrdup(\"\");\n \t}\n+\t/*\n+\t * NEEDSWORK: handle conflicts from merges with\n+\t * merge.renormalize set, too\n+\t */\n \tll_merge(&result, path, &mmfile[0], NULL,\n \t\t &mmfile[1], \"ours\",\n-\t\t &mmfile[2], \"theirs\", 0);\n+\t\t &mmfile[2], \"theirs\", 0, 0);\n \tfor (i = 0; i < 3; i++)\n \t\tfree(mmfile[i].ptr);\n \n@@ -361,7 +365,7 @@ static int find_conflict(struct string_list *conflict)\n \treturn 0;\n }\n \n-static int merge(const char *name, const char *path)\n+static int merge(const char *name, int renormalize, const char *path)\n {\n \tint ret;\n \tmmfile_t cur = {NULL, 0}, base = {NULL, 0}, other = {NULL, 0};\n@@ -376,7 +380,8 @@ static int merge(const char *name, const char *path)\n \t\tret = 1;\n \t\tgoto out;\n \t}\n-\tret = ll_merge(&result, path, &base, NULL, &cur, \"\", &other, \"\", 0);\n+\tret = ll_merge(&result, path, &base, NULL,\n+\t\t\t&cur, \"\", &other, \"\", renormalize, 0);\n \tif (!ret) {\n \t\tFILE *f = fopen(path, \"w\");\n \t\tif (!f)\n@@ -424,7 +429,7 @@ static int update_paths(struct string_list *update)\n \treturn status;\n }\n \n-static int do_plain_rerere(struct string_list *rr, int fd)\n+static int do_plain_rerere(struct string_list *rr, int fd, int renormalize)\n {\n \tstruct string_list conflict = { NULL, 0, 0, 1 };\n \tstruct string_list update = { NULL, 0, 0, 1 };\n@@ -469,7 +474,7 @@ static int do_plain_rerere(struct string_list *rr, int fd)\n \t\tconst char *name = (const char *)rr->items[i].util;\n \n \t\tif (has_rerere_resolution(name)) {\n-\t\t\tif (!merge(name, path)) {\n+\t\t\tif (!merge(name, renormalize, path)) {\n \t\t\t\tif (rerere_autoupdate)\n \t\t\t\t\tstring_list_insert(path, &update);\n \t\t\t\tfprintf(stderr,\n@@ -553,7 +558,7 @@ int rerere(int flags)\n \tfd = setup_rerere(&merge_rr, flags);\n \tif (fd < 0)\n \t\treturn 0;\n-\treturn do_plain_rerere(&merge_rr, fd);\n+\treturn do_plain_rerere(&merge_rr, fd, merge_renormalize);\n }\n \n static int rerere_forget_one_path(const char *path, struct string_list *rr)\n-- \n1.7.2.1.544.ga752d.dirty\n"},{"id":"147091","messageId":"20100804032216.GE19699@burratino","threadId":"24269","inReplyTo":"20100804031935.GA19699@burratino","subject":"[PATCH 4/6] rerere: migrate to parse-options API","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2010-08-04T03:22:16Z","receivedAt":"2010-08-04T03:22:16Z","isPatch":true,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"Signed-off-by: Jonathan Nieder \n---\n builtin/rerere.c | 52 ++++++++++++++++++++++++++++------------------------\n 1 files changed, 28 insertions(+), 24 deletions(-)\n\ndiff --git a/builtin/rerere.c b/builtin/rerere.c\nindex 0048f9e..295fe75 100644\n--- a/builtin/rerere.c\n+++ b/builtin/rerere.c\n@@ -1,13 +1,16 @@\n #include \"builtin.h\"\n #include \"cache.h\"\n #include \"dir.h\"\n+#include \"parse-options.h\"\n #include \"string-list.h\"\n #include \"rerere.h\"\n #include \"xdiff/xdiff.h\"\n #include \"xdiff-interface.h\"\n \n-static const char git_rerere_usage[] =\n-\"git rerere [clear | status | diff | gc]\";\n+static const char * const rerere_usage[] = {\n+\t\"git rerere [clear | status | diff | gc]\",\n+\tNULL,\n+};\n \n /* these values are days */\n static int cutoff_noresolve = 15;\n@@ -103,25 +106,26 @@ static int diff_two(const char *file1, const char *label1,\n int cmd_rerere(int argc, const char **argv, const char *prefix)\n {\n \tstruct string_list merge_rr = { NULL, 0, 0, 1 };\n-\tint i, fd, flags = 0;\n-\n-\tif (2 < argc) {\n-\t\tif (!strcmp(argv[1], \"-h\"))\n-\t\t\tusage(git_rerere_usage);\n-\t\tif (!strcmp(argv[1], \"--rerere-autoupdate\"))\n-\t\t\tflags = RERERE_AUTOUPDATE;\n-\t\telse if (!strcmp(argv[1], \"--no-rerere-autoupdate\"))\n-\t\t\tflags = RERERE_NOAUTOUPDATE;\n-\t\tif (flags) {\n-\t\t\targc--;\n-\t\t\targv++;\n-\t\t}\n-\t}\n-\tif (argc < 2)\n+\tint i, fd, autoupdate = -1, flags = 0;\n+\n+\tstruct option options[] = {\n+\t\tOPT_SET_INT(0, \"rerere-autoupdate\", &autoupdate,\n+\t\t\t\"register clean resolutions in index\", 1),\n+\t\tOPT_END(),\n+\t};\n+\n+\targc = parse_options(argc, argv, prefix, options, rerere_usage, 0);\n+\n+\tif (autoupdate == 1)\n+\t\tflags = RERERE_AUTOUPDATE;\n+\tif (autoupdate == 0)\n+\t\tflags = RERERE_NOAUTOUPDATE;\n+\n+\tif (argc < 1)\n \t\treturn rerere(flags);\n \n-\tif (!strcmp(argv[1], \"forget\")) {\n-\t\tconst char **pathspec = get_pathspec(prefix, argv + 2);\n+\tif (!strcmp(argv[0], \"forget\")) {\n+\t\tconst char **pathspec = get_pathspec(prefix, argv + 1);\n \t\treturn rerere_forget(pathspec);\n \t}\n \n@@ -129,26 +133,26 @@ int cmd_rerere(int argc, const char **argv, const char *prefix)\n \tif (fd < 0)\n \t\treturn 0;\n \n-\tif (!strcmp(argv[1], \"clear\")) {\n+\tif (!strcmp(argv[0], \"clear\")) {\n \t\tfor (i = 0; i < merge_rr.nr; i++) {\n \t\t\tconst char *name = (const char *)merge_rr.items[i].util;\n \t\t\tif (!has_rerere_resolution(name))\n \t\t\t\tunlink_rr_item(name);\n \t\t}\n \t\tunlink_or_warn(git_path(\"rr-cache/MERGE_RR\"));\n-\t} else if (!strcmp(argv[1], \"gc\"))\n+\t} else if (!strcmp(argv[0], \"gc\"))\n \t\tgarbage_collect(&merge_rr);\n-\telse if (!strcmp(argv[1], \"status\"))\n+\telse if (!strcmp(argv[0], \"status\"))\n \t\tfor (i = 0; i < merge_rr.nr; i++)\n \t\t\tprintf(\"%s\\n\", merge_rr.items[i].string);\n-\telse if (!strcmp(argv[1], \"diff\"))\n+\telse if (!strcmp(argv[0], \"diff\"))\n \t\tfor (i = 0; i < merge_rr.nr; i++) {\n \t\t\tconst char *path = merge_rr.items[i].string;\n \t\t\tconst char *name = (const char *)merge_rr.items[i].util;\n \t\t\tdiff_two(rerere_path(name, \"preimage\"), path, path, path);\n \t\t}\n \telse\n-\t\tusage(git_rerere_usage);\n+\t\tusage_with_options(rerere_usage, options);\n \n \tstring_list_clear(&merge_rr, 1);\n \treturn 0;\n-- \n1.7.2.1.544.ga752d.dirty\n"},{"id":"147092","messageId":"20100804032338.GF19699@burratino","threadId":"24269","inReplyTo":"20100804031935.GA19699@burratino","subject":"[PATCH 5/6] rerere: let caller decide whether to renormalize","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2010-08-04T03:23:38Z","receivedAt":"2010-08-04T03:23:38Z","isPatch":true,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"Add a RERERE_RENORMALIZE flag to rerere so callers can decide\ncase-by-case whether the merge is likely to have overlapped with a\nchange in smudge/clean rules.\n\nThis is only a change in internal plumbing. Many callers do not have\na way to use that setting; this patch does not change that.\n\nNEEDSWORK: this is a step in the wrong direction. rerere needs\nan -s option to use an arbitrary merge strategy and a -X option to\npass arbitrary options to that driver. And maybe the options used\nin a merge should be recorded somewhere to help rerere repeat it.\n\nThis treats the renormalize option specially anyway, to support the\nexisting \"git merge\" behavior of using -Xrenormalize in its rerere\ncall when configured to do so. It reluctant exposes that option in\nthe rerere command as --renormalize for the sake of experimentation,\ntoo, but it does not advertise it.