Drobe :: The archives
About Drobe | Contact | RSS | Twitter | Tech docs | Downloads | BBC Micro

GCCSDK team trumpets module support

By Chris Williams. Published: 9th Jul 2005, 20:41:50 | Permalink | Printable

Another piece in the jigsaw

GCC logoA new version of the GCCSDK released this weekend is said to include improved support for compiling RISC OS modules. As well as various improvements and critical bug fixes, GCC 3.4.4 release 3 can now successfully build modules, which are used by applications to extend the functionality of the underlying operating system.

Full module support in GCC for RISC OS arguably narrows the gap between itself and the commercial C/C++ compiler package Norcroft, which also supports module building. Programmers who employ GCC can now use the compiler package along with bundled utility CMunge to produce working modules.

Experimental module support first appeared in GCC 3.4.4 release 2, in May, for developers to test and provide feedback on. Peter Naulls, the GCCSDK contributor who wrapped together the GCC 3.4.4 release 3, dismissed inquiries on whether or not the module support had been through enough testing. He diplomatically quipped yesterday, "Spare me the patronising questions."

At the start of the month, John Tytgat, another leading GCCSDK contributor, listed six points of order that needed to be addressed before he felt module support could be considered tested and stable. He confirmed with us that these issues had been attended to in time for the third release of GCC 3.4.4.

He commented, "I honestly believe that GCC 3.4.4 release 3 code generation is pretty solid, especially when you know it is being used by Peter Naulls' Firefox build." He also pointed out that the NetSurf team use GCCSDK for building their home grown web browser and are quick to spot faults in the compiler.

John says module generation is "stable and reasonably tested", as various modules including iconv, RemotePrinterFS, Sunfish and JetDirectFS have been successfully created with GCC 3.4.4 release 3. Other modules, such as Memphis, CryptRandom and EtherY are expected to be successfully created too.

He added: "I think more RISC OS GCC specific documentation delivered with the GCCSDK kit couldn't hurt. The problem is that this information is available but spread over the mailing list and a couple of webpages which makes it difficult."

GCCSDK is a C/C++ compiler package for RISC OS that uses source code from the mainstream GNU Compiler Collection. It also supports other programming languages, including Fortran. Its project leader, Nick Burrett, recently quit from the development team although made progress in porting GCC 4.1 to RISC OS. The GCCSDK website urges developers to use GCC 3.4.4 release 3 for building software.

Links

GCCSDK for RISC OS

Previous: Work starts on DataPower 3
Next: MicroDigital sought by bailiffs

Discussion

Viewing threaded comments | View comments unthreaded, listed by date | Skip to the end

Sad that Chris needs to (inaccurately) report my response to _his_ paranoid questions, and went behind my back to ask John instead (not that John knew). Chris has in the past expressed considerable concerns, against all reason, about the state of GCCSDK.

Sad that he couldn't just stick to the facts, once again and also a shame that he's completely forgotten what it's like to be a developer and try and meet the requirements of an extremely demanding user base. None of this helps RISC OS.

RISC OS news reporting? More like Heat magazine.

 is a RISC OS Usermrchocky on 9/7/05 9:05PM
[ Reply | Permalink | Report ]

So moderating down my comment will make the issue go away? GCC(SDK) is an increasingly important part of RISC OS development. Trying to create fictional issues about it isn't going to help anyone. We went to a great deal of effort to ensure module support was correct after people asked for it for 6 years. A single positive comment wouldn't be out of order.

 is a RISC OS Usermrchocky on 9/7/05 9:10PM
[ Reply | Permalink | Report ]

Sounds more like the classic producer vs. consumer press-release-driven punditry of the Acorn User era rather than what I'd imagine Heat magazine to be like.

 is a RISC OS Userguestx on 9/7/05 9:57PM
[ Reply | Permalink | Report ]

More great work from John Tytgat, Peter Naulls et al. Thanks guys!

mrchocky:

What's inaccurate about the article?

 is a RISC OS Userfylfot on 9/7/05 10:51PM
[ Reply | Permalink | Report ]

Well done Peter and all other contributors! This is great news! :-)

 is a RISC OS UserJGZimmerle on 9/7/05 10:54PM
[ Reply | Permalink | Report ]

I agree this is great news, and my sincere thanks to all involved in the GCC project.

