coreboot August 2016

coreboot@coreboot.org

53 participants
132 discussions

Re: [coreboot] [RFC] Deciding on style for multi-line comments
by Julius Werner 27 Aug '16

27 Aug '16

>> But I'm really with Linus when it comes to the leading, single asterisk. >> It looks totally weird, IMHO >> >> /* A small, concise comment >> that doesn't fit a single line */ >> >> is easier for the eye than >> >> /* A small, concise comment >> * that doesn't fit a single line */ >> >> where your eyes somehow stick on the asterisk first and then have to >> search for the real start of the line. FWIW I don't care about the difference between the two, I'd be happy to switch to the former form if we decide that's what we want to do. I only care about having full extra lines in there. We've currently already merged --ignore=BLOCK_COMMENT_STYLE, so all of these forms get accepted by Gerrit right now. I don't think a tool that automatically reformats comments would work that well (that would probably end up causing other unintended side effects), and I don't think we should set a hard line limit to decide between comments either. (For example, in a header file where you're describing the API for every function prototype you'd probably want all of those to use the same style, even if some of them are longer and some shorter than the hardcoded limit.) I think we should decide on the two styles of comments we want to allow, and then we can either submit a patch to checkpatch (I'd hope they'd accept it if it adds a configurable switch, such as --comment-styles=blank_line_top_bottom,short_balanced_stars) or enforce it through a separate script in util/lint/. And update the Wiki page, of course.

1 0

Re: [coreboot] [RFC] Deciding on style for multi-line comments
by Nico Huber 26 Aug '16

26 Aug '16

On 26.08.2016 17:56, Vadim Bendebury wrote: > I actually tend to agree with Julius that it does not make sense to waste 4 > lines for a two line comment. So, ideally the tool should be enforcing the > verbose style for comments longer than say 2 lines. Well, I too prefer the concise style for shorter comments. But I wouldn't enforce a number of lines. Maybe just say it's allowed for single-paragraph comments? And thanks, Julius, for bringing up the readability. This is what we should focus on. Actually I think a verbose-style comment shouldn't be allowed between code lines. That's something for longer explanations like a function description. But I'm really with Linus when it comes to the leading, single asterisk. It looks totally weird, IMHO /* A small, concise comment that doesn't fit a single line */ is easier for the eye than /* A small, concise comment * that doesn't fit a single line */ where your eyes somehow stick on the asterisk first and then have to search for the real start of the line. I wouldn't go as far as changing current code, but enforcing a style (we can agree on) for new code might prevent future discussions. Nico

2 1

[RFC] Deciding on style for multi-line comments
by Paul Menzel 26 Aug '16

26 Aug '16

Dear coreboot folks, The coding style currently demands the following style of multi-line comments [1]. > The preferred style for long (multi-line) comments is: /* * This is the preferred style for multi-line * comments in the Linux kernel source code. * Please use it consistently. * * Description: A column of asterisks on the left side, * with beginning and ending almost-blank lines. */ This is straight from the Linux Kernel coding style [2]. Certain parts of the code do not follow this style, so the question is how to deal with this. There has been some discussion on Gerrit about that [3]. But the list is the forum for such discussions, so I am bringing it up here. Julius’ last comment: > No offense, but that part of the Wiki literally reads: > > > For files in net/ and drivers/net/ the preferred style for long > > (multi-line) comments is a little different. > > ...so I'm not really sure why we should take something that has > obviously been carelessly bulk-copy&pasted into there from Linux > kernel sources eons ago as more authoritative than living development > practice of the last few years. > > The coreboot wiki is, sorry I have to say it, for the most part > pretty awful and outdated. It would probably be a good thing to fix > if somebody has the time for it, but until then I don't think we > should put too much stake into it (at least in the stuff that hasn't > been updated and maintained recently). I think that the extra blank lines in multi-line comments make them stand out better, which I prefer over having more lines on the screen. Also staying close to the Linux Kernel coding style makes it easier to use their tools, and not having to adapt them, and people only have to remember one style. But in the end it’s a matter of taste. So what should be done? Adapt the coreboot coding syle, or gradually change the style in the existing code, but require the style in new commits? Thanks, Paul [1] https://www.coreboot.org/Coding_Style#Commenting [2] https://www.kernel.org/doc/Documentation/CodingStyle [3] https://review.coreboot.org/16060

