Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 1 | HOW TO GET YOUR CODE ACCEPTED IN HAPROXY |
| 2 | READ THIS CAREFULLY BEFORE SUBMITTING CODE |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 3 | |
| 4 | THIS DOCUMENT PROVIDES SOME RULES TO FOLLOW WHEN SENDING CONTRIBUTIONS. PATCHES |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 5 | NOT FOLLOWING THESE RULES WILL SIMPLY BE IGNORED IN ORDER TO PROTECT ALL OTHER |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 6 | RESPECTFUL CONTRIBUTORS' VALUABLE TIME. |
| 7 | |
| 8 | |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 9 | Abstract |
| 10 | -------- |
| 11 | |
| 12 | If you have never contributed to HAProxy before, or if you did so and noticed |
| 13 | that nobody seems to be interested in reviewing your submission, please do read |
| 14 | this long document carefully. HAProxy maintainers are particularly demanding on |
| 15 | respecting certain simple rules related to general code and documentation style |
| 16 | as well as splitting your patches and providing high quality commit messages. |
| 17 | The reason behind this is that your patch will be met multiple times in the |
| 18 | future, when doing some backporting work or when bisecting a bug, and it is |
| 19 | critical that anyone can quickly decide if the patch is right, wrong, if it |
| 20 | misses something, if it must be reverted or needs to be backported. Maintainers |
| 21 | are generally benevolent with newcomers and will help them provided their work |
| 22 | indicates they have at least read this document. Some have improved over time, |
| 23 | to the point of being totally trusted and gaining commit access so they don't |
| 24 | need to depend on anyone to pick their code. On the opposite, those who insist |
| 25 | not making minimal efforts however will simply be ignored. |
| 26 | |
| 27 | |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 28 | Background |
| 29 | ---------- |
| 30 | |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 31 | HAProxy is a community-driven project. But like most highly technical projects |
| 32 | it takes a lot of time to develop the skills necessary to be autonomous in the |
| 33 | project, and there is a very small core team helped by a small set of very |
| 34 | active participants. While most of the core team members work on the code as |
| 35 | part of their day job, most participants do it on a voluntary basis during |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 36 | their spare time. The ideal model for developers is to spend their time: |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 37 | 1) developing new features |
| 38 | 2) fixing bugs |
| 39 | 3) doing maintenance backports |
| 40 | 4) reviewing other people's code |
| 41 | |
| 42 | It turns out that on a project like HAProxy, like many other similarly complex |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 43 | projects, the time spent is exactly the opposite: |
| 44 | 1) reviewing other people's code |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 45 | 2) doing maintenance backports |
| 46 | 3) fixing bugs |
| 47 | 4) developing new features |
| 48 | |
| 49 | A large part of the time spent reviewing code often consists in giving basic |
| 50 | recommendations that are already explained in this file. In addition to taking |
| 51 | time, it is not appealing for people willing to spend one hour helping others |
| 52 | to do the same thing over and over instead of discussing the code design, and |
| 53 | it tends to delay the start of code reviews. |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 54 | |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 55 | Regarding backports, they are necessary to provide a set of stable branches |
| 56 | that are deployed in production at many places. Load balancers are complex and |
| 57 | new features often induce undesired side effects in other areas, which we will |
| 58 | call bugs. Thus it's common for users to stick to a branch featuring everything |
| 59 | they need and not to upgrade too often. This backporting job is critical to the |
| 60 | ecosystem's health and must be done regularly. Very often the person devoting |
| 61 | some time on backports has little to no information about the relevance (let |
| 62 | alone importance) of a patch and is unlikely to be an expert in the area |
| 63 | affected by the patch. It's the role of the commit message to explain WHAT |
| 64 | problem the patch tries to solve, WHY it is estimated that it is a problem, and |
| 65 | HOW it tries to address it. With these elements, the person in charge of the |
| 66 | backports can decide whether or not to pick the patch. And if the patch does |
| 67 | not apply (which is common for older versions) they have information in the |
| 68 | commit message about the principle and choices that the initial developer made |
| 69 | and will try to adapt the patch sticking to these principles. Thus, the time |
| 70 | spent backporting patches solely depends on the code quality and the commit |
| 71 | message details and accuracy. |
| 72 | |
| 73 | When it turns to fixing bugs, before declaring a bug, there is an analysis |
| 74 | phase. It starts with "is this behaviour expected", "is it normal", "under what |
| 75 | circumstances does it happen", "when did it start to happen", "was it intended", |
| 76 | "was it just overlooked", and "how to fix it without breaking the initial |
| 77 | intent". A utility called "git bisect" is usually involved in determining when |
| 78 | the behaviour started to happen. It determines the first patch which introduced |
| 79 | the new behaviour. If the patch is huge, touches many areas, is really difficult |
| 80 | to read because it needlessly reindents code or adds/removes line breaks out of |
| 81 | context, it will be very difficult to figure what part of this patch broke the |
| 82 | behaviour. Then once the part is figured, if the commit message doesn't provide |
| 83 | a detailed description about the intent of the patch, i.e. the problem it was |
| 84 | trying to solve, why and how, the developer landing on that patch will really |
| 85 | feel powerless. And very often in this case, the fix for the problem will break |
| 86 | something else or something that depended on the original patch. |
| 87 | |
| 88 | But contrary to what it could look like, providing great quality patches is not |
| 89 | difficult, and developers will always help contributors improve their patches |
| 90 | quality because it's in their interest as well. History has shown that first |
| 91 | time contributors can provide an excellent work when they have carefully read |
| 92 | this document, and that people coming from projects with different practices |
| 93 | can grow from first-time contributor to trusted committer in about 6 months. |
| 94 | |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 95 | |
| 96 | Preparation |
| 97 | ----------- |
| 98 | |
| 99 | It is possible that you'll want to add a specific feature to satisfy your needs |
| 100 | or one of your customers'. Contributions are welcome, however maintainers are |
| 101 | often very picky about changes. Patches that change massive parts of the code, |
| 102 | or that touch the core parts without any good reason will generally be rejected |
| 103 | if those changes have not been discussed first. |
| 104 | |
| 105 | The proper place to discuss your changes is the HAProxy Mailing List. There are |
| 106 | enough skilled readers to catch hazardous mistakes and to suggest improvements. |
| 107 | There is no other place where you'll find as many skilled people on the project, |
| 108 | and these people can help you get your code integrated quickly. You can |
| 109 | subscribe to it by sending an empty e-mail at the following address : |
| 110 | |
| 111 | haproxy+subscribe@formilux.org |
| 112 | |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 113 | It is not even necessary to subscribe, you can post there and verify via the |
| 114 | public list archives that your message was properly delivered. In this case you |
| 115 | should indicate in your message that you'd like responders to keep you CCed. |
| 116 | Please visit http://haproxy.org/ to figure available options to join the list. |
| 117 | |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 118 | If you have an idea about something to implement, *please* discuss it on the |
| 119 | list first. It has already happened several times that two persons did the same |
| 120 | thing simultaneously. This is a waste of time for both of them. It's also very |
| 121 | common to see some changes rejected because they're done in a way that will |
| 122 | conflict with future evolutions, or that does not leave a good feeling. It's |