\n\nCc: Eyvind Bernhardsen \nSigned-off-by: Jonathan Nieder \n---\n builtin/commit.c | 4 ++++\n builtin/merge.c | 7 +++++--\n builtin/rerere.c | 8 ++++++--\n builtin/revert.c | 11 +++++++++--\n rerere.c | 5 +++--\n rerere.h | 1 +\n 6 files changed, 28 insertions(+), 8 deletions(-)\n\ndiff --git a/builtin/commit.c b/builtin/commit.c\nindex 3d99cf9..56c998f 100644\n--- a/builtin/commit.c\n+++ b/builtin/commit.c\n@@ -1377,6 +1377,10 @@ int cmd_commit(int argc, const char **argv, const char *prefix)\n \t\t \"new_index file. Check that disk is not full or quota is\\n\"\n \t\t \"not exceeded, and then \\\"git reset HEAD\\\" to recover.\");\n \n+\t/*\n+\t * NEEDSWORK: use RERERE_RENORMALIZE after a merge with\n+\t * merge.renormalize set\n+\t */\n \trerere(0);\n \trun_hook(get_index_file(), \"post-commit\", NULL);\n \tif (amend && !no_post_rewrite) {\ndiff --git a/builtin/merge.c b/builtin/merge.c\nindex b836e9c..823b76b 100644\n--- a/builtin/merge.c\n+++ b/builtin/merge.c\n@@ -822,7 +822,7 @@ static int finish_automerge(struct commit_list *common,\n static int suggest_conflicts(void)\n {\n \tFILE *fp;\n-\tint pos;\n+\tint pos, flag;\n \n \tfp = fopen(git_path(\"MERGE_MSG\"), \"a\");\n \tif (!fp)\n@@ -841,7 +841,10 @@ static int suggest_conflicts(void)\n \t\t}\n \t}\n \tfclose(fp);\n-\trerere(allow_rerere_auto);\n+\tflag = allow_rerere_auto;\n+\tif (merge_renormalize)\n+\t\tflag |= RERERE_RENORMALIZE;\n+\trerere(flag);\n \tprintf(\"Automatic merge failed; \"\n \t\t\t\"fix conflicts and then commit the result.\\n\");\n \treturn 1;\ndiff --git a/builtin/rerere.c b/builtin/rerere.c\nindex 295fe75..4009c01 100644\n--- a/builtin/rerere.c\n+++ b/builtin/rerere.c\n@@ -111,15 +111,19 @@ int cmd_rerere(int argc, const char **argv, const char *prefix)\n \tstruct option options[] = {\n \t\tOPT_SET_INT(0, \"rerere-autoupdate\", &autoupdate,\n \t\t\t\"register clean resolutions in index\", 1),\n+\t\t{ OPTION_BIT, 0, \"renormalize-before-merge\", &flags, NULL,\n+\t\t\t\"simplify conflict hunks for normalization changes\",\n+\t\t\tPARSE_OPT_NOARG | PARSE_OPT_HIDDEN, NULL,\n+\t\t\tRERERE_RENORMALIZE },\n \t\tOPT_END(),\n \t};\n \n \targc = parse_options(argc, argv, prefix, options, rerere_usage, 0);\n \n \tif (autoupdate == 1)\n-\t\tflags = RERERE_AUTOUPDATE;\n+\t\tflags |= RERERE_AUTOUPDATE;\n \tif (autoupdate == 0)\n-\t\tflags = RERERE_NOAUTOUPDATE;\n+\t\tflags |= RERERE_NOAUTOUPDATE;\n \n \tif (argc < 1)\n \t\treturn rerere(flags);\ndiff --git a/builtin/revert.c b/builtin/revert.c\nindex 853e9e4..c694801 100644\n--- a/builtin/revert.c\n+++ b/builtin/revert.c\n@@ -338,7 +338,7 @@ static void do_recursive_merge(struct commit *base, struct commit *next,\n \trollback_lock_file(&index_lock);\n \n \tif (!clean) {\n-\t\tint i;\n+\t\tint i, flags;\n \t\tstrbuf_addstr(msgbuf, \"\\nConflicts:\\n\\n\");\n \t\tfor (i = 0; i < active_nr;) {\n \t\t\tstruct cache_entry *ce = active_cache[i++];\n@@ -354,7 +354,10 @@ static void do_recursive_merge(struct commit *base, struct commit *next,\n \t\twrite_message(msgbuf, defmsg);\n \t\tfprintf(stderr, \"Automatic %s failed.%s\\n\",\n \t\t\tme, help_msg());\n-\t\trerere(allow_rerere_auto);\n+\t\tflags = allow_rerere_auto;\n+\t\tif (o.renormalize)\n+\t\t\tflags |= RERERE_RENORMALIZE;\n+\t\trerere(flags);\n \t\texit(1);\n \t}\n \twrite_message(msgbuf, defmsg);\n@@ -481,6 +484,10 @@ static int do_pick_commit(void)\n \t\tif (res) {\n \t\t\tfprintf(stderr, \"Automatic %s with strategy %s failed.