3 2

Re: [coreboot] [RFC] Deciding on style for multi-line comments
by Matt DeVillier 26 Aug '16

26 Aug '16

agree w/Julius as well. > 2 lines use verbose seems like a reasonable rule. On Fri, Aug 26, 2016 at 10:56 AM, Vadim Bendebury <vbendeb(a)chromium.org> wrote: > I actually tend to agree with Julius that it does not make sense to waste > 4 lines for a two line comment. So, ideally the tool should be enforcing > the verbose style for comments longer than say 2 lines. > > --vb > > On Fri, Aug 26, 2016 at 8:48 AM, Martin Roth <gaumless(a)gmail.com> wrote: > >> Can we please just decide on some tool and setting that we can run all >> the changes through that will "correctly" format it so we can stop arguing >> about the format? >> >> >> Martin >> >> On Thu, Aug 25, 2016 at 12:59 PM, Julius Werner <jwerner(a)chromium.org> >> wrote: >> >>> First of all, let's dispel the notion that there was one "correct" and >>> one "wrong" comment style in coreboot right now. >>> >>> If we count the amount of multi-line comments in the concise style >>> >>> > coreboot $ egrep '^\s*/\* [^/]+$' src/ | wc >>> > 5230 50584 491997 >>> >>> against the amount of multi-line comments in the verbose style >>> >>> > coreboot $ egrep '^\s*/\*$' src/ | wc >>> > 16103 18607 827285 >>> >>> and subtract the counts from the FSF license header from that, which >>> uses the verbose style and is just copy&pasted into every file >>> >>> > coreboot $ grep 'FITNESS FOR A PARTICULAR PURPOSE' src/ | wc >>> > 11857 133958 1424055 >>> >>> we find that the concise style is actually more common in coreboot >>> right now. I find the idea of overruling thousands of instances of >>> established practice on behalf of some ancient Wiki text that the >>> author obviously didn't even read before copy&pasting it in there >>> kinda ridiculous. >>> >>> Secondly, Linux cannot even decide for itself which comment style they >>> prefer (as exemplified by their net/ directory exemption, which they >>> have maintained for years despite the fact that enforcing different >>> styles in different parts of the repo is truly crazy), so I don't >>> think "do what Linux does because Linux is obviously always right" >>> makes much sense here either. As far as I can tell, they also don't >>> have any reasonable justification for their decision other than "it's >>> what Linus (and the net/ maintainer) like". If we really wanted to >>> blindly follow the Linux way here, we'd have to pick one directory >>> under coreboot/src/ that arbitrarily enforces a different style than >>> the others for no reason whatsoever. >>> >>> Look, I'm not saying that we should only be using the concise style. >>> I'm just saying that both styles have their valid uses and we >>> shouldn't be forcing one over the other in every possible case. Yes, >>> the verbose style separates comments and makes them stand out better, >>> and sometimes that's a good and desirable thing! If you want to put in >>> a big explanation that applies to a whole code block or something, it >>> may well be the better choice. >>> >>> But in other cases you maybe just want to explain a tricky detail >>> about a single statement in a way that you can't quite fit it on one >>> line, and forcing people to waste four lines and disrupt the whole >>> flow of the reader at a point where this wouldn't make sense is not a >>> good decision in that case. It would be similar to decreeing something >>> like "there may be at most three statements in a row without a blank >>> line in between them"... yes, blank lines often increase readability >>> and are encouraged by our coding style, but only in places that make >>> sense in the context of the code. Screen real estate is a real thing >>> and can make an already complicated piece of code way harder to >>> understand because you can't fit it all on one screen anymore. We'd be >>> actively making code less readable by forcing such an inflexible >>> style. >>> >>> All I'm saying is that the programmer writing the comment and the >>> surrounding code probably has a better idea about which amount of >>> blank space would make it most readable and understandable than some >>> outdated Wiki page, so this decision should be left to them. We're >>> writing comments with the explicit intention of helping others >>> understand our code, after all, so we're naturally going to write them >>> with the style that works best for each individual case. Forcing the >>> verbose style everywhere would just cause people to unnecessarily >>> shorten comments to one line when a big 4+ multi-line comment would be >>> too disruptive to readability, and thus decrease the overall quality >>> of our comments. >>> >>> -- >>> coreboot mailing list: coreboot(a)coreboot.org >>> https://www.coreboot.org/mailman/listinfo/coreboot >>> >> >> >> -- >> coreboot mailing list: coreboot(a)coreboot.org >> https://www.coreboot.org/mailman/listinfo/coreboot >> > > > -- > coreboot mailing list: coreboot(a)coreboot.org > https://www.coreboot.org/mailman/listinfo/coreboot >