Incidentally, is there anywhere I can download a working copy of Sunfish compiled under GCC?

 is a RISC OS Usermd0u80c9 on 9/7/05 11:40PM
[ Reply | Permalink | Report ]

Even though I don't program myself, I surely do appreciate the work invested into development of GCC SDK recently and over the years! I believe this progress is strongly reflected in the appearance of programmes of which some would perhaps never been possible without these developments, such is the significance of these developments.

Why is mrchocky modded down when his comment is strongly On Topic and he is in a central position to comment on this?

 is a RISC OS UserhEgelia on 10/7/05 10:59AM
[ Reply | Permalink | Report ]

Sadly, this article does seem to have negative undertones which I can't see any justification for.

md0u80c9: In terms of functionality and speed the GCC compiled Sunfish is indistinguishable from the Norcroft compiled version. You can always check out the latest source and compile it yourself.

 is a RISC OS UserAJW on 10/7/05 12:00PM
[ Reply | Permalink | Report ]

What negative undertones?

It's great, it's GCC, it uses source from the main stream, it has module support, it's tested and stable, various modules have already be correctly buildwith it, etc.

The only things that could be found negative in the article is a answer not gotten (although the reply seems to mean "yes it's tested") and the project leader leaving.

The article itself isn't negative to me!

Then again I don't realy fit into the RISC OS scene as I'm not depressed all the time :)

 is a RISC OS UserJaco on 10/7/05 12:42PM
[ Reply | Permalink | Report ]

and now someting on topic:

Congratulations to the GCCSDK team!!!!!!!!!!!!!! Thanks for the hard work.

 is a RISC OS UserJaco on 10/7/05 12:46PM
[ Reply | Permalink | Report ]

Good news indeed. Full congratulations to both Peter Naulls and the GCCSDK team for a job well done.

I feel a download coming on ;)

 is a RISC OS UserAMS on 10/7/05 12:58PM
[ Reply | Permalink | Report ]

This is excellent stuff. I've always felt (rightly or wrongly) that writing modules was a rather inaccessible activity since one either needed to buy Norcroft or learn to write assembler. Now I'd love to see how it's actually done given I have the tools now. Many many thanks to the GCCSDK team! -- Spriteman (looking forward to getting back to some RISC OS programming soon)

 is a RISC OS UserSpriteman on 10/7/05 3:27PM
[ Reply | Permalink | Report ]

Congratulations to all involved. I guess mrchocky is paying the price of fame. ;(

 is a RISC OS Usersteelpillow on 10/7/05 10:12PM
[ Reply | Permalink | Report ]

This is great news. My thanks goes to the GCCSDK team. I like many others have been put off writing modules before now, it's nice to have some freely available tools for the job. I couldn't see anything wrong with the article.. It doesn't bode well though for drobe.co.uk if their headlining writers are having a public spat..

 is a RISC OS UserSnig on 11/7/05 11:54AM
[ Reply | Permalink | Report ]

Snig: you don't think it's wrong that I was quoted out of context for the clear purpose of inciting annoyance? Sticking to facts would be much more constructive.

One of the facts it gets quite wrong is the second paragraph, where it says that GCC "narrows the gap". Given that module support was the _only_ feature (that I'm aware of) that GCC previously didn't support and that Norcroft did, _and_ that GCC does loads of other things that Norcroft may never do, as well as generate significantly faster code, the statement is nonsense.

"widens the gap" would be more accurate, if not terribly helpful. And since compiler comparisons/benchmarks are potentially contentious, it would be much better to state things as they are.