| 123 | always unpleasant for the person who did the work, and it is unpleasant in |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 124 | general because people's time and efforts are valuable and would be better |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 125 | spent working on something else. That would not happen if these were discussed |
| 126 | first. There is no problem posting work in progress to the list, it happens |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 127 | quite often in fact. Just prefix your mail subject with "RFC" (it stands for |
| 128 | "request for comments") and everyone will understand you'd like some opinion |
| 129 | on your work in progress. Also, don't waste your time with the doc when |
| 130 | submitting patches for review, only add the doc with the patch you consider |
| 131 | ready to merge (unless you need some help on the doc itself, of course). |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 132 | |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 133 | Another important point concerns code portability. HAProxy requires gcc as the |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 134 | C compiler, and may or may not work with other compilers. However it's known to |
| 135 | build using gcc 2.95 or any later version. As such, it is important to keep in |
| 136 | mind that certain facilities offered by recent versions must not be used in the |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 137 | code: |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 138 | |
| 139 | - declarations mixed in the code (requires gcc >= 3.x and is a bad practice) |
| 140 | - GCC builtins without checking for their availability based on version and |
| 141 | architecture ; |
| 142 | - assembly code without any alternate portable form for other platforms |
| 143 | - use of stdbool.h, "bool", "false", "true" : simply use "int", "0", "1" |
| 144 | - in general, anything which requires C99 (such as declaring variables in |
| 145 | "for" statements) |
| 146 | |
| 147 | Since most of these restrictions are just a matter of coding style, it is |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 148 | normally not a problem to comply. Please read doc/coding-style.txt for all the |
| 149 | details. |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 150 | |
Willy Tarreau | 9d84cd6 | 2017-07-18 06:56:40 +0200 | [diff] [blame] | 151 | When modifying some optional subsystem (SSL, Lua, compression, device detection |
| 152 | engines), please make sure the code continues to build (and to work) when these |
| 153 | features are disabled. Similarly, when modifying the SSL stack, please always |
| 154 | ensure that supported OpenSSL versions continue to build and to work, especially |
| 155 | if you modify support for alternate libraries. Clean support for the legacy |
| 156 | OpenSSL libraries is mandatory, support for its derivatives is a bonus and may |
| 157 | occasionally break eventhough a great care is taken. In other words, if you |
| 158 | provide a patch for OpenSSL you don't need to test its derivatives, but if you |
| 159 | provide a patch for a derivative you also need to test with OpenSSL. |
| 160 | |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 161 | If your work is very confidential and you can't publicly discuss it, you can |
| 162 | also mail willy@haproxy.org directly about it, but your mail may be waiting |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 163 | several days in the queue before you get a response, if you get a response at |
| 164 | all. Retransmit if you don't get a response by one week. Please note that |
| 165 | direct sent e-mails to this address for non-confidential subjects may simply |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 166 | be forwarded to the list or be deleted without notification. An auto-responder |
| 167 | bot is in place to try to detect e-mails from people asking for help and to |
| 168 | redirect them to the mailing list. Do not be surprised if this happens to you. |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 169 | |
| 170 | If you'd like a feature to be added but you think you don't have the skills to |
| 171 | implement it yourself, you should follow these steps : |
| 172 | |
| 173 | 1. discuss the feature on the mailing list. It is possible that someone |
| 174 | else has already implemented it, or that someone will tell you how to |
| 175 | proceed without it, or even why not to do it. It is also possible that |
| 176 | in fact it's quite easy to implement and people will guide you through |
| 177 | the process. That way you'll finally have YOUR patch merged, providing |
| 178 | the feature YOU need. |
| 179 | |
| 180 | 2. if you really can't code it yourself after discussing it, then you may |
| 181 | consider contacting someone to do the job for you. Some people on the |
| 182 | list might sometimes be OK with trying to do it. |
| 183 | |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 184 | The version control system used by the project (Git) keeps authorship |
| 185 | information in the form of the patch author's e-mail address. This way you will |
| 186 | be credited for your work in the project's history. If you contract with |
| 187 | someone to implement your idea you may have to discuss such modalities with |
| 188 | the person doing the work as by default this person will be mentioned as the |
| 189 | work's author. |
| 190 | |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 191 | |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 192 | Rules: the 12 laws of patch contribution |
| 193 | ---------------------------------------- |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 194 | |
| 195 | People contributing patches must apply the following rules. That may sound heavy |
| 196 | at the beginning but it's common sense more than anything else and contributors |
| 197 | do not think about them anymore after a few patches. |
| 198 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 199 | 1) Comply with the license |
| 200 | |
| 201 | Before modifying some code, you have read the LICENSE file ("main license") |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 202 | coming with the sources, and all the files this file references. Certain |
| 203 | files may be covered by different licenses, in which case it will be |
| 204 | indicated in the files themselves. In any case, you agree to respect these |
| 205 | licenses and to contribute your changes under the same licenses. If you want |
| 206 | to create new files, they will be under the main license, or any license of |
| 207 | your choice that you have verified to be compatible with the main license, |
Tim Düsterhus | 4896c44 | 2016-11-29 02:15:19 +0100 | [diff] [blame] | 208 | and that will be explicitly mentioned in the affected files. The project's |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 209 | maintainers are free to reject contributions proposing license changes they |
| 210 | feel are not appropriate or could cause future trouble. |
| 211 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 212 | 2) Develop on development branch, not stable ones |
| 213 | |
| 214 | Your work may only be based on the latest development version. No development |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 215 | is made on a stable branch. If your work needs to be applied to a stable |
| 216 | branch, it will first be applied to the development branch and only then will |
| 217 | be backported to the stable branch. You are responsible for ensuring that |
| 218 | your work correctly applies to the development version. If at any moment you |
| 219 | are going to work on restructuring something important which may impact other |
| 220 | contributors, the rule that applies is that the first sent is the first |
| 221 | served. However it is considered good practice and politeness to warn others |
| 222 | in advance if you know you're going to make changes that may force them to |
| 223 | re-adapt their code, because they did probably not expect to have to spend |
| 224 | more time discovering your changes and rebasing their work. |
| 225 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 226 | 3) Read and respect the coding style |
| 227 | |
| 228 | You have read and understood "doc/coding-style.txt", and you're actively |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 229 | determined to respect it and to enforce it on your coworkers if you're going |
| 230 | to submit a team's work. We don't care what text editor you use, whether it's |
| 231 | an hex editor, cat, vi, emacs, Notepad, Word, or even Eclipse. The editor is |
| 232 | only the interface between you and the text file. What matters is what is in |