1 0

Re: [coreboot] [RFC] Deciding on style for multi-line comments
by Martin Roth 26 Aug '16

26 Aug '16

And I don't really care either way. Let's either update the documentation or follow it though. If we have coding standards, let's follow them. If we feel like the individual programmer knows best for everything, let's SAY that so I can stop trying to fix whitespace in the reviews. On Fri, Aug 26, 2016 at 9:56 AM, Vadim Bendebury <vbendeb(a)chromium.org> wrote: > I actually tend to agree with Julius that it does not make sense to waste > 4 lines for a two line comment. So, ideally the tool should be enforcing > the verbose style for comments longer than say 2 lines. > > --vb > > On Fri, Aug 26, 2016 at 8:48 AM, Martin Roth <gaumless(a)gmail.com> wrote: > >> Can we please just decide on some tool and setting that we can run all >> the changes through that will "correctly" format it so we can stop arguing >> about the format? >> >> >> Martin >> >> On Thu, Aug 25, 2016 at 12:59 PM, Julius Werner <jwerner(a)chromium.org> >> wrote: >> >>> First of all, let's dispel the notion that there was one "correct" and >>> one "wrong" comment style in coreboot right now. >>> >>> If we count the amount of multi-line comments in the concise style >>> >>> > coreboot $ egrep '^\s*/\* [^/]+$' src/ | wc >>> > 5230 50584 491997 >>> >>> against the amount of multi-line comments in the verbose style >>> >>> > coreboot $ egrep '^\s*/\*$' src/ | wc >>> > 16103 18607 827285 >>> >>> and subtract the counts from the FSF license header from that, which >>> uses the verbose style and is just copy&pasted into every file >>> >>> > coreboot $ grep 'FITNESS FOR A PARTICULAR PURPOSE' src/ | wc >>> > 11857 133958 1424055 >>> >>> we find that the concise style is actually more common in coreboot >>> right now. I find the idea of overruling thousands of instances of >>> established practice on behalf of some ancient Wiki text that the >>> author obviously didn't even read before copy&pasting it in there >>> kinda ridiculous. >>> >>> Secondly, Linux cannot even decide for itself which comment style they >>> prefer (as exemplified by their net/ directory exemption, which they >>> have maintained for years despite the fact that enforcing different >>> styles in different parts of the repo is truly crazy), so I don't >>> think "do what Linux does because Linux is obviously always right" >>> makes much sense here either. As far as I can tell, they also don't >>> have any reasonable justification for their decision other than "it's >>> what Linus (and the net/ maintainer) like". If we really wanted to >>> blindly follow the Linux way here, we'd have to pick one directory >>> under coreboot/src/ that arbitrarily enforces a different style than >>> the others for no reason whatsoever. >>> >>> Look, I'm not saying that we should only be using the concise style. >>> I'm just saying that both styles have their valid uses and we >>> shouldn't be forcing one over the other in every possible case. Yes, >>> the verbose style separates comments and makes them stand out better, >>> and sometimes that's a good and desirable thing! If you want to put in >>> a big explanation that applies to a whole code block or something, it >>> may well be the better choice. >>> >>> But in other cases you maybe just want to explain a tricky detail >>> about a single statement in a way that you can't quite fit it on one >>> line, and forcing people to waste four lines and disrupt the whole >>> flow of the reader at a point where this wouldn't make sense is not a >>> good decision in that case. It would be similar to decreeing something >>> like "there may be at most three statements in a row without a blank >>> line in between them"... yes, blank lines often increase readability >>> and are encouraged by our coding style, but only in places that make >>> sense in the context of the code. Screen real estate is a real thing >>> and can make an already complicated piece of code way harder to >>> understand because you can't fit it all on one screen anymore. We'd be >>> actively making code less readable by forcing such an inflexible >>> style. >>> >>> All I'm saying is that the programmer writing the comment and the >>> surrounding code probably has a better idea about which amount of >>> blank space would make it most readable and understandable than some >>> outdated Wiki page, so this decision should be left to them. We're >>> writing comments with the explicit intention of helping others >>> understand our code, after all, so we're naturally going to write them >>> with the style that works best for each individual case. Forcing the >>> verbose style everywhere would just cause people to unnecessarily >>> shorten comments to one line when a big 4+ multi-line comment would be >>> too disruptive to readability, and thus decrease the overall quality >>> of our comments. >>> >>> -- >>> coreboot mailing list: coreboot(a)coreboot.org >>> https://www.coreboot.org/mailman/listinfo/coreboot >>> >> >> >> -- >> coreboot mailing list: coreboot(a)coreboot.org >> https://www.coreboot.org/mailman/listinfo/coreboot >> > >