%s\\n\",\n \t\t\t\tme, strategy, help_msg());\n+\t\t\t/*\n+\t\t\t * NEEDSWORK: use RERERE_RENORMALIZE after\n+\t\t\t * a merge with merge.renormalize set\n+\t\t\t */\n \t\t\trerere(allow_rerere_auto);\n \t\t\texit(1);\n \t\t}\ndiff --git a/rerere.c b/rerere.c\nindex 17dcc3c..8767024 100644\n--- a/rerere.c\n+++ b/rerere.c\n@@ -553,12 +553,13 @@ int setup_rerere(struct string_list *merge_rr, int flags)\n int rerere(int flags)\n {\n \tstruct string_list merge_rr = { NULL, 0, 0, 1 };\n-\tint fd;\n+\tint fd, renormalize;\n \n \tfd = setup_rerere(&merge_rr, flags);\n \tif (fd < 0)\n \t\treturn 0;\n-\treturn do_plain_rerere(&merge_rr, fd, merge_renormalize);\n+\trenormalize = (flags & RERERE_RENORMALIZE) ? 1 : 0;\n+\treturn do_plain_rerere(&merge_rr, fd, renormalize);\n }\n \n static int rerere_forget_one_path(const char *path, struct string_list *rr)\ndiff --git a/rerere.h b/rerere.h\nindex eaa9004..a73555f 100644\n--- a/rerere.h\n+++ b/rerere.h\n@@ -5,6 +5,7 @@\n \n #define RERERE_AUTOUPDATE 01\n #define RERERE_NOAUTOUPDATE 02\n+#define RERERE_RENORMALIZE 04\n \n extern int setup_rerere(struct string_list *, int);\n extern int rerere(int);\n-- \n1.7.2.1.544.ga752d.dirty\n"},{"id":"147093","messageId":"20100804032937.GG19699@burratino","threadId":"24269","inReplyTo":"20100804031935.GA19699@burratino","subject":"[PATCH 6/6] merge-recursive: add -Xrenormalize option","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2010-08-04T03:29:37Z","receivedAt":"2010-08-04T03:29:37Z","isPatch":true,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"This adds a -Xrenormalize option to override the\nmerge.renormalize configuration. There is no way to override\nit in the negative yet, but hopefully this suggests how.\n\nThe good part: this destroys the global merge_renormalize\nvariable.\n\nThe bad part: merge.renormalize is still not honored for\nmost commands. And it reveals lots of places that -X has not\nbeen plumbed in.\n\nCc: Eyvind Bernhardsen \nSigned-off-by: Jonathan Nieder \n---\nThere you have it. As I said, this is pretty rough. Places I would\nbe happy to see this go:\n\n - add tests and documentation\n - -Xno-renormalize to override [merge] renormalize in the negative\n - checkout -X and rerere -X\n\nI probably will not have time to work on this topic again for a while,\nso I would be very happy if someone takes ideas from it and takes it\nsomewhere else before I get a chance. If there are just minor\ncleanups to do, I can maintain this series for a short while, though.\n\nThoughts?\n\n Documentation/merge-strategies.txt | 8 ++++++++\n builtin/checkout.c | 7 +++++++\n builtin/merge-recursive.c | 2 ++\n builtin/merge.c | 19 +++++++++++++------\n builtin/revert.c | 7 +++++++\n cache.h | 1 -\n environment.c | 1 -\n merge-recursive.c | 2 +-\n 8 files changed, 38 insertions(+), 9 deletions(-)\n\ndiff --git a/Documentation/merge-strategies.txt b/Documentation/merge-strategies.txt\nindex a5bc1db..350c3c7 100644\n--- a/Documentation/merge-strategies.txt\n+++ b/Documentation/merge-strategies.txt\n@@ -40,6 +40,14 @@ the other tree did, declaring 'our' history contains all that happened in it.\n theirs;;\n \tThis is opposite of 'ours'.