1Checklist (and a short version for the impatient): 2 3 - make commits of logical units 4 - check for unnecessary whitespace with "git diff --check" 5 before committing 6 - do not check in commented out code or unneeded files 7 - provide a meaningful commit message 8 - the first line of the commit message should be a short 9 description and should skip the full stop 10 - if you want your work included in git.git, add a 11 "Signed-off-by: Your Name <your@email.com>" line to the 12 commit message (or just use the option "-s" when 13 committing) to confirm that you agree to the Developer's 14 Certificate of Origin 15 - do not PGP sign your patch 16 - use "git format-patch -M" to create the patch 17 - do not attach your patch, but read in the mail 18 body, unless you cannot teach your mailer to 19 leave the formatting of the patch alone. 20 - be careful doing cut & paste into your mailer, not to 21 corrupt whitespaces. 22 - provide additional information (which is unsuitable for 23 the commit message) between the "---" and the diffstat 24 - send the patch to the list _and_ the maintainer 25 - if you change, add, or remove a command line option or 26 make some other user interface change, the associated 27 documentation should be updated as well. 28 29Long version: 30 31I started reading over the SubmittingPatches document for Linux 32kernel, primarily because I wanted to have a document similar to 33it for the core GIT to make sure people understand what they are 34doing when they write "Signed-off-by" line. 35 36But the patch submission requirements are a lot more relaxed 37here on the technical/contents front, because the core GIT is 38thousand times smaller ;-). So here is only the relevant bits. 39 40 41(1) Make separate commits for logically separate changes. 42 43Unless your patch is really trivial, you should not be sending 44out a patch that was generated between your working tree and 45your commit head. Instead, always make a commit with complete 46commit message and generate a series of patches from your 47repository. It is a good discipline. 48 49Describe the technical detail of the change(s). 50 51If your description starts to get too long, that's a sign that you 52probably need to split up your commit to finer grained pieces. 53 54Oh, another thing. I am picky about whitespaces. Make sure your 55changes do not trigger errors with the sample pre-commit hook shipped 56in templates/hooks--pre-commit. To help ensure this does not happen, 57run git diff --check on your changes before you commit. 58 59 60(2) Generate your patch using git tools out of your commits. 61 62git based diff tools (git, Cogito, and StGIT included) generate 63unidiff which is the preferred format. 64 65You do not have to be afraid to use -M option to "git diff" or 66"git format-patch", if your patch involves file renames. The 67receiving end can handle them just fine. 68 69Please make sure your patch does not include any extra files 70which do not belong in a patch submission. Make sure to review 71your patch after generating it, to ensure accuracy. Before 72sending out, please make sure it cleanly applies to the "master" 73branch head. If you are preparing a work based on "next" branch, 74that is fine, but please mark it as such. 75 76 77(3) Sending your patches. 78 79People on the git mailing list need to be able to read and 80comment on the changes you are submitting. It is important for 81a developer to be able to "quote" your changes, using standard 82e-mail tools, so that they may comment on specific portions of 83your code. For this reason, all patches should be submitted 84"inline". WARNING: Be wary of your MUAs word-wrap 85corrupting your patch. Do not cut-n-paste your patch; you can 86lose tabs that way if you are not careful. 87 88It is a common convention to prefix your subject line with 89[PATCH]. This lets people easily distinguish patches from other 90e-mail discussions. 91 92"git format-patch" command follows the best current practice to 93format the body of an e-mail message. At the beginning of the 94patch should come your commit message, ending with the 95Signed-off-by: lines, and a line that consists of three dashes, 96followed by the diffstat information and the patch itself. If 97you are forwarding a patch from somebody else, optionally, at 98the beginning of the e-mail message just before the commit 99message starts, you can put a "From: " line to name that person. 100 101You often want to add additional explanation about the patch, 102other than the commit message itself. Place such "cover letter" 103material between the three dash lines and the diffstat. 104 105Do not attach the patch as a MIME attachment, compressed or not. 106Do not let your e-mail client send quoted-printable. Do not let 107your e-mail client send format=flowed which would destroy 108whitespaces in your patches. Many 109popular e-mail applications will not always transmit a MIME 110attachment as plain text, making it impossible to comment on 111your code. A MIME attachment also takes a bit more time to 112process. This does not decrease the likelihood of your 113MIME-attached change being accepted, but it makes it more likely 114that it will be postponed. 