| 233 | the text file in the end. The editor is not an excuse for submitting poorly |
| 234 | indented code, which only proves that the person has no consideration for |
| 235 | quality and/or has done it in a hurry (probably worse). Please note that most |
| 236 | bugs were found in low-quality code. Reviewers know this and tend to be much |
Ilya Shipitsin | 2a950d0 | 2020-03-06 13:07:38 +0500 | [diff] [blame^] | 237 | more reluctant to accept poorly formatted code because by experience they |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 238 | won't trust their author's ability to write correct code. It is also worth |
| 239 | noting that poor quality code is painful to read and may result in nobody |
| 240 | willing to waste their time even reviewing your work. |
| 241 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 242 | 4) Present clean work |
| 243 | |
| 244 | The time it takes for you to polish your code is always much smaller than the |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 245 | time it takes others to do it for you, because they always have to wonder if |
| 246 | what they see is intended (meaning they didn't understand something) or if it |
| 247 | is a mistake that needs to be fixed. And since there are less reviewers than |
| 248 | submitters, it is vital to spread the effort closer to where the code is |
| 249 | written and not closer to where it gets merged. For example if you have to |
| 250 | write a report for a customer that your boss wants to review before you send |
| 251 | it to the customer, will you throw on his desk a pile of paper with stains, |
| 252 | typos and copy-pastes everywhere ? Will you say "come on, OK I made a mistake |
| 253 | in the company's name but they will find it by themselves, it's obvious it |
| 254 | comes from us" ? No. When in doubt, simply ask for help on the mailing list. |
| 255 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 256 | 5) Documentation is very important |
| 257 | |
| 258 | There are four levels of importance of quality in the project : |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 259 | |
| 260 | - The most important one, and by far, is the quality of the user-facing |
| 261 | documentation. This is the first contact for most users and it immediately |
| 262 | gives them an accurate idea of how the project is maintained. Dirty docs |
| 263 | necessarily belong to a dirty project. Be careful to the way the text you |
| 264 | add is presented and indented. Be very careful about typos, usual mistakes |
| 265 | such as double consonants when only one is needed or "it's" instead of |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 266 | "its", don't mix US English and UK English in the same paragraph, etc. |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 267 | When in doubt, check in a dictionary. Fixes for existing typos in the doc |
| 268 | are always welcome and chasing them is a good way to become familiar with |
| 269 | the project and to get other participants' respect and consideration. |
| 270 | |
| 271 | - The second most important level is user-facing messages emitted by the |
| 272 | code. You must try to see all the messages your code produces to ensure |
| 273 | they are understandable outside of the context where you wrote them, |
| 274 | because the user often doesn't expect them. That's true for warnings, and |
| 275 | that's even more important for errors which prevent the program from |
| 276 | working and which require an immediate and well understood fix in the |
| 277 | configuration. It's much better to say "line 35: compression level must be |
| 278 | an integer between 1 and 9" than "invalid argument at line 35". In HAProxy, |
| 279 | error handling roughly represents half of the code, and that's about 3/4 of |
| 280 | the configuration parser. Take the time to do something you're proud of. A |
| 281 | good rule of thumb is to keep in mind that your code talks to a human and |
| 282 | tries to teach him/her how to proceed. It must then speak like a human. |
| 283 | |
| 284 | - The third most important level is the code and its accompanying comments, |
| 285 | including the commit message which is a complement to your code and |
| 286 | comments. It's important for all other contributors that the code is |
| 287 | readable, fluid, understandable and that the commit message describes what |
| 288 | was done, the choices made, the possible alternatives you thought about, |
| 289 | the reason for picking this one and its limits if any. Comments should be |
| 290 | written where it's easy to have a doubt or after some error cases have been |
| 291 | wiped out and you want to explain what possibilities remain. All functions |
| 292 | must have a comment indicating what they take on input and what they |
| 293 | provide on output. Please adjust the comments when you copy-paste a |
| 294 | function or change its prototype, this type of lazy mistake is too common |
| 295 | and very confusing when reading code later to debug an issue. Do not forget |
| 296 | that others will feel really angry at you when they have to dig into your |
| 297 | code for a bug that your code caused and they feel like this code is dirty |
| 298 | or confusing, that the commit message doesn't explain anything useful and |
| 299 | that the patch should never have been accepted in the first place. That |
| 300 | will strongly impact your reputation and will definitely affect your |
| 301 | chances to contribute again! |
| 302 | |
| 303 | - The fourth level of importance is in the technical documentation that you |
| 304 | may want to add with your code. Technical documentation is always welcome |
| 305 | as it helps others make the best use of your work and to go exactly in the |
| 306 | direction you thought about during the design. This is also what reduces |
| 307 | the risk that your design gets changed in the near future due to a misuse |
| 308 | and/or a poor understanding. All such documentation is actually considered |
| 309 | as a bonus. It is more important that this documentation exists than that |
| 310 | it looks clean. Sometimes just copy-pasting your draft notes in a file to |
| 311 | keep a record of design ideas is better than losing them. Please do your |
| 312 | best so that other ones can read your doc. If these docs require a special |
| 313 | tool such as a graphics utility, ensure that the file name makes it |
| 314 | unambiguous how to process it. So there are no rules here for the contents, |
| 315 | except one. Please write the date in your file. Design docs tend to stay |
| 316 | forever and to remain long after they become obsolete. At this point that |
| 317 | can cause harm more than it can help. Writing the date in the document |
| 318 | helps developers guess the degree of validity and/or compare them with the |
| 319 | date of certain commits touching the same area. |
| 320 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 321 | 6) US-ASCII only! |
| 322 | |
| 323 | All text files and commit messages are written using the US-ASCII charset. |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 324 | Please be careful that your contributions do not contain any character not |
| 325 | printable using this charset, as they will render differently in different |
| 326 | editors and/or terminals. Avoid latin1 and more importantly UTF-8 which some |
| 327 | editors tend to abuse to replace some US-ASCII characters with their |
| 328 | typographic equivalent which aren't readable anymore in other editors. The |
| 329 | only place where alternative charsets are tolerated is in your name in the |
| 330 | commit message, but it's at your own risk as it can be mangled during the |
| 331 | merge. Anyway if you have an e-mail address, you probably have a valid |
| 332 | US-ASCII representation for it as well. |
| 333 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 334 | 7) Comments |
| 335 | |
| 336 | Be careful about comments when you move code around. It's not acceptable that |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 337 | a block of code is moved to another place leaving irrelevant comments at the |
| 338 | old place, just like it's not acceptable that a function is duplicated without |
| 339 | the comments being adjusted. The example below started to become quite common |
| 340 | during the 1.6 cycle, it is not acceptable and wastes everyone's time : |
| 341 | |
| 342 | /* Parse switching <str> to build rule <rule>. Returns 0 on error. */ |
| 343 | int parse_switching_rule(const char *str, struct rule *rule) |
| 344 | { |
| 345 | ... |
| 346 | } |
| 347 | |
| 348 | /* Parse switching <str> to build rule <rule>. Returns 0 on error. */ |
| 349 | void execute_switching_rule(struct rule *rule) |
| 350 | { |
| 351 | ... |
| 352 | } |
| 353 | |