\n \n+renormalize;;\n+\tThis runs a virtual check-out and check-in of all three stages\n+\tof a file when resolving a three-way merge. This option is\n+\tmeant to be used when merging branches with different clean\n+\tfilters or end-of-line normalization rules. See \"Merging\n+\tbranches with differing checkin/checkout attributes\" in\n+\tlinkgit:gitattributes[5] for details.\n+\n subtree[=path];;\n \tThis option is a more advanced form of 'subtree' strategy, where\n \tthe strategy makes a guess on how two trees must be shifted to\ndiff --git a/builtin/checkout.c b/builtin/checkout.c\nindex 5a7ae03..00fd4cd 100644\n--- a/builtin/checkout.c\n+++ b/builtin/checkout.c\n@@ -437,6 +437,13 @@ static int merge_working_tree(struct checkout_opts *opts,\n \t\t\t */\n \n \t\t\tadd_files_to_cache(NULL, NULL, 0);\n+\t\t\t/*\n+\t\t\t * NEEDSWORK: carrying over local changes\n+\t\t\t * when branches have different end-of-line\n+\t\t\t * normalization (or clean+smudge rules) is\n+\t\t\t * a pain; plumb in an option to set\n+\t\t\t * o.renormalize?\n+\t\t\t */\n \t\t\tinit_merge_options(&o);\n \t\t\to.verbosity = 0;\n \t\t\twork = write_tree_from_memory(&o);\ndiff --git a/builtin/merge-recursive.c b/builtin/merge-recursive.c\nindex d8875d5..2cc84d9 100644\n--- a/builtin/merge-recursive.c\n+++ b/builtin/merge-recursive.c\n@@ -45,6 +45,8 @@ int cmd_merge_recursive(int argc, const char **argv, const char *prefix)\n \t\t\t\to.subtree_shift = \"\";\n \t\t\telse if (!prefixcmp(arg+2, \"subtree=\"))\n \t\t\t\to.subtree_shift = arg + 10;\n+\t\t\telse if (!strcmp(arg+2, \"renormalize\"))\n+\t\t\t\to.renormalize = 1;\n \t\t\telse\n \t\t\t\tdie(\"Unknown option %s\", arg);\n \t\t\tcontinue;\ndiff --git a/builtin/merge.c b/builtin/merge.c\nindex 823b76b..ad0aa5a 100644\n--- a/builtin/merge.c\n+++ b/builtin/merge.c\n@@ -54,6 +54,7 @@ static size_t use_strategies_nr, use_strategies_alloc;\n static const char **xopts;\n static size_t xopts_nr, xopts_alloc;\n static const char *branch;\n+static int option_renormalize;\n static int verbosity;\n static int allow_rerere_auto;\n \n@@ -503,9 +504,8 @@ static int git_merge_config(const char *k, const char *v, void *cb)\n \t\treturn git_config_string(&pull_octopus, k, v);\n \telse if (!strcmp(k, \"merge.log\") || !strcmp(k, \"merge.summary\"))\n \t\toption_log = git_config_bool(k, v);\n-\telse if (!strcmp(k, \"merge.renormalize\")) {\n-\t\tmerge_renormalize = git_config_bool(k, v);\n-\t}\n+\telse if (!strcmp(k, \"merge.renormalize\"))\n+\t\toption_renormalize = git_config_bool(k, v);\n \treturn git_diff_ui_config(k, v, cb);\n }\n \n@@ -627,6 +627,11 @@ static int try_merge_strategy(const char *strategy, struct commit_list *common,\n \t\tif (!strcmp(strategy, \"subtree\"))\n \t\t\to.subtree_shift = \"\";\n \n+\t\to.renormalize = option_renormalize;\n+\n+\t\t/*\n+\t\t * NEEDSWORK: merge with table in builtin/merge-recursive\n+\t\t */\n \t\tfor (x = 0; x < xopts_nr; x++) {\n \t\t\tif (!strcmp(xopts[x], \"ours\"))\n \t\t\t\to.recursive_variant = MERGE_RECURSIVE_OURS;\n@@ -636,6 +641,8 @@ static int try_merge_strategy(const char *strategy, struct commit_list *common,\n \t\t\t\to.subtree_shift = \"\";\n \t\t\telse if (!prefixcmp(xopts[x], \"subtree=\"))\n \t\t\t\to.subtree_shift = xopts[x]+8;\n+\t\t\telse if (!strcmp(xopts[x], \"renormalize\"))\n+\t\t\t\to.renormalize = 1;\n \t\t\telse\n \t\t\t\tdie(\"Unknown option for merge-recursive: -X%s\", xopts[x]);\n \t\t}\n@@ -819,7 +826,7 @@ static int finish_automerge(struct commit_list *common,\n \treturn 0;\n }\n \n-static int suggest_conflicts(void)\n+static int suggest_conflicts(int renormalizing)\n {\n \tFILE *fp;\n \tint pos, flag;\n@@ -842,7 +849,7 @@ static int suggest_conflicts(void)\n \t}\n \tfclose(fp);\n \tflag = allow_rerere_auto;\n-\tif (merge_renormalize)\n+\tif (renormalizing)\n \t\tflag |= RERERE_RENORMALIZE;\n \trerere(flag);\n \tprintf(\"Automatic merge failed; \"\n@@ -1307,5 +1314,5 @@ int cmd_merge(int argc, const char **argv, const char *prefix)\n \t\t\t\"stopped before committing as requested\\n\");\n \t\treturn 0;\n \t} else\n-\t\treturn suggest_conflicts();\n+\t\treturn suggest_conflicts(option_renormalize);\n }\ndiff --git a/builtin/revert.c b/builtin/revert.c\nindex c694801..1999252 100644\n--- a/builtin/revert.c\n+++ b/builtin/revert.c\n@@ -318,6 +318,13 @@ static void do_recursive_merge(struct commit *base, struct commit *next,\n \tindex_fd = hold_locked_index(&index_lock, 1);\n \n \tread_cache();\n+\n+\t/*\n+\t * NEEDSWORK: cherry-picking between branches with\n+\t * different end-of-line normalization is a pain;\n+\t * plumb in an option to set o.renormalize?\n+\t * (or better: arbitrary -X options)\n+\t */\n \tinit_merge_options(&o);\n \to.ancestor = base ? base_label : \"(empty tree)\";\n \to.branch1 = \"HEAD\";\ndiff --git a/cache.h b/cache.h\nindex ed73da8..aa725b0 100644\n--- a/cache.h\n+++ b/cache.h\n@@ -551,7 +551,6 @@ extern int read_replace_refs;\n extern int fsync_object_files;\n extern int core_preload_index;\n extern int core_apply_sparse_checkout;\n-extern int merge_renormalize;\n \n enum safe_crlf {\n \tSAFE_CRLF_FALSE = 0,\ndiff --git a/environment.c b/environment.c\nindex 81a3682..83d38d3 100644\n--- a/environment.c\n+++ b/environment.c\n@@ -53,7 +53,6 @@ enum object_creation_mode object_creation_mode = OBJECT_CREATION_MODE;\n char *notes_ref_name;\n int grafts_replace_parents = 1;\n int core_apply_sparse_checkout;\n-int merge_renormalize;\n \n /* Parallel index stat data preload? */\n int core_preload_index = 0;\ndiff --git a/merge-recursive.c b/merge-recursive.c\nindex 4838939..73f2768 100644\n--- a/merge-recursive.c\n+++ b/merge-recursive.c\n@@ -1485,7 +1485,7 @@ void init_merge_options(struct merge_options *o)\n \to->buffer_output = 1;\n \to->diff_rename_limit = -1;\n \to->merge_rename_limit = -1;\n-\to->renormalize = merge_renormalize;\n+\to->renormalize = 0;\n \tgit_config(merge_recursive_config, o);\n \tif (getenv(\"GIT_MERGE_VERBOSITY\"))\n \t\to->verbosity =\n-- \n1.7.2.1.544.ga752d.dirty\n"},{"id":"147119","messageId":"7v39uugses.fsf@alter.siamese.dyndns.org","threadId":"24269","inReplyTo":"20100804031935.GA19699@burratino","subject":"Re: [PATCH/RFC eb/double-convert-before-merge 0/6] merge -Xrenormalize","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2010-08-04T18:10:51Z","receivedAt":"2010-08-04T18:10:51Z","isPatch":true,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Jonathan Nieder writes:\n\n> But for later, they tell a nice story. The idea is that the\n> \"merge.renormalize\" is not really about the behavior of the \"git\n> merge\" command but about the merge-recursive driver; and so it\n> should be usable everywhere else that driver is used, too.\n> ...\n> So in an ideal world, you would be able to, e.g.:\n>\n> \tgit checkout -m -Xrenormalize otherbranch\n>\n> or\n>\n> \tgit revert -Xrenormalize otherpatch\n>\n> or\n>\n> \tgit pull --rebase -Xrenormalize\n>\n> Of course, the plumbing is not all there yet; many commands that\n> use merge drivers cannot pass through arbitrary options through\n> yet.