Whether or not the article is amended to correct this and all its other indiscresctions is another matter. I'm willing to bet it won't be.

 is a RISC OS Usermrchocky on 11/7/05 12:16PM
[ Reply | Permalink | Report ]

mrchocky: The article says "arguably narrows the gap", which I believe is fair. Being the self assumed GCC project team leader, I would expect you to argue that GCC is more advanced than Norcorft. Conversely, I know people who feel Norcroft is superior to GCC. Hence the word "arguably".

You've also failed to email Drobe where there are other inaccuracies or indiscresctions in the article. If you do not tell me, I cannot correct the article, which I am always willing to do. The fact that you have not pointed out any specific errors and yet continue to mouth off about the article speaks volumes. Put up, or shut up.

snig: Peter has not been a headline writer for Drobe for a while now. See the 'About us' page.

 is a RISC OS Userdiomus on 11/7/05 12:33PM
[ Reply | Permalink | Report ]

No, the statement is completely nonsensical, it gives no information and is misleading. Norcroft may have the edge in things as such as speed, but the paragraph quite clearly refers to _features_ not opinions. At the very least, the paragraph requires some qualification, otherwise the picture is very skewed.

Why would I need to email drobe? Don't pretend you accidentally put in things to purposely wind me up. Don't pretend you didn't know about the features GCC offers. Don't try and shift the blame onto me because you didn't make an effort with your Cartmell style reporting. Emailing me is one thing, but your question had nothing to do with this point.

Also, as drobe editor, do you think it's appropriate to abuse your powers and moderate me down by so much?

 is a RISC OS Usermrchocky on 11/7/05 12:41PM
[ Reply | Permalink | Report ]

There's nothing negative about GCCSDK in that article (having just re-read it). The nearest thing is the debatable point in para 2 - but that's an opinion, not a factual error. It might be a wrong opinion but it has already been defended in these comments. If the "gap" doesn't exist or is indeed in the other direction, the best way to point that out is to do just that: point it out, don't start down an emotional and ad hominem route.

As for thing about quotations - if that is something you said in an interview, you can hardly complain if it's used in the article; that's why one must always be careful when talking to the Press. I can see nothing wrong with also interviewing another team member, and I can see no reason why Drobe should ask your permission to do so (as you imply they should do by your "behind my back" comment above).

AFAICS, you have taken offence at the second paragraph and the quotation they chose to use from you, and then taken the entire article negatively when there is little that can really be said to be negative in it. It seems to be a fair piece of reporting that talks GCCSDK up.

pax et bonum

 is a RISC OS Userjohnpettigrew on 11/7/05 12:55PM
[ Reply | Permalink | Report ]

Rah, rah, rah. It's easy to talk tough on the Internet, isn't it, Peter? In my view, as an impartial editor, the paragraph in question is fine as it stands although I'll consider writing a Norcroft vs. GCC article, if I get time. And as I suspected, you can't think of any other problems with the article so you were in fact, lying.

There was no attempt to wind you up, there has been no rigging of the moderation system, and you're just full of hot air. This is why people can't take what you say online seriously anymore.

 is a RISC OS Userdiomus on 11/7/05 12:56PM
[ Reply | Permalink | Report ]

mrchocky> Chris asked you about the GCC release, and you had ample opportunity to give him decent answers, however answers such as "Spare me the patronising questions." are hardly helpful in this respect, you should know from experience that that attitude only irritates people. I can certainly understand Chris' fustration when trying to get information from you, even if his questions can on occasion be poorly presented.

As to others moderating you down (and a fair few back up, it really is a tug-o-war as your privs let you see), your initial comment belongs in an email to the article writer (in this case Chris) and only in the comments system if you have exhausted that means of contacting him, which I believe you have not done.