| 354 | This patch is not acceptable either (and it's unfortunately not that rare) : |
| 355 | |
| 356 | + if (!session || !arg || list_is_empty(&session->rules->head)) |
| 357 | + return 0; |
| 358 | + |
| 359 | /* Check if session->rules is valid before dereferencing it */ |
| 360 | if (!session->rules_allocated) |
| 361 | return 0; |
| 362 | |
| 363 | - if (!arg || list_is_empty(&session->rules->head)) |
| 364 | - return 0; |
| 365 | - |
| 366 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 367 | 8) Short, readable identifiers |
| 368 | |
| 369 | Limit the length of your identifiers in the code. When your identifiers start |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 370 | to sound like sentences, it's very hard for the reader to keep on track with |
| 371 | what operation they are observing. Also long names force expressions to fit |
| 372 | on several lines which also cause some difficulties to the reader. See the |
| 373 | example below : |
| 374 | |
| 375 | int file_name_len_including_global_path; |
| 376 | int file_name_len_without_global_path; |
| 377 | int global_path_len_or_zero_if_default; |
| 378 | |
| 379 | if (global_path) |
| 380 | global_path_len_or_zero_if_default = strlen(global_path); |
| 381 | else |
| 382 | global_path_len_or_zero_if_default = 0; |
| 383 | |
| 384 | file_name_len_without_global_path = strlen(file_name); |
| 385 | file_name_len_including_global_path = |
| 386 | file_name_len_without_global_path + 1 + /* for '/' */ |
| 387 | global_path_len_or_zero_if_default ? |
| 388 | global_path_len_or_zero_if_default : default_path_len; |
| 389 | |
| 390 | Compare it to this one : |
| 391 | |
| 392 | int f, p; |
| 393 | |
| 394 | p = global_path ? strlen(global_path) : default_path_len; |
| 395 | f = p + 1 + strlen(file_name); /* 1 for '/' */ |
| 396 | |
| 397 | A good rule of thumb is that if your identifiers start to contain more than |
| 398 | 3 words or more than 15 characters, they can become confusing. For function |
| 399 | names it's less important especially if these functions are rarely used or |
Bertrand Jacquin | d5e4de8 | 2018-10-13 16:06:18 +0100 | [diff] [blame] | 400 | are used in a complex context where it is important to differentiate between |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 401 | their multiple variants. |
| 402 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 403 | 9) Unified diff only |
| 404 | |
| 405 | The best way to build your patches is to use "git format-patch". This means |
| 406 | that you have committed your patch to a local branch, with an appropriate |
| 407 | subject line and a useful commit message explaining what the patch attempts |
| 408 | to do. It is not strictly required to use git, but what is strictly required |
Bertrand Jacquin | d5e4de8 | 2018-10-13 16:06:18 +0100 | [diff] [blame] | 409 | is to have all these elements in the same mail, easily distinguishable, and |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 410 | a patch in "diff -up" format (which is also the format used by Git). This |
| 411 | means the "unified" diff format must be used exclusively, and with the |
| 412 | function name printed in the diff header of each block. That significantly |
| 413 | helps during reviews. Keep in mind that most reviews are done on the patch |
| 414 | and not on the code after applying the patch. Your diff must keep some |
| 415 | context (3 lines above and 3 lines below) so that there's no doubt where the |
| 416 | code has to be applied. Don't change code outside of the context of your |
| 417 | patch (eg: take care of not adding/removing empty lines once you remove |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 418 | your debugging code). If you are using Git (which is strongly recommended), |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 419 | always use "git show" after doing a commit to ensure it looks good, and |
| 420 | enable syntax coloring that will automatically report in red the trailing |
| 421 | spaces or tabs that your patch added to the code and that must absolutely be |
| 422 | removed. These ones cause a real pain to apply patches later because they |
| 423 | mangle the context in an invisible way. Such patches with trailing spaces at |
| 424 | end of lines will be rejected. |
| 425 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 426 | 10) One patch per feature |
| 427 | |
| 428 | Please cut your work in series of patches that can be independently reviewed |
| 429 | and merged. Each patch must do something on its own that you can explain to |
| 430 | someone without being ashamed of what you did. For example, you must not say |
| 431 | "This is the patch that implements SSL, it was tricky". There's clearly |
| 432 | something wrong there, your patch will be huge, will definitely break things |
| 433 | and nobody will be able to figure what exactly introduced the bug. However |
| 434 | it's much better to say "I needed to add some fields in the session to store |
| 435 | the SSL context so this patch does this and doesn't touch anything else, so |
| 436 | it's safe". Also when dealing with series, you will sometimes fix a bug that |
| 437 | one of your patches introduced. Please do merge these fixes (eg: using git |
| 438 | rebase -i and squash or fixup), as it is not acceptable to see patches which |
| 439 | introduce known bugs even if they're fixed later. Another benefit of cleanly |
| 440 | splitting patches is that if some of your patches need to be reworked after |
| 441 | a review, the other ones can still be merged so that you don't need to care |
| 442 | about them anymore. When sending multiple patches for review, prefer to send |
| 443 | one e-mail per patch than all patches in a single e-mail. The reason is that |
| 444 | not everyone is skilled in all areas nor has the time to review everything |
| 445 | at once. With one patch per e-mail, it's easy to comment on a single patch |
| 446 | without giving an opinion on the other ones, especially if a long thread |
| 447 | starts about one specific patch on the mailing list. "git send-email" does |
| 448 | that for you though it requires a few trials before getting it right. |
| 449 | |
| 450 | If you can, please always put all the bug fixes at the beginning of the |
| 451 | series. This often makes it easier to backport them because they will not |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 452 | depend on context that your other patches changed. As a hint, if you can't |
| 453 | do this, there are little chances that your bug fix can be backported. |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 454 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 455 | 11) Real commit messages please! |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 456 | |
Willy Tarreau | 41f638c | 2019-07-26 15:21:54 +0200 | [diff] [blame] | 457 | The commit message is how you're trying to convince a maintainer to adopt |
| 458 | your work and maintain it as long as possible. A dirty commit message almost |
| 459 | always comes with dirty code. Too short a commit message indicates that too |
| 460 | short an analysis was done and that side effects are extremely likely to be |
| 461 | encountered. It's the maintainer's job to decide to accept this work in its |
| 462 | current form or not, with the known constraints. Some patches which rework |
| 463 | architectural parts or fix sensitive bugs come with 20-30 lines of design |
| 464 | explanations, limitations, hypothesis or even doubts, and despite this it |
| 465 | happens when reading them 6 months later while trying to identify a bug that |
| 466 | developers still miss some information about corner cases. |
| 467 | |
| 468 | So please properly format your commit messages. To get an idea, just run |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 469 | "git log" on the file you've just modified. Patches always have the format |
| 470 | of an e-mail made of a subject, a description and the actual patch. If you |
| 471 | are sending a patch as an e-mail formatted this way, it can quickly be |
| 472 | applied with limited effort so that's acceptable : |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 473 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 474 | - A subject line (may wrap to the next line, but please read below) |
| 475 | - an empty line (subject delimiter) |
| 476 | - a non-empty description (the body of the e-mail) |
| 477 | - the patch itself |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 478 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 479 | The subject describes the "What" of the change ; the description explains |
| 480 | the "why", the "how" and sometimes "what next". For example a commit message |