\n>\n> Well, enough talking. Here are the patches; what do you think?\n\nNicely summarized.\n"},{"id":"147120","messageId":"7vvd7qfdsa.fsf@alter.siamese.dyndns.org","threadId":"24269","inReplyTo":"20100804032144.GD19699@burratino","subject":"Re: [PATCH 3/6] ll-merge: let caller decide whether to renormalize","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2010-08-04T18:12:05Z","receivedAt":"2010-08-04T18:12:05Z","isPatch":true,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Jonathan Nieder writes:\n\n> Add a \"renormalize\" parameter to ll_merge so callers can decide on a\n> case-by-case basis whether the merge is likely to have overlapped with\n> a change in smudge/clean rules. This reveals a few commands that have\n> not been taking that situation into account, though it does not fix\n> them.\n>\n> No change in behavior intended.\n\nLooking nice so far; I however think a separate \"renormalize\" parameter\nshould be folded into \"flag\" word, perhaps with a few new macros to make\nit easier to access renormalize, virtual-ancestor, and favor-side parts\nit.\n"},{"id":"147121","messageId":"7vocdifdrk.fsf@alter.siamese.dyndns.org","threadId":"24269","inReplyTo":"20100804032338.GF19699@burratino","subject":"Re: [PATCH 5/6] rerere: let caller decide whether to renormalize","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2010-08-04T18:12:31Z","receivedAt":"2010-08-04T18:12:31Z","isPatch":true,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Jonathan Nieder writes:\n\n> NEEDSWORK: this is a step in the wrong direction. rerere needs\n> an -s option to use an arbitrary merge strategy and a -X option to\n> pass arbitrary options to that driver.\n\nIf you are talking about \"rerere\", I strongly disagree with that.\n\n1. \"-s\"\n\n A \"merge strategy\" deals with the shape of the history (e.g. common\n ancestor selection or synthesis for the purpose of 3-way merge) and the\n shape of the trees (e.g. rename detection, subtree shifting). Starting\n from two commits, it decides, based on the shape of the history, what\n three trees your tree-level 3-way merge would operate on, and then\n based on the shape of the trees, decides the pairing of blobs to run\n the 3-way merge at the file-content level.\n\n When \"rerere\" is invoked, a strategy already has dealt with all of the\n above, and \"rerere\" only works on the (half-completed) result of that.\n It is purely a three-way merge at the file-contents level and there is\n no room for a \"strategy\" to get involved. It makes direct calls to\n ll_merge() exactly for this reason.\n\n2. \"-X\"\n\n In the \"merge -X\" syntax, \"-X\" does not stand for \"low level\n details\"; it just means \"eXternal callout\". It is there just to tell\n the \"merge\" front-end \"You may not understand this yourself, but the\n program you call does, so just pass it along\".\n\n IOW, we want to be able to pass -- through the frontend to the\n backend, without having to tell all the options any possible backends\n may know to the frontend. Just to make it easier to parse and tell\n which ones are the front-end options and which ones are not (for both\n machines and humans), we say -X to the frontend. Then \"merge\"\n turns that -X into \"--\" and give it to the strategy.\n\n A command that natively knows about an option is correct to take an\n option as \"--