1 0

Re: [coreboot] [RFC] Deciding on style for multi-line comments
by Vadim Bendebury 26 Aug '16

26 Aug '16

I actually tend to agree with Julius that it does not make sense to waste 4 lines for a two line comment. So, ideally the tool should be enforcing the verbose style for comments longer than say 2 lines. --vb On Fri, Aug 26, 2016 at 8:48 AM, Martin Roth <gaumless(a)gmail.com> wrote: > Can we please just decide on some tool and setting that we can run all the > changes through that will "correctly" format it so we can stop arguing > about the format? > > > Martin > > On Thu, Aug 25, 2016 at 12:59 PM, Julius Werner <jwerner(a)chromium.org> > wrote: > >> First of all, let's dispel the notion that there was one "correct" and >> one "wrong" comment style in coreboot right now. >> >> If we count the amount of multi-line comments in the concise style >> >> > coreboot $ egrep '^\s*/\* [^/]+$' src/ | wc >> > 5230 50584 491997 >> >> against the amount of multi-line comments in the verbose style >> >> > coreboot $ egrep '^\s*/\*$' src/ | wc >> > 16103 18607 827285 >> >> and subtract the counts from the FSF license header from that, which >> uses the verbose style and is just copy&pasted into every file >> >> > coreboot $ grep 'FITNESS FOR A PARTICULAR PURPOSE' src/ | wc >> > 11857 133958 1424055 >> >> we find that the concise style is actually more common in coreboot >> right now. I find the idea of overruling thousands of instances of >> established practice on behalf of some ancient Wiki text that the >> author obviously didn't even read before copy&pasting it in there >> kinda ridiculous. >> >> Secondly, Linux cannot even decide for itself which comment style they >> prefer (as exemplified by their net/ directory exemption, which they >> have maintained for years despite the fact that enforcing different >> styles in different parts of the repo is truly crazy), so I don't >> think "do what Linux does because Linux is obviously always right" >> makes much sense here either. As far as I can tell, they also don't >> have any reasonable justification for their decision other than "it's >> what Linus (and the net/ maintainer) like". If we really wanted to >> blindly follow the Linux way here, we'd have to pick one directory >> under coreboot/src/ that arbitrarily enforces a different style than >> the others for no reason whatsoever. >> >> Look, I'm not saying that we should only be using the concise style. >> I'm just saying that both styles have their valid uses and we >> shouldn't be forcing one over the other in every possible case. Yes, >> the verbose style separates comments and makes them stand out better, >> and sometimes that's a good and desirable thing! If you want to put in >> a big explanation that applies to a whole code block or something, it >> may well be the better choice. >> >> But in other cases you maybe just want to explain a tricky detail >> about a single statement in a way that you can't quite fit it on one >> line, and forcing people to waste four lines and disrupt the whole >> flow of the reader at a point where this wouldn't make sense is not a >> good decision in that case. It would be similar to decreeing something >> like "there may be at most three statements in a row without a blank >> line in between them"... yes, blank lines often increase readability >> and are encouraged by our coding style, but only in places that make >> sense in the context of the code. Screen real estate is a real thing >> and can make an already complicated piece of code way harder to >> understand because you can't fit it all on one screen anymore. We'd be >> actively making code less readable by forcing such an inflexible >> style. >> >> All I'm saying is that the programmer writing the comment and the >> surrounding code probably has a better idea about which amount of >> blank space would make it most readable and understandable than some >> outdated Wiki page, so this decision should be left to them. We're >> writing comments with the explicit intention of helping others >> understand our code, after all, so we're naturally going to write them >> with the style that works best for each individual case. Forcing the >> verbose style everywhere would just cause people to unnecessarily >> shorten comments to one line when a big 4+ multi-line comment would be >> too disruptive to readability, and thus decrease the overall quality >> of our comments. >> >> -- >> coreboot mailing list: coreboot(a)coreboot.org >> https://www.coreboot.org/mailman/listinfo/coreboot >> > > > -- > coreboot mailing list: coreboot(a)coreboot.org > https://www.coreboot.org/mailman/listinfo/coreboot >

1 0

Re: [coreboot] [RFC] Deciding on style for multi-line comments
by Martin Roth 26 Aug '16