| 481 | looking like this will be rejected : |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 482 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 483 | | From: Mr Foobar <foobar@example.com> |
| 484 | | Subject: BUG: fix typo in ssl_sock |
| 485 | | |
| 486 | |
| 487 | This one as well (too long subject, not the right place for the details) : |
| 488 | |
| 489 | | From: Mr Foobar <foobar@example.com> |
| 490 | | Subject: BUG/MEDIUM: ssl: use an error flag to prevent ssl_read() from |
| 491 | | returning 0 when dealing with large buffers because that can cause |
| 492 | | an infinite loop |
| 493 | | |
| 494 | |
| 495 | This one ought to be used instead : |
| 496 | |
| 497 | | From: Mr Foobar <foobar@example.com> |
| 498 | | Subject: BUG/MEDIUM: ssl: fix risk of infinite loop in ssl_sock |
| 499 | | |
| 500 | | ssl_read() must not return 0 on error or the caller may loop forever. |
| 501 | | Instead we add a flag to the connection to notify about the error and |
| 502 | | check it at all call places. This situation can only happen with large |
| 503 | | buffers so a workaround is to limit buffer sizes. Another option would |
| 504 | | have been to return -1 but it required to use signed ints everywhere |
| 505 | | and would have made the patch larger and riskier. This fix should be |
| 506 | | backported to versions 1.2 and upper. |
| 507 | |
| 508 | It is important to understand that for any reader to guess the text above |
| 509 | when it's absent, it will take a huge amount of time. If you made the |
| 510 | analysis leading to your patch, you must explain it, including the ideas |
| 511 | you dropped if you had a good reason for this. |
| 512 | |
| 513 | While it's not strictly required to use Git, it is strongly recommended |
| 514 | because it helps you do the cleanest job with the least effort. But if you |
| 515 | are comfortable with writing clean e-mails and inserting your patches, you |
| 516 | don't need to use Git. |
| 517 | |
| 518 | But in any case, it is important that there is a clean description of what |
| 519 | the patch does, the motivation for what it does, why it's the best way to do |
Willy Tarreau | 41f638c | 2019-07-26 15:21:54 +0200 | [diff] [blame] | 520 | it, its impacts, and what it does not yet cover. And this is particularly |
| 521 | important for bugs. A patch tagged "BUG" must absolutely explain what the |
| 522 | problem is, why it is considered as a bug. Anybody, even non-developers, |
| 523 | should be able to tell whether or not a patch is likely to address an issue |
| 524 | they are facing. Indicating what the code will do after the fix doesn't help |
| 525 | if it does not say what problem is encountered without the patch. Note that |
| 526 | in some cases the bug is purely theorical and observed by reading the code. |
| 527 | In this case it's perfectly fine to provide an estimate about possible |
| 528 | effects. Also, in HAProxy, like many projects which take a great care of |
| 529 | maintaining stable branches, patches are reviewed later so that some of them |
| 530 | can be backported to stable releases. |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 531 | |
| 532 | While reviewing hundreds of patches can seem cumbersome, with a proper |
| 533 | formatting of the subject line it actually becomes very easy. For example, |
| 534 | here's how one can find patches that need to be reviewed for backports (bugs |
| 535 | and doc) between since commit ID 827752e : |
| 536 | |
| 537 | $ git log --oneline 827752e.. | grep 'BUG\|DOC' |
| 538 | 0d79cf6 DOC: fix function name |
| 539 | bc96534 DOC: ssl: missing LF |
Joseph Herlant | e07bc14 | 2018-11-09 17:44:10 -0800 | [diff] [blame] | 540 | 10ec214 BUG/MEDIUM: lua: the lua function Channel:close() causes a segf |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 541 | bdc97a8 BUG/MEDIUM: lua: outgoing connection was broken since 1.6-dev2 |
| 542 | ba56d9c DOC: mention support for RFC 5077 TLS Ticket extension in start |
| 543 | f1650a8 DOC: clarify some points about SSL and the proxy protocol |
| 544 | b157d73 BUG/MAJOR: peers: fix current table pointer not re-initialized |
| 545 | e1ab808 BUG/MEDIUM: peers: fix wrong message id on stick table updates |
| 546 | cc79b00 BUG/MINOR: ssl: TLS Ticket Key rotation broken via socket comma |
| 547 | d8e42b6 DOC: add new file intro.txt |
| 548 | c7d7607 BUG/MEDIUM: lua: bad error processing |
| 549 | 386a127 DOC: match several lua configuration option names to those impl |
| 550 | 0f4eadd BUG/MEDIUM: counters: ensure that src_{inc,clr}_gpc0 creates a |
| 551 | |
| 552 | It is made possible by the fact that subject lines are properly formatted and |
| 553 | always respect the same principle : one part indicating the nature and |
| 554 | severity of the patch, another one to indicate which subsystem is affected, |
| 555 | and the last one is a succinct description of the change, with the important |
| 556 | part at the beginning so that it's obvious what it does even when lines are |
| 557 | truncated like above. The whole stable maintenance process relies on this. |
| 558 | For this reason, it is mandatory to respect some easy rules regarding the |
| 559 | way the subject is built. Please see the section below for more information |
| 560 | regarding this formatting. |
| 561 | |
Willy Tarreau | 9d84cd6 | 2017-07-18 06:56:40 +0200 | [diff] [blame] | 562 | As a rule of thumb, your patch MUST NEVER be made only of a subject line, |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 563 | it *must* contain a description. Even one or two lines, or indicating |
| 564 | whether a backport is desired or not. It turns out that single-line commits |
| 565 | are so rare in the Git world that they require special manual (hence |
| 566 | painful) handling when they are backported, and at least for this reason |
| 567 | it's important to keep this in mind. |
| 568 | |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 569 | Maintainers who pick your patch may slightly adjust the description as they |
| 570 | see fit. Do not see this as a failure to do a clean job, it just means they |
| 571 | think it will help them do their daily job this way. The code may also be |
| 572 | slightly adjusted before being merged (non-functional changes only, fix for |
| 573 | typos, tabs vs spaces for example), unless your patch contains a |
| 574 | Signed-off-By tag, in which case they will either modify it and mention the |
| 575 | changes after your Signed-off-By line, or (more likely) ask you to perform |
| 576 | these changes yourself. This ability to slightly adjust a patch before |
| 577 | merging is is the main reason for not using pull requests which do not |
| 578 | provide this facility and will require to iterate back and forth with the |
| 579 | submitter and significantly delay the patch inclusion. |
| 580 | |
Willy Tarreau | 9d84cd6 | 2017-07-18 06:56:40 +0200 | [diff] [blame] | 581 | Each patch fixing a bug MUST be tagged with "BUG", a severity level, an |
| 582 | indication of the affected subsystem and a brief description of the nature |
| 583 | of the issue in the subject line, and a detailed analysis in the message |
| 584 | body. The explanation of the user-visible impact and the need for |
| 585 | backporting to stable branches or not are MANDATORY. Bug fixes with no |
| 586 | indication will simply be rejected as they are very likely to cause more |
| 587 | harm when nobody is able to tell whether or not the patch needs to be |
| 588 | backported or can be reverted in case of regression. |
| 589 | |
Frédéric Lécaille | 4891e40 | 2018-06-19 14:06:07 +0200 | [diff] [blame] | 590 | When fixing a bug which is reproducible, if possible, the contributors are |
| 591 | strongly encouraged to write a regression testing VTC file for varnishtest |
| 592 | to add to reg-tests directory. More information about varnishtest may be |
| 593 | found in README file of reg-tests directory and in doc/regression-testing.txt |
| 594 | file. |
| 595 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 596 | 12) Discuss on the mailing list |
| 597 | |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 598 | Note, some first-time contributors might feel impressed or scared by posting |
| 599 | to a list. This list is frequented only by nice people who are willing to |
| 600 | help you polish your work so that it is perfect and can last long. What you |
| 601 | think could be perceived as a proof of incompetence or lack of care will |
| 602 | instead be a proof of your ability to work with a community. You will not be |
| 603 | judged nor blamed for making mistakes. The project maintainers are the ones |
| 604 | creating the most bugs and mistakes anyway, and nobody knows the project in |
| 605 | its entirety anymore so you're just like anyone else. And people who have no |
| 606 | consideration for other's work are quickly ejected from the list so the |