115 116Exception: If your mailer is mangling patches then someone may ask 117you to re-send them using MIME, that is OK. 118 119Do not PGP sign your patch, at least for now. Most likely, your 120maintainer or other people on the list would not have your PGP 121key and would not bother obtaining it anyway. Your patch is not 122judged by who you are; a good patch from an unknown origin has a 123far better chance of being accepted than a patch from a known, 124respected origin that is done poorly or does incorrect things. 125 126If you really really really really want to do a PGP signed 127patch, format it as "multipart/signed", not a text/plain message 128that starts with '-----BEGIN PGP SIGNED MESSAGE-----'. That is 129not a text/plain, it's something else. 130 131Note that your maintainer does not necessarily read everything 132on the git mailing list. If your patch is for discussion first, 133send it "To:" the mailing list, and optionally "cc:" him. If it 134is trivially correct or after the list reached a consensus, send 135it "To:" the maintainer and optionally "cc:" the list. 136 137Also note that your maintainer does not actively involve himself in 138maintaining what are in contrib/ hierarchy. When you send fixes and 139enhancements to them, do not forget to "cc: " the person who primarily 140worked on that hierarchy in contrib/. 141 142 143(4) Sign your work 144 145To improve tracking of who did what, we've borrowed the 146"sign-off" procedure from the Linux kernel project on patches 147that are being emailed around. Although core GIT is a lot 148smaller project it is a good discipline to follow it. 149 150The sign-off is a simple line at the end of the explanation for 151the patch, which certifies that you wrote it or otherwise have 152the right to pass it on as a open-source patch. The rules are 153pretty simple: if you can certify the below: 154 155 Developer's Certificate of Origin 1.1 156 157 By making a contribution to this project, I certify that: 158 159 (a) The contribution was created in whole or in part by me and I 160 have the right to submit it under the open source license 161 indicated in the file; or 162 163 (b) The contribution is based upon previous work that, to the best 164 of my knowledge, is covered under an appropriate open source 165 license and I have the right under that license to submit that 166 work with modifications, whether created in whole or in part 167 by me, under the same open source license (unless I am 168 permitted to submit under a different license), as indicated 169 in the file; or 170 171 (c) The contribution was provided directly to me by some other 172 person who certified (a), (b) or (c) and I have not modified 173 it. 174 175 (d) I understand and agree that this project and the contribution 176 are public and that a record of the contribution (including all 177 personal information I submit with it, including my sign-off) is 178 maintained indefinitely and may be redistributed consistent with 179 this project or the open source license(s) involved. 180 181then you just add a line saying 182 183 Signed-off-by: Random J Developer <random@developer.example.org> 184 185This line can be automatically added by git if you run the git-commit 186command with the -s option. 187 188Some people also put extra tags at the end. They'll just be ignored for 189now, but you can do this to mark internal company procedures or just 190point out some special detail about the sign-off. 191 192 193------------------------------------------------ 194MUA specific hints 195 196Some of patches I receive or pick up from the list share common 197patterns of breakage. Please make sure your MUA is set up 198properly not to corrupt whitespaces. Here are two common ones 199I have seen: 200 201* Empty context lines that do not have _any_ whitespace. 202 203* Non empty context lines that have one extra whitespace at the 204 beginning. 205 206One test you could do yourself if your MUA is set up correctly is: 207 208* Send the patch to yourself, exactly the way you would, except 209 To: and Cc: lines, which would not contain the list and 210 maintainer address. 211 212* Save that patch to a file in UNIX mailbox format. Call it say 213 a.patch. 214 215* Try to apply to the tip of the "master" branch from the 216 git.git public repository: 217 218 $ git fetch http://kernel.org/pub/scm/git/git.git master:test-apply 219 $ git checkout test-apply 220 $ git reset --hard 221 $ git applymbox a.patch 222 223If it does not apply correctly, there can be various reasons. 224 225* Your patch itself does not apply cleanly. That is _bad_ but 226 does not have much to do with your MUA. Please rebase the 227 patch appropriately. 