26 Aug '16

Can we please just decide on some tool and setting that we can run all the changes through that will "correctly" format it so we can stop arguing about the format? Martin On Thu, Aug 25, 2016 at 12:59 PM, Julius Werner <jwerner(a)chromium.org> wrote: > First of all, let's dispel the notion that there was one "correct" and > one "wrong" comment style in coreboot right now. > > If we count the amount of multi-line comments in the concise style > > > coreboot $ egrep '^\s*/\* [^/]+$' src/ | wc > > 5230 50584 491997 > > against the amount of multi-line comments in the verbose style > > > coreboot $ egrep '^\s*/\*$' src/ | wc > > 16103 18607 827285 > > and subtract the counts from the FSF license header from that, which > uses the verbose style and is just copy&pasted into every file > > > coreboot $ grep 'FITNESS FOR A PARTICULAR PURPOSE' src/ | wc > > 11857 133958 1424055 > > we find that the concise style is actually more common in coreboot > right now. I find the idea of overruling thousands of instances of > established practice on behalf of some ancient Wiki text that the > author obviously didn't even read before copy&pasting it in there > kinda ridiculous. > > Secondly, Linux cannot even decide for itself which comment style they > prefer (as exemplified by their net/ directory exemption, which they > have maintained for years despite the fact that enforcing different > styles in different parts of the repo is truly crazy), so I don't > think "do what Linux does because Linux is obviously always right" > makes much sense here either. As far as I can tell, they also don't > have any reasonable justification for their decision other than "it's > what Linus (and the net/ maintainer) like". If we really wanted to > blindly follow the Linux way here, we'd have to pick one directory > under coreboot/src/ that arbitrarily enforces a different style than > the others for no reason whatsoever. > > Look, I'm not saying that we should only be using the concise style. > I'm just saying that both styles have their valid uses and we > shouldn't be forcing one over the other in every possible case. Yes, > the verbose style separates comments and makes them stand out better, > and sometimes that's a good and desirable thing! If you want to put in > a big explanation that applies to a whole code block or something, it > may well be the better choice. > > But in other cases you maybe just want to explain a tricky detail > about a single statement in a way that you can't quite fit it on one > line, and forcing people to waste four lines and disrupt the whole > flow of the reader at a point where this wouldn't make sense is not a > good decision in that case. It would be similar to decreeing something > like "there may be at most three statements in a row without a blank > line in between them"... yes, blank lines often increase readability > and are encouraged by our coding style, but only in places that make > sense in the context of the code. Screen real estate is a real thing > and can make an already complicated piece of code way harder to > understand because you can't fit it all on one screen anymore. We'd be > actively making code less readable by forcing such an inflexible > style. > > All I'm saying is that the programmer writing the comment and the > surrounding code probably has a better idea about