| 607 | place is as safe and welcoming to new contributors as it is to long time |
| 608 | ones. |
| 609 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 610 | When submitting changes, please always CC the mailing list address so that |
| 611 | everyone gets a chance to spot any issue in your code. It will also serve |
| 612 | as an advertisement for your work, you'll get more testers quicker and |
| 613 | you'll feel better knowing that people really use your work. It's often |
| 614 | convenient to prepend "[PATCH]" in front of your mail's subject to mention |
| 615 | that this e-mail contains a patch (or a series of patches), because it will |
| 616 | easily catch reviewer's attention. It's automatically done by tools such as |
| 617 | "git format-patch" and "git send-email". If you don't want your patch to be |
| 618 | merged yet and prefer to show it for discussion, better tag it as "[RFC]" |
| 619 | (stands for "Request For Comments") and it will be reviewed but not merged |
| 620 | without your approval. It is also important to CC any author mentioned in |
| 621 | the file you change, or a subsystem maintainers whose address is mentioned |
| 622 | in a MAINTAINERS file. Not everyone reads the list on a daily basis so it's |
| 623 | very easy to miss some changes. Don't consider it as a failure when a |
| 624 | reviewer tells you you have to modify your patch, actually it's a success |
| 625 | because now you know what is missing for your work to get accepted. That's |
| 626 | why you should not hesitate to CC enough people. Don't copy people who have |
| 627 | no deal with your work area just because you found their address on the |
| 628 | list. That's the best way to appear careless about their time and make them |
| 629 | reject your changes in the future. |
| 630 | |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 631 | |
| 632 | Patch classifying rules |
| 633 | ----------------------- |
| 634 | |
| 635 | There are 3 criteria of particular importance in any patch : |
| 636 | - its nature (is it a fix for a bug, a new feature, an optimization, ...) |
| 637 | - its importance, which generally reflects the risk of merging/not merging it |
| 638 | - what area it applies to (eg: http, stats, startup, config, doc, ...) |
| 639 | |
| 640 | It's important to make these 3 criteria easy to spot in the patch's subject, |
| 641 | because it's the first (and sometimes the only) thing which is read when |
| 642 | reviewing patches to find which ones need to be backported to older versions. |
| 643 | It also helps when trying to find which patch is the most likely to have caused |
| 644 | a regression. |
| 645 | |
| 646 | Specifically, bugs must be clearly easy to spot so that they're never missed. |
| 647 | Any patch fixing a bug must have the "BUG" tag in its subject. Most common |
| 648 | patch types include : |
| 649 | |
| 650 | - BUG fix for a bug. The severity of the bug should also be indicated |
| 651 | when known. Similarly, if a backport is needed to older versions, |
Willy Tarreau | 41f638c | 2019-07-26 15:21:54 +0200 | [diff] [blame] | 652 | it should be indicated on the last line of the commit message. The |
| 653 | commit message MUST ABSOLUTELY describe the problem and its impact |
| 654 | to non-developers. Any user must be able to guess if this patch is |
| 655 | likely to fix a problem they are facing. Even if the bug was |
| 656 | discovered by accident while reading the code or running an |
| 657 | automated tool, it is mandatory to try to estimate what potential |
| 658 | issue it might cause and under what circumstances. There may even |
| 659 | be security implications sometimes so a minimum analysis is really |
| 660 | required. Also please think about stable maintainers who have to |
| 661 | build the release notes, they need to have enough input about the |
| 662 | bug's impact to explain it. If the bug has been identified as a |
| 663 | regression brought by a specific patch or version, this indication |
| 664 | will be appreciated too. New maintenance releases are generally |
| 665 | emitted when a few of these patches are merged. If the bug is a |
| 666 | vulnerability for which a CVE identifier was assigned before you |
| 667 | publish the fix, you can mention it in the commit message, it will |
| 668 | help distro maintainers. |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 669 | |
Tim Düsterhus | 4896c44 | 2016-11-29 02:15:19 +0100 | [diff] [blame] | 670 | - CLEANUP code cleanup, silence of warnings, etc... theoretically no impact. |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 671 | These patches will rarely be seen in stable branches, though they |
| 672 | may appear when they remove some annoyance or when they make |
| 673 | backporting easier. By nature, a cleanup is always of minor |
| 674 | importance and it's not needed to mention it. |
| 675 | |
| 676 | - DOC updates to any of the documentation files, including README. Many |
| 677 | documentation updates are backported since they don't impact the |
| 678 | product's stability and may help users avoid bugs. So please |
| 679 | indicate in the commit message if a backport is desired. When a |
| 680 | feature gets documented, it's preferred that the doc patch appears |
| 681 | in the same patch or after the feature patch, but not before, as it |
| 682 | becomes confusing when someone working on a code base including |
| 683 | only the doc patch won't understand why a documented feature does |
| 684 | not work as documented. |
| 685 | |
| 686 | - REORG code reorganization. Some blocks may be moved to other places, |
| 687 | some important checks might be swapped, etc... These changes |
| 688 | always present a risk of regression. For this reason, they should |
| 689 | never be mixed with any bug fix nor functional change. Code is |
| 690 | only moved as-is. Indicating the risk of breakage is highly |
| 691 | recommended. Minor breakage is tolerated in such patches if trying |
| 692 | to fix it at once makes the whole change even more confusing. That |
| 693 | may happen for example when some #ifdefs need to be propagated in |
| 694 | every file consecutive to the change. |
| 695 | |
| 696 | - BUILD updates or fixes for build issues. Changes to makefiles also fall |
| 697 | into this category. The risk of breakage should be indicated if |
| 698 | known. It is also appreciated to indicate what platforms and/or |
| 699 | configurations were tested after the change. |
| 700 | |
| 701 | - OPTIM some code was optimised. Sometimes if the regression risk is very |
| 702 | low and the gains significant, such patches may be merged in the |
| 703 | stable branch. Depending on the amount of code changed or replaced |
| 704 | and the level of trust the author has in the change, the risk of |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 705 | regression should be indicated. If the optimization depends on the |
| 706 | architecture or on build options, it is important to verify that |
| 707 | the code continues to work without it. |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 708 | |
| 709 | - RELEASE release of a new version (development or stable). |
| 710 | |
| 711 | - LICENSE licensing updates (may impact distro packagers). |
| 712 | |
Frédéric Lécaille | a8cf95d | 2018-06-20 10:14:01 +0200 | [diff] [blame] | 713 | - REGTEST updates to any of the regression testing files found in reg-tests |
| 714 | directory, including README or any documentation file. |
| 715 | |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 716 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 717 | When the patch cannot be categorized, it's best not to put any type tag, and to |
| 718 | only use a risk or complexity information only as below. This is commonly the |
| 719 | case for new features, which development versions are mostly made of. |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 720 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 721 | The importance, complexity of the patch, or severity of the bug it fixes must |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 722 | be indicated when relevant. A single upper-case word is preferred, among : |
| 723 | |
| 724 | - MINOR minor change, very low risk of impact. It is often the case for |
| 725 | code additions that don't touch live code. As a rule of thumb, a |
| 726 | patch tagged "MINOR" is safe enough to be backported to stable |
| 727 | branches. For a bug, it generally indicates an annoyance, nothing |
| 728 | more. |
| 729 | |
| 730 | - MEDIUM medium risk, may cause unexpected regressions of low importance or |
| 731 | which may quickly be discovered. In short, the patch is safe but |
| 732 | touches working areas and it is always possible that you missed |