228 229* Your MUA corrupted your patch; applymbox would complain that 230 the patch does not apply. Look at .dotest/ subdirectory and 231 see what 'patch' file contains and check for the common 232 corruption patterns mentioned above. 233 234* While you are at it, check what are in 'info' and 235 'final-commit' files as well. If what is in 'final-commit' is 236 not exactly what you would want to see in the commit log 237 message, it is very likely that your maintainer would end up 238 hand editing the log message when he applies your patch. 239 Things like "Hi, this is my first patch.\n", if you really 240 want to put in the patch e-mail, should come after the 241 three-dash line that signals the end of the commit message. 242 243 244Pine 245---- 246 247(Johannes Schindelin) 248 249I don't know how many people still use pine, but for those poor 250souls it may be good to mention that the quell-flowed-text is 251needed for recent versions. 252 253... the "no-strip-whitespace-before-send" option, too. AFAIK it 254was introduced in 4.60. 255 256(Linus Torvalds) 257 258And 4.58 needs at least this. 259 260--- 261diff-tree 8326dd8350be64ac7fc805f6563a1d61ad10d32c (from e886a61f76edf5410573e92e38ce22974f9c40f1) 262Author: Linus Torvalds <torvalds@g5.osdl.org> 263Date: Mon Aug 15 17:23:51 2005 -0700 264 265 Fix pine whitespace-corruption bug 266 267 There's no excuse for unconditionally removing whitespace from 268 the pico buffers on close. 269 270diff --git a/pico/pico.c b/pico/pico.c 271--- a/pico/pico.c 272+++ b/pico/pico.c 273@@ -219,7 +219,9 @@ PICO *pm; 274 switch(pico_all_done){ /* prepare for/handle final events */ 275 case COMP_EXIT : /* already confirmed */ 276 packheader(); 277+#if 0 278 stripwhitespace(); 279+#endif 280 c |= COMP_EXIT; 281 break; 282 283 284(Daniel Barkalow) 285 286> A patch to SubmittingPatches, MUA specific help section for 287> users of Pine 4.63 would be very much appreciated. 288 289Ah, it looks like a recent version changed the default behavior to do the 290right thing, and inverted the sense of the configuration option. (Either 291that or Gentoo did it.) So you need to set the 292"no-strip-whitespace-before-send" option, unless the option you have is 293"strip-whitespace-before-send", in which case you should avoid checking 294it. 295 296 297Thunderbird 298----------- 299 300(A Large Angry SCM) 301 302Here are some hints on how to successfully submit patches inline using 303Thunderbird. 304 305This recipe appears to work with the current [*1*] Thunderbird from Suse. 306 307The following Thunderbird extensions are needed: 308 AboutConfig 0.5 309 http://aboutconfig.mozdev.org/ 310 External Editor 0.7.2 311 http://globs.org/articles.php?lng=en&pg=8 312 3131) Prepare the patch as a text file using your method of choice. 314 3152) Before opening a compose window, use Edit->Account Settings to 316uncheck the "Compose messages in HTML format" setting in the 317"Composition & Addressing" panel of the account to be used to send the 318patch. [*2*] 319 3203) In the main Thunderbird window, _before_ you open the compose window 321for the patch, use Tools->about:config to set the following to the 322indicated values: 323 mailnews.send_plaintext_flowed => false 324 mailnews.wraplength => 0 325 3264) Open a compose window and click the external editor icon. 327 3285) In the external editor window, read in the patch file and exit the 329editor normally. 330 3316) Back in the compose window: Add whatever other text you wish to the 332message, complete the addressing and subject fields, and press send. 333 3347) Optionally, undo the about:config/account settings changes made in 335steps 2 & 3. 336 337 338[Footnotes] 339*1* Version 1.0 (20041207) from the MozillaThunderbird-1.0-5 rpm of Suse 3409.3 professional updates. 341 342*2* It may be possible to do this with about:config and the following 343settings but I haven't tried, yet. 344 mail.html_compose => false 345 mail.identity.default.compose_html => false 346 mail.identity.id?.compose_html => false 347 348 349Gnus 350---- 351 352'|' in the *Summary* buffer can be used to pipe the current 353message to an external program, and this is a handy way to drive 354"git am". However, if the message is MIME encoded, what is 355piped into the program is the representation you see in your 356*Article* buffer after unwrapping MIME. This is often not what 357you would want for two reasons. It tends to screw up non ASCII 358characters (most notably in people's names), and also 359whitespaces (fatal in patches). Running 'C-u g' to display the 360message in raw form before using '|' to run the pipe can work 361this problem around. 362 363 364KMail 365----- 366 367This should help you to submit patches inline using KMail. 368 3691) Prepare the patch as a text file. 370 3712) Click on New Mail. 372 3733) Go under "Options" in the Composer window and be sure that 374"Word wrap" is not set. 375 3764) Use Message -> Insert file... and insert the patch. 377 3785) Back in the compose window: add whatever other text you wish to the 379message, complete the addressing and subject fields, and press send.