Chris> The article is obviously incorrect in part, or at the very least *potentially* *misleading*, and this is causing problems, so I suggest you humble a bit and fix it.

 is a RISC OS Userpiemmm on 11/7/05 12:57PM
[ Reply | Permalink | Report ]

It's a shame you can't admit you might have made a mistake, rather than dissmising my claims with hand waving. I can assure you that what I'm saying is taken plenty seriously, despite your unfounded claim otherwise.

Saying I'm lying and using insults such as "full of hot air" is rather immature for a news editor. Sadly, no more or less so than that of the other RISC OS news portals.

John: Emotional? No, I'm sticking to facts. There was no excuse for not getting it right in the first place. This is precisely one of the qualities that drobe has prided itself on in the past, and now is sadly lacking.

I certainly _can_ complain about quotations if I please, especially given that it's out of context.

As I said, much earlier, there's little now to encourage people to make a real effort to improve RISC OS, despite the considerable technical achievements, when so much seems driven by selfishness - not just here, but on other forums too.

 is a RISC OS Usermrchocky on 11/7/05 1:07PM
[ Reply | Permalink | Report ]

Chris: I can think of one rather big problem with the article: saying "GCCSDK is a C/C++ compiler package for RISC OS" is wrong, misleading and incomplete. I'm also rather unhappy with the "narrowing the gap" phrase. I'm also not sure if it is sensible to ever ask a software developer if a piece of software is "tested enough".

 is a RISC OS Userhubersn on 11/7/05 1:12PM
[ Reply | Permalink | Report ]

pie: please don't trying accuse me of not giving information about when (a) I've been exceptinally open about all my work, and (b) you all too often ignore problems completely such by not repsonding to emails or enquiries.

This issue has already been discussed privately with Chris, who refused to take the matter seriously.

As I said, the quote was taken out of context, so it's no good saying I shouldn't have said it.

I accept that given my visibility, I'm going to get a fair bit of flak, regardless of what I do. But please don't expect me to be quiet when something is incorrect. RISC OS has deeply suffered for misinformation in the years immediately after Acorn, and prentending eveything will be hunky-dory in such situations is not ok.

 is a RISC OS Usermrchocky on 11/7/05 1:13PM
[ Reply | Permalink | Report ]

hubersn: ok, I've added a line that says it supports other languages, including Fortran. Saying the sentence is "wrong, misleading and incomplete" is a little strong, in my opinion, as the sentence in question is substantially correct. Again, the line about Norcroft and GCC is down to personal opinion, hence the word "arguably". It's like a StrongED vs. Zap flamewar.

As for the testing bit, I understand your point of view, and if I asked you that question and you answered with something like 'well, there's no such thing as fully tested, in the same way that nothing can be fully safe', then that's fine and I'd quote it. It's sad that Peter chose to interpret it as a wind up. It was an honest question.

mrchocky: Oh come on, take what you dish out. You've called people plenty of things worse than "full of hot air". The quote was not taken out of context.

 is a RISC OS Userdiomus on 11/7/05 1:25PM
[ Reply | Permalink | Report ]

Why articles/threads with Mr. Peter Naulls taking part are very similar to a battlefield given a fair amount of time?

And now mod me down, don't bother... if you can't accept my 2 cents (Euro cents!!) I don't see any reason to bother....

 is a RISC OS Userbernie on 11/7/05 1:37PM
[ Reply | Permalink | Report ]

diomus:"there has been no rigging of the moderation system" Yeah, right.

 is a RISC OS Useradamr on 11/7/05 1:59PM
[ Reply | Permalink | Report ]

what ever work is done on the compiler, surely must be good!