| 733 | something you didn't know existed (eg: adding a "case" entry or |
| 734 | an error message after adding an error code to an enum). For a bug, |
| 735 | it generally indicates something odd which requires changing the |
| 736 | configuration in an undesired way to work around the issue. |
| 737 | |
| 738 | - MAJOR major risk of hidden regression. This happens when large parts of |
| 739 | the code are rearranged, when new timeouts are introduced, when |
| 740 | sensitive parts of the session scheduling are touched, etc... We |
| 741 | should only exceptionally find such patches in stable branches when |
| 742 | there is no other option to fix a design issue. For a bug, it |
| 743 | indicates severe reliability issues for which workarounds are |
| 744 | identified with or without performance impacts. |
| 745 | |
| 746 | - CRITICAL medium-term reliability or security is at risk and workarounds, |
| 747 | if they exist, might not always be acceptable. An upgrade is |
| 748 | absolutely required. A maintenance release may be emitted even if |
| 749 | only one of these bugs are fixed. Note that this tag is only used |
| 750 | with bugs. Such patches must indicate what is the first version |
| 751 | affected, and if known, the commit ID which introduced the issue. |
| 752 | |
| 753 | The expected length of the commit message grows with the importance of the |
| 754 | change. While a MINOR patch may sometimes be described in 1 or 2 lines, MAJOR |
| 755 | or CRITICAL patches cannot have less than 10-15 lines to describe exactly the |
| 756 | impacts otherwise the submitter's work will be considered as rough sabotage. |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 757 | If you are sending a new patch series after a review, it is generally good to |
| 758 | enumerate at the end of the commit description what changed from the previous |
| 759 | one as it helps reviewers quickly glance over such changes and not re-read the |
| 760 | rest. |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 761 | |
| 762 | For BUILD, DOC and CLEANUP types, this tag is not always relevant and may be |
| 763 | omitted. |
| 764 | |
| 765 | The area the patch applies to is quite important, because some areas are known |
| 766 | to be similar in older versions, suggesting a backport might be desirable, and |
| 767 | conversely, some areas are known to be specific to one version. The area is a |
| 768 | single-word lowercase name the contributor find clear enough to describe what |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 769 | part is being touched. The following list of tags is suggested but not |
| 770 | exhaustive: |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 771 | |
| 772 | - examples example files. Be careful, sometimes these files are packaged. |
| 773 | |
| 774 | - tests regression test files. No code is affected, no need to upgrade. |
| 775 | |
Frédéric Lécaille | 4891e40 | 2018-06-19 14:06:07 +0200 | [diff] [blame] | 776 | - reg-tests regression test files for varnishtest. No code is affected, no |
| 777 | need to upgrade. |
| 778 | |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 779 | - init initialization code, arguments parsing, etc... |
| 780 | |
| 781 | - config configuration parser, mostly used when adding new config keywords |
| 782 | |
| 783 | - http the HTTP engine |
| 784 | |
| 785 | - stats the stats reporting engine |
| 786 | |
| 787 | - cli the stats socket CLI |
| 788 | |
| 789 | - checks the health checks engine (eg: when adding new checks) |
| 790 | |
| 791 | - sample the sample fetch system (new fetch or converter functions) |
| 792 | |
| 793 | - acl the ACL processing core or some ACLs from other areas |
| 794 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 795 | - filters everything related to the filters core |
| 796 | |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 797 | - peers the peer synchronization engine |
| 798 | |
| 799 | - lua the Lua scripting engine |
| 800 | |
| 801 | - listeners everything related to incoming connection settings |
| 802 | |
| 803 | - frontend everything related to incoming connection processing |
| 804 | |
| 805 | - backend everything related to LB algorithms and server farm |
| 806 | |
| 807 | - session session processing and flags (very sensible, be careful) |
| 808 | |
| 809 | - server server connection management, queueing |
| 810 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 811 | - spoe SPOE code |
| 812 | |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 813 | - ssl the SSL/TLS interface |
| 814 | |
| 815 | - proxy proxy maintenance (start/stop) |
| 816 | |
| 817 | - log log management |
| 818 | |
| 819 | - poll any of the pollers |
| 820 | |
| 821 | - halog the halog sub-component in the contrib directory |
| 822 | |
| 823 | - contrib any addition to the contrib directory |
| 824 | |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 825 | - htx general HTX subsystem |
| 826 | |
| 827 | - mux-h1 HTTP/1.x multiplexer/demultiplexer |
| 828 | |
| 829 | - mux-h2 HTTP/2 multiplexer/demultiplexer |
| 830 | |
| 831 | - h1 general HTTP/1.x protocol parser |
| 832 | |
| 833 | - h2 general HTTP/2 protocol parser |
| 834 | |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 835 | Other names may be invented when more precise indications are meaningful, for |
| 836 | instance : "cookie" which indicates cookie processing in the HTTP core. Last, |
| 837 | indicating the name of the affected file is also a good way to quickly spot |
| 838 | changes. Many commits were already tagged with "stream_sock" or "cfgparse" for |
| 839 | instance. |
| 840 | |
| 841 | It is required that the type of change and the severity when relevant are |
| 842 | indicated, as well as the touched area when relevant as well in the patch |
| 843 | subject. Normally, we would have the 3 most often. The two first criteria should |
| 844 | be present before a first colon (':'). If both are present, then they should be |
| 845 | delimited with a slash ('/'). The 3rd criterion (area) should appear next, also |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 846 | followed by a colon. Thus, all of the following subject lines are valid : |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 847 | |
Willy Tarreau | 138544f | 2017-03-31 16:24:44 +0200 | [diff] [blame] | 848 | Examples of subject lines : |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 849 | - DOC: document options forwardfor to logasap |
| 850 | - DOC/MAJOR: reorganize the whole document and change indenting |
| 851 | - BUG: stats: connection reset counters must be plain ascii, not HTML |
| 852 | - BUG/MINOR: stats: connection reset counters must be plain ascii, not HTML |
| 853 | - MEDIUM: checks: support multi-packet health check responses |
| 854 | - RELEASE: Released version 1.4.2 |
| 855 | - BUILD: stats: stdint is not present on solaris |
| 856 | - OPTIM/MINOR: halog: make fgets parse more bytes by blocks |
| 857 | - REORG/MEDIUM: move syscall redefinition to specific places |
| 858 | |
| 859 | Please do not use square brackets anymore around the tags, because they induce |
| 860 | more work when merging patches, which need to be hand-edited not to lose the |
| 861 | enclosed part. |
| 862 | |
| 863 | In fact, one of the only square bracket tags that still makes sense is '[RFC]' |
| 864 | at the beginning of the subject, when you're asking for someone to review your |
| 865 | change before getting it merged. If the patch is OK to be merged, then it can |
| 866 | be merge as-is and the '[RFC]' tag will automatically be removed. If you don't |
| 867 | want it to be merged at all, you can simply state it in the message, or use an |
| 868 | alternate 'WIP/' prefix in front of your tag tag ("work in progress"). |
| 869 | |
| 870 | The tags are not rigid, follow your intuition first, and they may be readjusted |
| 871 | when your patch is merged. It may happen that a same patch has a different tag |
| 872 | in two distinct branches. The reason is that a bug in one branch may just be a |
| 873 | cleanup or safety measure in the other one because the code cannot be triggered. |
| 874 | |
| 875 | |
| 876 | Working with Git |
| 877 | ---------------- |
| 878 | |
| 879 | For a more efficient interaction between the mainline code and your code, you |
| 880 | are strongly encouraged to try the Git version control system : |
| 881 | |
| 882 | http://git-scm.com/ |
| 883 | |
| 884 | It's very fast, lightweight and lets you undo/redo your work as often as you |
| 885 | want, without making your mistakes visible to the rest of the world. It will |
| 886 | definitely help you contribute quality code and take other people's feedback |
| 887 | in consideration. In order to clone the HAProxy Git repository : |
| 888 | |