which amount of > blank space would make it most readable and understandable than some > outdated Wiki page, so this decision should be left to them. We're > writing comments with the explicit intention of helping others > understand our code, after all, so we're naturally going to write them > with the style that works best for each individual case. Forcing the > verbose style everywhere would just cause people to unnecessarily > shorten comments to one line when a big 4+ multi-line comment would be > too disruptive to readability, and thus decrease the overall quality > of our comments. > > -- > coreboot mailing list: coreboot(a)coreboot.org > https://www.coreboot.org/mailman/listinfo/coreboot >

1 0

New on blogs.coreboot.org: [GSoC] Multiple status registers, block protection and OTP support, wrap-up (1/2)
by WordPress 26 Aug '16

26 Aug '16

1 0

New Defects reported by Coverity Scan for coreboot
by scan-admin＠coverity.com 26 Aug '16

26 Aug '16

Hi, Please find the latest report on new defect(s) introduced to coreboot found with Coverity Scan. 2 new defect(s) introduced to coreboot found with Coverity Scan. New defect(s) Reported-by: Coverity Scan Showing 2 of 2 defect(s) ** CID 1361947: Code maintainability issues (UNUSED_VALUE) /src/arch/riscv/trap_handler.c: 37 in handle_supervisor_call() ________________________________________________________________________________________________________ *** CID 1361947: Code maintainability issues (UNUSED_VALUE) /src/arch/riscv/trap_handler.c: 37 in handle_supervisor_call() 31 printk(BIOS_DEBUG, "Getting hart id...\n"); 32 returnValue = read_csr(mhartid); 33 break; 34 case SBI_ECALL_NUM_HARTS: 35 /* TODO: parse the hardware-supplied config string and 36 return the correct value */ >>> CID 1361947: Code maintainability issues (UNUSED_VALUE) >>> Assigning value "1ULL" to "returnValue" here, but that stored value is overwritten before it can be used. 37 returnValue = 1; 38 case SBI_ECALL_CONSOLE_PUT: 39 returnValue = mcall_console_putchar(arg0); 40 break; 41 case SBI_ECALL_SEND_DEVICE_REQUEST: 42 printk(BIOS_DEBUG, "Sending device request...\n"); ** CID 1361946: Control flow issues (MISSING_BREAK) /src/arch/riscv/trap_handler.c: 38 in handle_supervisor_call() ________________________________________________________________________________________________________ *** CID 1361946: Control flow issues (MISSING_BREAK) /src/arch/riscv/trap_handler.c: 38 in handle_supervisor_call() 32 returnValue = read_csr(mhartid); 33 break; 34 case SBI_ECALL_NUM_HARTS: 35 /* TODO: parse the hardware-supplied config string and 36 return the correct value */ 37 returnValue = 1; >>> CID 1361946: Control flow issues (MISSING_BREAK) >>> The above case falls through to this one. 38 case SBI_ECALL_CONSOLE_PUT: 39 returnValue = mcall_console_putchar(arg0); 40 break; 41 case SBI_ECALL_SEND_DEVICE_REQUEST: 42 printk(BIOS_DEBUG, "Sending device request...\n"); 43 returnValue = mcall_dev_req((sbi_device_message*) arg0); ________________________________________________________________________________________________________ To view the defects in Coverity Scan visit, https://u2389337.ct.sendgrid.net/wf/click?upn=08onrYu34A-2BWcWUl-2F-2BfV0V0… To manage Coverity Scan email notifications for "coreboot(a)coreboot.org", click https://u2389337.ct.sendgrid.net/wf/click?upn=08onrYu34A-2BWcWUl-2F-2BfV0V0…