Well done to the team....I would help, but my VB coding gets this sort of remark "corr blimy! hacking that about a bit aint ya?" :@P so i would be no use lol.

let us not forget that this work that is being done, needs to be done....annoying developers is one sure way of killing the platform, dont you all agree?

 is a RISC OS Userem2ac on 11/7/05 2:07PM
[ Reply | Permalink | Report ]

Adamr> Feel free to click on your profile and see how your behaviour on drove has affected your ratings, and how other people have rated your comments. Not my fault if you do not appear to be capable of making constructive comments that the majority might actually moderate up for a change.

 is a RISC OS Userpiemmm on 11/7/05 2:08PM
[ Reply | Permalink | Report ]

Chris: saying "narrowing the gap" says that one product is more advanced than the other. For nearly all complex pieces of software, it is very unusual that one has a superset of capabilities of the other. Unsurprisingly, when comparing GCC with Norcroft, this is also not the case.

This is not a matter of personal opinion. It is not in any way arguable. The gap analogy is just wrong and misleading. What is written in the article is not equivalent to the StrongED vs. Zap situation.

 is a RISC OS Userhubersn on 11/7/05 2:51PM
[ Reply | Permalink | Report ]

piemmm: Oh, those stats are new! That is at least step in the right direction :-) (I was about to moan about 'Drobe doesn't like this user' being meaningless.)

As for me having a poor rating - I'm quite content with that.[1] What I object to is "unfair" interference in the comment rating system by members of the drobe team.

Adam

[1] For instance, this post is ludicrously off topic so I won't be very surprised if/when it gets modded down! That doesn't stop me wanting to put in my 2c though.

 is a RISC OS Useradamr on 11/7/05 2:55PM
[ Reply | Permalink | Report ]

Adamr> Thats what these forums are for, there's a hardcoded limit of 30 comments on the article, whereby when it hits that people have gone sufficiently off topic from the article to continue on in here without getting modded up/down/sideways.

I should put a 'x new posts' thing in the forum link on the sidebar though, to make things a bit more obvious when there is activity on there.

 is a RISC OS Userpiemmm on 11/07/05 2:57PM
[ Reply | Permalink | Report ]

I'll comment now then shall I? Apparrently because 14% of my comments were considered "bad" I can't post. My evil ways have destroyed the drobe website.

Anyway, it's obvious Chris is trying to be more like a "real" "reporter"

In other news drobe.co.uk rss bloglines subscriptions down 20%

Anyway, what does Acorn/Castle C provide now that GCC doesn't? Toolbox stuff?

 is a RISC OS Usermavhc on 15/07/05 11:45PM
[ Reply | Permalink | Report ]

Please login before posting a comment. Use the form on the right to do so or create a free account.

Search the archives

Today's featured article

  • Writing internationalised software
    John-Mark Bell dives into the nitty-gritty of Unicode and writing RISC OS software for users worldwide
     4 comments, latest by caliston2 on 13/4/05 2:04PM. Published: 13 Apr 2005

  • Random article

  • Olympia Linux Expo 2003
    Mark Stephens discovers RISC OS players aiming at a wider audience
     4 comments, latest by Nodoid on 25/11/03 8:32PM. Published: 8 Oct 2003

  • Useful links

    News and media:
    IconbarMyRISCOSArcSiteRISCOScodeANSC.S.A.AnnounceArchiveQercusRiscWorldDrag'n'DropGAG-News

    Top developers:
    RISCOS LtdRISC OS OpenMW SoftwareR-CompAdvantage SixVirtualAcorn

    Dealers:
    CJE MicrosAPDLCastlea4X-AmpleLiquid SiliconWebmonster

    Usergroups:
    WROCCRONENKACCIRUGSASAUGROUGOLRONWUGMUGWAUGGAGRISCOS.be

    Useful:
    RISCOS.org.ukRISCOS.orgRISCOS.infoFilebaseChris Why's Acorn/RISC OS collectionNetSurf

    Non-RISC OS:
    The RegisterThe InquirerApple InsiderBBC NewsSky NewsGoogle Newsxkcddiodesign


    © 1999-2009 The Drobe Team. Some rights reserved, click here for more information
    Powered by MiniDrobeCMS, based on J4U | Statistics
    "I was actually referring to drobe itself as the 'young upstart' rather than the age of the contributors"
    Page generated in 0.2033 seconds.