| 889 | $ git clone http://git.haproxy.org/git/haproxy.git/ (development) |
| 890 | |
| 891 | If you decide to use Git for your developments, then your commit messages will |
| 892 | have the subject line in the format described above, then the whole description |
| 893 | of your work (mainly why you did it) will be in the body. You can directly send |
| 894 | your commits to the mailing list, the format is convenient to read and process. |
| 895 | |
| 896 | It is recommended to create a branch for your work that is based on the master |
| 897 | branch : |
| 898 | |
| 899 | $ git checkout -b 20150920-fix-stats master |
| 900 | |
| 901 | You can then do your work and even experiment with multiple alternatives if you |
| 902 | are not completely sure that your solution is the best one : |
| 903 | |
| 904 | $ git checkout -b 20150920-fix-stats-v2 |
| 905 | |
| 906 | Then reorder/merge/edit your patches : |
| 907 | |
| 908 | $ git rebase -i master |
| 909 | |
| 910 | When you think you're ready, reread your whole patchset to ensure there is no |
Tim Düsterhus | 4896c44 | 2016-11-29 02:15:19 +0100 | [diff] [blame] | 911 | formatting or style issue : |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 912 | |
| 913 | $ git show master.. |
| 914 | |
| 915 | And once you're satisfied, you should update your master branch to be sure that |
Thiago Farina | 9f72a39 | 2016-04-01 16:43:50 -0300 | [diff] [blame] | 916 | nothing changed during your work (only needed if you left it unattended for days |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 917 | or weeks) : |
| 918 | |
| 919 | $ git checkout -b 20150920-fix-stats-rebased |
| 920 | $ git fetch origin master:master |
| 921 | $ git rebase master |
| 922 | |
Thiago Farina | 9f72a39 | 2016-04-01 16:43:50 -0300 | [diff] [blame] | 923 | You can build a list of patches ready for submission like this : |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 924 | |
| 925 | $ git format-patch master |
| 926 | |
| 927 | The output files are the patches ready to be sent over e-mail, either via a |
| 928 | regular e-mail or via git send-email (carefully check the man page). Don't |
| 929 | destroy your other work branches until your patches get merged, it may happen |
| 930 | that earlier designs will be preferred for various reasons. Patches should be |
| 931 | sent to the mailing list : haproxy@formilux.org and CCed to relevant subsystem |
| 932 | maintainers or authors of the modified files if their address appears at the |
| 933 | top of the file. |
| 934 | |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 935 | Please don't send pull requests, they are really inconvenient as they make it |
| 936 | much more complicate to perform minor adjustments, and nobody benefits from |
| 937 | any comment on the code while on a list all subscribers learn a little bit on |
| 938 | each review of anyone else's code. |
| 939 | |
| 940 | |
| 941 | What to do if your patch is ignored |
| 942 | ----------------------------------- |
| 943 | |
| 944 | All patches merged are acknowledged by the maintainer who picked it. If you |
| 945 | didn't get an acknowledgement, check the mailing list archives to see if your |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 946 | mail was properly delivered there and possibly if anyone responded and you did |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 947 | not get their response (please look at http://haproxy.org/ for the mailing list |
| 948 | archive's address). |
| 949 | |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 950 | If you see that your mail is there but nobody responded, please recheck: |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 951 | - was the subject clearly indicating that it was a patch and/or that you were |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 952 | seeking some review? |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 953 | |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 954 | - was your email mangled by your mail agent? If so it's possible that |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 955 | nobody had the willingness yet to mention it. |
| 956 | |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 957 | - was your email sent as HTML? If so it definitely ended in spam boxes |
| 958 | regardless of the archives. |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 959 | |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 960 | - did the patch violate some of the principles explained in this document? |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 961 | |
| 962 | If none of these cases matches, it might simply be that everyone was busy when |
| 963 | your patch was sent and that it was overlooked. In this case it's fine to |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 964 | either resubmit it or respond to your own email asking if anything's wrong |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 965 | about it. In general don't expect a response after one week of silence, just |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 966 | because your email will not appear in anyone else's current window. So after |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 967 | one week it's time to resubmit. |
| 968 | |
| 969 | Among the mistakes that tend to make reviewers not respond are those who send |
| 970 | multiple versions of a patch in a row. It's natural for others then to wait for |
| 971 | the series to stabilize. And once it doesn't move anymore everyone forgot about |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 972 | it. As a rule of thumb, if you have to update your original email more than |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 973 | twice, first double-check that your series is really ready for submission, and |
| 974 | second, start a new thread and stop responding to the previous one. In this |
| 975 | case it is well appreciated to mention a version of your patch set in the |
| 976 | subject such as "[PATCH v2]", so that reviewers can immediately spot the new |
| 977 | version and not waste their time on the old one. |
| 978 | |
| 979 | If you still do not receive any response, it is possible that you've already |
| 980 | played your last card by not respecting the basic principles multiple times |
| 981 | despite being told about it several times, and that nobody is willing to spend |
| 982 | more of their time than normally needed with your work anymore. Your best |
| 983 | option at this point probably is to ask "did I do something wrong" than to |
| 984 | resend the same patches. |
| 985 | |
| 986 | |
| 987 | How to be sure to irritate everyone |
| 988 | ----------------------------------- |
| 989 | |
| 990 | Among the best ways to quickly lose everyone's respect, there is this small |
| 991 | selection, which should help you improve the way you work with others, if |
Tim Duesterhus | 2847f14 | 2019-06-15 19:47:29 +0200 | [diff] [blame] | 992 | you notice you're already practising some of them: |
Ilya Shipitsin | 2a950d0 | 2020-03-06 13:07:38 +0500 | [diff] [blame^] | 993 | - repeatedly send improperly formatted commit messages, with no type or |
Willy Tarreau | 09e0d74 | 2019-06-15 17:15:12 +0200 | [diff] [blame] | 994 | severity, or with no commit message body. These ones require manual |
| 995 | edition, maintainers will quickly learn to recognize your name. |
| 996 | |
| 997 | - repeatedly send patches which break something, and disappear or take a long |
| 998 | time to provide a fix. |
| 999 | |
| 1000 | - fail to respond to questions related to features you have contributed in |
| 1001 | the past, which can further lead to the feature being declared unmaintained |
| 1002 | and removed in a future version. |
| 1003 | |
| 1004 | - send a new patch iteration without taking *all* comments from previous |
| 1005 | review into consideration, so that the reviewer discovers he/she has to do |
| 1006 | the exact same work again. |
| 1007 | |
| 1008 | - "hijack" an existing thread to discuss something different or promote your |
| 1009 | work. This will generally make you look like a fool so that everyone wants |
| 1010 | to stay away from your e-mails. |
| 1011 | |
| 1012 | - continue to send pull requests after having been explained why they are not |
| 1013 | welcome. |
| 1014 | |
| 1015 | - give wrong advices to people asking for help, or sending them patches to |
| 1016 | try which make no sense, waste their time, and give them a bad impression |
| 1017 | of the people working on the project. |
| 1018 | |
| 1019 | - be disrespectful to anyone asking for help or contributing some work. This |
| 1020 | may actually even get you kicked out of the list and banned from it. |
Willy Tarreau | 11e334d9 | 2015-09-20 22:31:42 +0200 | [diff] [blame] | 1021 | |
| 1022 | -- end |