1 0

Re: [coreboot] [RFC] Deciding on style for multi-line comments
by Julius Werner 25 Aug '16

25 Aug '16

First of all, let's dispel the notion that there was one "correct" and one "wrong" comment style in coreboot right now. If we count the amount of multi-line comments in the concise style > coreboot $ egrep '^\s*/\* [^/]+$' src/ | wc > 5230 50584 491997 against the amount of multi-line comments in the verbose style > coreboot $ egrep '^\s*/\*$' src/ | wc > 16103 18607 827285 and subtract the counts from the FSF license header from that, which uses the verbose style and is just copy&pasted into every file > coreboot $ grep 'FITNESS FOR A PARTICULAR PURPOSE' src/ | wc > 11857 133958 1424055 we find that the concise style is actually more common in coreboot right now. I find the idea of overruling thousands of instances of established practice on behalf of some ancient Wiki text that the author obviously didn't even read before copy&pasting it in there kinda ridiculous. Secondly, Linux cannot even decide for itself which comment style they prefer (as exemplified by their net/ directory exemption, which they have maintained for years despite the fact that enforcing different styles in different parts of the repo is truly crazy), so I don't think "do what Linux does because Linux is obviously always right" makes much sense here either. As far as I can tell, they also don't have any reasonable justification for their decision other than "it's what Linus (and the net/ maintainer) like". If we really wanted to blindly follow the Linux way here, we'd have to pick one directory under coreboot/src/ that arbitrarily enforces a different style than the others for no reason whatsoever. Look, I'm not saying that we should only be using the concise style. I'm just saying that both styles have their valid uses and we shouldn't be forcing one over the other in every possible case. Yes, the verbose style separates comments and makes them stand out better, and sometimes that's a good and desirable thing! If you want to put in a big explanation that applies to a whole code block or something, it may well be the better choice. But in other cases you maybe just want to explain a tricky detail about a single statement in a way that you can't quite fit it on one line, and forcing people to waste four lines and disrupt the whole flow of the reader at a point where this wouldn't make sense is not a good decision in that case. It would be similar to decreeing something like "there may be at most three statements in a row without a blank line in between them"... yes, blank lines often increase readability and are encouraged by our coding style, but only in places that make sense in the context of the code. Screen real estate is a real thing and can make an already complicated piece of code way harder to understand because you can't fit it all on one screen anymore. We'd be actively making code less readable by forcing such an inflexible style. All I'm saying is that the programmer writing the comment and the surrounding code probably has a better idea about which amount of blank space would make it most readable and understandable than some outdated Wiki page, so this decision should be left to them. We're writing comments with the explicit intention of helping others understand our code, after all, so we're naturally going to write them with the style that works best for each individual case. Forcing the verbose style everywhere would just cause people to unnecessarily shorten comments to one line when a big 4+ multi-line comment would be too disruptive to readability, and thus decrease the overall quality of our comments.

1 0

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

coreboot August 2016